Modulation of fungal phosphate homeostasis by the plant hormone strigolactone.

Inter-kingdom communication through small molecules is essential to the coexistence of organisms in an ecosystem. In soil communities, the plant root is a nexus of interactions for a remarkable number of fungi and is a source of small-molecule plant hormones that shape fungal compositions. Although hormone signaling pathways are established in plants, how fungi perceive and respond to molecules is unclear because many plant-associated fungi are recalcitrant to experimentation. Here, we develop an approach using the model fungus, Saccharomyces cerevisiae, to elucidate mechanisms of fungal response to plant hormones. Two plant hormones, strigolactone and methyl jasmonate, produce unique transcript profiles in yeast, affecting phosphate and sugar metabolism, respectively. Genetic analysis in combination with structural studies suggests that SLs require the high-affinity transporter Pho84 to modulate phosphate homeostasis. The ability to study small-molecule plant hormones in a tractable genetic system should have utility in understanding fungal-plant interactions.

BIONIC: biological network integration using convolutions.

Biological networks constructed from varied data can be used to map cellular function, but each data type has limitations. Network integration promises to address these limitations by combining and automatically weighting input information to obtain a more accurate and comprehensive representation of the underlying biology. We developed a deep learning-based network integration algorithm that incorporates a graph convolutional network framework. Our method, BIONIC (Biological Network Integration using Convolutions), learns features that contain substantially more functional information compared to existing approaches. BIONIC has unsupervised and semisupervised learning modes, making use of available gene function annotations. BIONIC is scalable in both size and quantity of the input networks, making it feasible to integrate numerous networks on the scale of the human genome. To demonstrate the use of BIONIC in identifying new biology, we predicted and experimentally validated essential gene chemical-genetic interactions from nonessential gene profiles in yeast.

Structural and Functional Analyses of Inhibition of Human Dihydroorotate Dehydrogenase by Antiviral Furocoumavirin.

Natural products are important sources of seed compounds for drug discovery. However, it has become difficult in recent years to discover new compounds with valuable pharmacological activities. On the other hand, among the vast number of natural products that have been isolated so far, a considerable number of compounds with specific biological activities are thought to be overlooked in screening that uses biological activity as an index. Therefore, it is conceivable that such overlooked useful compounds may be found by screening compound libraries that have been amassed previously through specific assays. Previously, NPD723, a member of the Natural Products Depository library comprised of a mixture of natural and non-natural products developed at RIKEN, and its metabolite H-006 were found to inhibit growth of various cancer cells at low nanomolar half-maximal inhibitory concentration. Subsequent analysis revealed that H-006 strongly inhibited human dihydroorotate dehydrogenase (DHODH), the rate-limiting enzyme in the de novo pyrimidine biosynthetic pathway. Here, we elucidated the crystal structure of the DHODH-flavin mononucleotide-orotic acid-H-006 complex at 1.7 Å resolution to determine that furocoumavirin, the S-enantiomer of H-006, was the actual inhibitor. The overall mode of interaction of furocoumavirin with the inhibitor binding pocket was similar to that described for previously reported tight-binding inhibitors. However, the structural information together with kinetic characterizations of site-specific mutants identified key unique features that are considered to contribute to the sub-nanomolar inhibition of DHODH by furocoumavirin. Our finding identified new chemical features that could improve the design of human DHODH inhibitors.

Identification of antimycin A as a c-Myc degradation accelerator via high-throughput screening.

c-Myc is a critical regulator of cell proliferation and growth. Elevated levels of c-Myc cause transcriptional amplification, leading to various types of cancers. Small molecules that specifically inhibit c-Myc-dependent regulation are potentially invaluable for anticancer therapy. Because c-Myc does not have enzymatic activity or targetable pockets, researchers have attempted to obtain small molecules that inhibit c-Myc cofactors, activate c-Myc repressors, or target epigenetic modifications to regulate the chromatin of c-Myc-addicted cancer without any clinical success. In this study, we screened for c-Myc inhibitors using a cell-dependent assay system in which the expression of c-Myc and its transcriptional activity can be inferred from monomeric Keima and enhanced GFP fluorescence, respectively. We identified one mitochondrial inhibitor, antimycin A, as a hit compound. The compound enhanced the c-Myc phosphorylation of threonine-58, consequently increasing the proteasome-mediated c-Myc degradation. The mechanistic analysis of antimycin A revealed that it enhanced the degradation of c-Myc protein through the activation of glycogen synthetic kinase 3 by reactive oxygen species (ROS) from damaged mitochondria. Furthermore, we found that the inhibition of cell growth by antimycin A was caused by both ROS-dependent and ROS-independent pathways. Interestingly, ROS-dependent growth inhibition occurred only in the presence of c-Myc, which may reflect the representative features of cancer cells. Consistently, the antimycin A sensitivity of cells was correlated to the endogenous c-Myc levels in various cancer cells. Overall, our study provides an effective strategy for identifying c-Myc inhibitors and proposes a novel concept for utilizing ROS inducers for cancer therapy.

Amiodarone inhibits the Toll-like receptor 3-mediated nuclear factor κB signaling pathway by blocking organelle acidification.

Toll-like receptor (TLR) agonists or pro-inflammatory cytokines converge to activate the nuclear factor κB (NF-κB) signaling pathway, which provokes inflammatory responses. In the present study, we identified amiodarone hydrochloride as a selective inhibitor of the TLR3-mediated NF-κB signaling pathway by screening the RIKEN NPDepo Chemical Library. In human umbilical vein endothelial cells (HUVEC), amiodarone selectively inhibited the expression of intercellular adhesion molecule-1 (ICAM-1) induced by polyinosinic-polycytidylic acid (Poly(I:C)), but not tumor necrosis factor-α, interleukin-1α, or lipopolysaccharide. In response to a Poly(I:C) stimulation, amiodarone at 20 µM reduced the up-regulation of mRNA expression encoding ICAM-1, vascular cell adhesion molecule-1, and E-selectin. The nuclear translocation of the NF-κB subunit RelA was inhibited by amiodarone at 15-20 µM in Poly(I:C)-stimulated HUVEC. Amiodarone diminished the fluorescent dots of LysoTracker® Red DND-99 scattered over the cytoplasm of HUVEC. Therefore, the present study revealed that amiodarone selectively inhibited the TLR3-mediated NF-κB signaling pathway by blocking the acidification of intracellular organelles.

Total synthesis, stereochemical assignment, and biological evaluation of opantimycin A and analogues thereof.

Opantimycin A, a rare antimycin-class antibiotic without the macrolide core, was isolated from Streptomyces sp. RK88-1355 in 2017. In this study, we explored the total synthesis and stereochemical assignment of opantimycin A. The synthesis of all potential diastereomers has been accomplished via traceless Staudinger ligation. A comparison of the spectroscopic data of the synthesized compounds with that reported for the natural product confirmed that the absolute configuration of the natural product was (14S,17R,21R). Two analogous compounds were prepared, where the Dhb ((Z)-dehydrobutyrine) moiety was replaced with Dha (dehydroalanine) or ΔVal moieties, respectively. The inhibitory activities of these synthetic compounds against the production of the anti-inflammatory cytokine IL-6 were evaluated, and two potential candidates for further development as anti-inflammatory agents were identified.

Production of Kinanthraquinone D with Antimalarial Activity by Heterologous Gene Expression and Biotransformation in Streptomyces lividans TK23.

Two new compounds, kinanthraquinone C (1) and kinanthraquinone D (2), were isolated along with two known compounds, kinanthraquinone (3) and kinanthraquinone B (4), produced by the heterologous expression of the kiq biosynthetic gene cluster and its pathway-specific regulator, kiqA, in Streptomyces lividans TK23. The chemical structures of compounds 1 and 2 were determined using mass spectrometry and nuclear magnetic resonance analyses. To examine a biosynthetic pathway of compounds 1 and 2, incubation experiments were conducted using S. lividans TK23 to supply the compounds 3 and 4. These experiments indicated that compounds 3 and 4 were converted to compounds 2 and 1, respectively, by the endogenous enzymes of S. lividans TK23. Compounds 2, 3, and 4 had antimalarial activities at half-maximal inhibitory concentration values of 0.91, 1.2, and 15 µM, respectively, without cytotoxicity up to 30 µM.

Expression of Syo_1.56 SARP Regulator Unveils Potent Elasnin Derivatives with Antibacterial Activity.

Actinomycetes are prolific producers of natural products, particularly antibiotics. However, a significant proportion of its biosynthetic gene clusters (BGCs) remain silent under typical laboratory conditions. This limits the effectiveness of conventional isolation methods for the discovery of novel natural products. Genetic interventions targeting the activation of silent gene clusters are necessary to address this challenge. Streptomyces antibiotic regulatory proteins (SARPs) act as cluster-specific activators and can be used to target silent BGCs for the discovery of new antibiotics. In this study, the expression of a previously uncharacterized SARP protein, Syo_1.56, in Streptomyces sp. RK18-A0406 significantly enhanced the production of known antimycins and led to the discovery of 12 elasnins (1-12), 10 of which were novel. The absolute stereochemistry of elasnin A1 was assigned for the first time to be 6S. Unexpectedly, Syo_1.56 seems to function as a pleiotropic rather than cluster-specific SARP regulator, with the capability of co-regulating two distinct biosynthetic pathways, simultaneously. All isolated elasnins were active against wild-type and methicillin-resistant Staphylococcus aureus with IC50 values of 0.5-20 µg/mL, some of which (elasnins A1, B2, and C1 and proelasnins A1, and C1) demonstrated moderate to strong antimalarial activities against Plasmodium falciparum 3D7. Elasnins A1, B3, and C1 also showed in vitro inhibition of the metallo-ß-lactamase responsible for the development of highly antibiotic-resistant bacterial strains.

Reclassification of a reveromycin-producer and the proposals of Actinacidiphila reveromycinica sp. nov. and Actinacidiphila acidipaludis comb. nov.

Streptomyces sp. SN-593, a reveromycin producer, was previously thought to belong to the genus Streptomyces based on its morphological and chemotaxonomic characteristics. In this paper, we re-considered its taxonomic position according to the current criteria. Phylogenetic analysis using 16S rRNA gene sequences showed that the strain belongs to the genus Actinacidiphila. In multilocus sequence and phylogenomic analyses, the strain SN-593 represented a distinct evolutionary lineage within this genus, and its closest neighbor was A. yanglinensis. Digital DNA-DNA hybridization indicated that the strain shares less than 32% DNA-DNA relatedness with the type strains of closely related species, confirming this strain is a novel genomospecies. According to its phenotypic distinctiveness from the closest neighbor, we propose Actinacidiphila reveromycinica sp. nov. for the strain SN-593T. Additionally, as Streptomyces acidipaludis belonged to the genus Actinacidiphila in these analyses, it should be transferred to the genus, for which Actinacidiphila acidipaludis comb. nov. is proposed.

Lonidamine and domperidone inhibit expansion of transformed cell areas by modulating motility of surrounding nontransformed cells.

Cancer cells intrinsically proliferate in an autonomous manner; however, the expansion of cancer cell areas in a tissue is known to be regulated by surrounding nontransformed cells. Whether these nontransformed cells can be targeted to control the spread of cancer cells is not understood. In this study, we established a system to evaluate the cancer-inhibitory activity of surrounding nontransformed cells and screened chemical compounds that could induce this activity. Our findings revealed that lonidamine (LND) and domperidone (DPD) inhibited expansion of oncogenic foci of KRASG12D-expressing transformed cells, whereas they did not inhibit the proliferation of monocultured KRASG12D-expressing cells. Live imaging revealed that LND and DPD suppressed the movement of nontransformed cells away from the attaching cancer cells. Moreover, we determined that LND and DPD promoted stress fiber formation, and the dominant-negative mutant of a small GTPase RhoA relieved the suppression of focus expansion, suggesting that RhoA-mediated stress fiber formation is involved in the inhibition of the movement of nontransformed cells and focus expansion. In conclusion, we suggest that elucidation of the mechanism of action of LND and DPD may lead to the development of a new type of drug that could induce the anticancer activity of surrounding nontransformed cells.

N-(3,4-dimethoxyphenethyl)-6-methyl-2,3,4,9-tetrahydro-1H-carbazol-1-amine inhibits bladder cancer progression by suppressing YAP1/TAZ.

Bladder cancer (BlC) is the fourth most common cancer in males worldwide, but few systemic chemotherapy options for its effective treatment exist. The development of new molecularly-targeted agents against BlC is therefore an urgent issue. The Hippo signaling pathway, with its upstream LATS kinases and downstream transcriptional co-activators YAP1 and TAZ, plays a pivotal role in diverse cell functions, including cell proliferation. Recent studies have shown that overexpression of YAP1 occurs in advanced BlCs and is associated with poor patient prognosis. Accessing data from our previous screening of a chemical library of compounds targeting the Hippo pathway, we identified DMPCA (N-(3,4-dimethoxyphenethyl)-6-methyl-2,3,4,9-tetrahydro-1H-carbazol-1-amine) as an agent able to induce the phosphorylation of LATS1 and YAP1/TAZ in BlC cells, thereby suppressing their viability both in vitro and in mouse xenografts. Our data indicate that DMPCA has a potent anti-tumor effect, and raise the possibility that this agent may represent a new and effective therapeutic option for BlC.

Establishment of a Monoclonal Antibody against Human NTCP That Blocks Hepatitis B Virus Infection.

Hepatitis B virus (HBV) infects 240 million people worldwide. Current therapy profoundly suppresses HBV replication but requires long-term maintenance therapy. Therefore, there is still a medical need for an efficient HBV cure. HBV enters host cells by binding via the preS1 domain of the viral L protein to the Na+/taurocholate cotransporting polypeptide (NTCP). Thus, NTCP should be a key target for the development of anti-HBV therapeutics. Indeed, myrcludex B, a synthetic form of the myristoylated preS1 peptide, effectively reduces HBV/hepatitis D virus (HDV) infection and has been approved as Hepcludex in Europe for the treatment of patients with chronic HDV infection. We established a monoclonal antibody (MAb), N6HB426-20, that recognizes the extracellular domain of human NTCP and blocks HBV entry in vitro into human liver cells but has much less of an inhibitory effect on bile acid uptake. In vivo, administration of the N6HB426-20 MAb prevented HBV viremia for an extended period of time after HBV inoculation in a mouse model system without strongly inhibiting bile acid absorption. Among the extracellular loops (ECLs) of NTCP, regions of amino acids (aa) 84 to 87 in ECL1 and aa 157 to 165 near ECL2 of transmembrane domain 5 are critically important for HBV/HDV infection. Epitope mapping and the three-dimensional (3D) model of the NTCP structure suggested that the N6HB426-20 MAb may recognize aa 276/277 at the tip of ECL4 and interfere with binding of HBV to the region from aa 84 to 87. In summary, we identified an in vivo neutralizing NTCP-targeting antibody capable of preventing HBV infection. Further improvements in efficacy of this drug will pave the way for its clinical applications. IMPORTANCE A number of entry inhibitors are being developed to enhance the treatment of HBV patients with oral nucleoside/nucleotide analogues (NA). To amplify the effectiveness of NA therapy, several efforts have been made to develop therapeutic MAbs with neutralizing activity against HBs antigens. However, the neutralizing effect of these MAbs may be muted by a large excess of HBsAg-positive noninfectious particles in the blood of infected patients. The advantage of NTCP-targeted HBV entry inhibitors is that they remain effective regardless of viral genotype, viral mutations, and the presence of subviral particles. Although N6HB426-20 requires a higher dose than myrcludex to obtain equivalent suppression of HBV in a model mouse system, it maintained the inhibitory effect for a long time postadministration in proportion to the half-life of an IgG MAb. We believe that further improvements will make this antibody a promising treatment option for patients with chronic hepatitis B.

Chemical reversal of abnormalities in cells carrying mitochondrial DNA mutations.

Mitochondrial DNA (mtDNA) mutations are the major cause of mitochondrial diseases. Cells harboring disease-related mtDNA mutations exhibit various phenotypic abnormalities, such as reduced respiration and elevated lactic acid production. Induced pluripotent stem cell (iPSC) lines derived from patients with mitochondrial disease, with high proportions of mutated mtDNA, exhibit defects in maturation into neurons or cardiomyocytes. In this study, we have discovered a small-molecule compound, which we name tryptolinamide (TLAM), that activates mitochondrial respiration in cybrids generated from patient-derived mitochondria and fibroblasts from patient-derived iPSCs. We found that TLAM inhibits phosphofructokinase-1 (PFK1), which in turn activates AMPK-mediated fatty-acid oxidation to promote oxidative phosphorylation, and redirects carbon flow from glycolysis toward the pentose phosphate pathway to reinforce anti-oxidative potential. Finally, we found that TLAM rescued the defect in neuronal differentiation of iPSCs carrying a high ratio of mutant mtDNA, suggesting that PFK1 represents a potential therapeutic target for mitochondrial diseases.

Identification of a derivative of the alkaloid emetine as an inhibitor of the YAP-TEAD interaction and its potential as an anticancer agent.

TEAD is a transcription factor responsible for the output of the tumor suppressor Hippo pathway. The transcriptional activity of TEAD requires molecular interaction with its transcriptional coactivator, YAP. Aberrant activation of TEAD is deeply involved in tumorigenesis and is associated with poor prognosis, suggesting that inhibitors targeting the YAP-TEAD system are promising as antitumor agents. In this study, we identified NPD689, an analog of the natural product alkaloid emetine, as an inhibitor of the YAP-TEAD interaction. NPD689 suppressed the transcriptional activity of TEAD and reduced the viability of human malignant pleural mesothelioma and non-small cell lung cancer cells but not the viability of normal human mesothelial cells. Our results suggest that NPD689 is not only a new useful chemical tool for elucidating the biological role of the YAP-TEAD system but also has potential as a starting compound for developing a cancer therapeutic agent that targets the YAP-TEAD interaction.

CSE1L promotes nuclear accumulation of transcriptional coactivator TAZ and enhances invasiveness of human cancer cells.

The transcriptional coactivator with PDZ-binding motif (TAZ) (WWTR1) induces epithelial-mesenchymal transition and enhances drug resistance in multiple cancers. TAZ has been shown to interact with transcription factors in the nucleus, but when phosphorylated, translocates to the cytoplasm and is degraded through proteasomes. Here, we identified a compound TAZ inhibitor 4 (TI-4) that shifted TAZ localization to the cytoplasm independently of its phosphorylation. We used affinity beads to ascertain a putative target of TI-4, chromosomal segregation 1 like (CSE1L), which is known to be involved in the recycling of importin α and as a biomarker of cancer malignancy. We found that TI-4 suppressed TAZ-mediated transcription in a CSE1L-dependent manner. CSE1L overexpression increased nuclear levels of TAZ, whereas CSE1L silencing delayed its nuclear import. We also found via the in vitro coimmunoprecipitation experiments that TI-4 strengthened the interaction between CSE1L and importin α5 and blocked the binding of importin α5 to TAZ. WWTR1 silencing attenuated CSE1L-promoted colony formation, motility, and invasiveness of human lung cancer and glioblastoma cells. Conversely, CSE1L silencing blocked TAZ-promoted colony formation, motility, and invasiveness in human lung cancer and glioblastoma cells. In human cancer tissues, the expression level of CSE1L was found to correlate with nuclear levels of TAZ. These findings support that CSE1L promotes the nuclear accumulation of TAZ and enhances malignancy in cancer cells.

New antimalarials identified by a cell-based phenotypic approach: Structure-activity relationships of 2,3,4,9-tetrahydro-1H-ß-carboline derivatives possessing a 2-((coumarin-5-yl)oxy)alkanoyl moiety.

The identification, structure-activity relationships (SARs), and biological effects of new antimalarials consisting of a 2,3,4,9-tetrahydro-1H-ß-carboline core, a coumarin ring, and an oxyalkanoyl linker are described. A cell-based phenotypic approach was employed in this search for novel antimalarial drugs with unique modes of action. Our screening campaign of the RIKEN compound library succeeded in the identification of the known tetrahydro-ß-carboline derivative (4e) as a hit compound showing significant in vitro activity. SAR studies on this chemical series led to the discovery of compound 4h having a (R)-methyl group on the oxyacetyl linker with potent inhibition of parasite growth (IC50 = 2.0 nM). Compound 4h was also found to exhibit significant in vivo antimalarial effects in mouse models. Furthermore, molecular modeling studies on 4e, 4h, and its diastereomer (4j) suggested that the (R)-methyl group of 4h forces the preferential adoption of a specific conformer which is considered to be an active conformer.

Dihydrolucilactaene, a Potent Antimalarial Compound from Fusarium sp. RK97-94.

A recently discovered secondary metabolism regulator, NPD938, was used to alter the secondary metabolite profile in Fusarium sp. RK97-94. Three lucilactaene analogues were detected via UPLC-ESI-MS analysis in NPD938-treated culture. The three metabolites were successfully purified and identified as dihydroNG391 (1), dihydrolucilactaene (2), and 13α-hydroxylucilactaene (3) via extensive spectroscopic analyses. DihydroNG391 (1) exhibited weak in vitro antimalarial activity (IC50 = 62 µM). In contrast, dihydrolucilactaene (2) and 13α-hydroxylucilactaene (3) showed very potent antimalarial activity (IC50 = 0.0015 and 0.68 µM, respectively). These findings provide insight into the structure-activity relationship of lucilactaene and its analogues as antimalarial lead compounds.

Unique features of the ketosynthase domain in a nonribosomal peptide synthetase-polyketide synthase hybrid enzyme, tenuazonic acid synthetase 1.

Many microbial secondary metabolites are produced by multienzyme complexes comprising nonribosomal peptide synthetases (NRPSs) and polyketide synthases (PKSs). The ketosynthase (KS) domains of polyketide synthase normally catalyze the decarboxylative Claisen condensation of acyl and malonyl blocks to extend the polyketide chain. However, the terminal KS domain in tenuazonic acid synthetase 1 (TAS1) from the fungus Pyricularia oryzae conducts substrate cyclization. Here, we report on the unique features of the KS domain in TAS1. We observed that this domain is monomeric, not dimeric as is typical for KSs. Analysis of a 1.68-Å resolution crystal structure suggests that the substrate cyclization is triggered via proton abstraction from the active methylene moiety in the substrate by a catalytic His-322 residue. Additionally, we show that TAS1 KS promiscuously accepts aminoacyl substrates and that this promiscuity can be increased by a single amino acid substitution in the substrate-binding pocket of the enzyme. These findings provide insight into a KS domain that accepts the amino acid-containing substrate in an NRPS-PKS hybrid enzyme and provide hints to the substrate cyclization mechanism performed by the KS domain in the biosynthesis of the mycotoxin tenuazonic acid.

Alantolactone is a natural product that potently inhibits YAP1/TAZ through promotion of reactive oxygen species accumulation.

Yes-associated protein 1 (YAP1) and its paralogue PDZ-binding motif (TAZ) play pivotal roles in cell proliferation, migration, and invasion, and abnormal activation of these TEAD transcriptional coactivators is found in diverse cancers in humans and mice. Targeting YAP1/TAZ signaling is thus a promising therapeutic avenue but, to date, few selective YAP1/TAZ inhibitors have been effective against cancer cells either in vitro or in vivo. We screened chemical libraries for potent YAP1/TAZ inhibitors using a highly sensitive luciferase reporter system to monitor YAP1/TAZ-TEAD transcriptional activity in cells. Among 29 049 low-molecular-weight compounds screened, we obtained nine hits, and the four of these that were the most effective shared a core structure with the natural product alantolactone (ALT). We also tested 16 other structural derivatives of ALT and found that natural ALT was the most efficient at increasing ROS-induced LATS kinase activities and thus YAP1/TAZ phosphorylation. Phosphorylated YAP1/TAZ proteins were subject to nuclear exclusion and proteosomic degradation such that the growth of ALT-treated tumor cells was inhibited both in vitro and in vivo. Our data show for the first time that ALT can be used to target the ROS-YAP pathway driving tumor cell growth and so could be a potent anticancer drug.

Improving Measures of Chemical Structural Similarity Using Machine Learning on Chemical-Genetic Interactions.

A common strategy for identifying molecules likely to possess a desired biological activity is to search large databases of compounds for high structural similarity to a query molecule that demonstrates this activity, under the assumption that structural similarity is predictive of similar biological activity. However, efforts to systematically benchmark the diverse array of available molecular fingerprints and similarity coefficients have been limited by a lack of large-scale datasets that reflect biological similarities of compounds. To elucidate the relative performance of these alternatives, we systematically benchmarked 11 different molecular fingerprint encodings, each combined with 13 different similarity coefficients, using a large set of chemical-genetic interaction data from the yeast Saccharomyces cerevisiae as a systematic proxy for biological activity. We found that the performance of different molecular fingerprints and similarity coefficients varied substantially and that the all-shortest path fingerprints paired with the Braun-Blanquet similarity coefficient provided superior performance that was robust across several compound collections. We further proposed a machine learning pipeline based on support vector machines that offered a fivefold improvement relative to the best unsupervised approach. Our results generally suggest that using high-dimensional chemical-genetic data as a basis for refining molecular fingerprints can be a powerful approach for improving prediction of biological functions from chemical structures.