Drug target identification is a crucial step in development, yet is also among the most complex. To address this, we develop BANDIT, a Bayesian machine-learning approach that integrates multiple data types to predict drug binding targets. Integrating public data, BANDIT benchmarked a ~90% accuracy on 2000+ small molecules. Applied to 14,000+ compounds without known targets, BANDIT generated ~4,000 previously unknown molecule-target predictions. From this set we validate 14 novel microtubule inhibitors, including 3 with activity on resistant cancer cells. We applied BANDIT to ONC201—an anti-cancer compound in clinical development whose target had remained elusive. We identified and validated DRD2 as ONC201’s target, and this information is now being used for precise clinical trial design. Finally, BANDIT identifies connections between different drug classes, elucidating previously unexplained clinical observations and suggesting new drug repositioning opportunities. Overall, BANDIT represents an efficient and accurate platform to accelerate drug discovery and direct clinical application.

The sarcin-ricin loop (SRL) is one of the longest conserved sequences in the 23S rRNA. The SRL has been accepted as crucial for the activity of the ribosome because it is targeted by cytotoxins such as α-sarcin and ricin that completely abolish translation. Nevertheless, the precise functional role of the SRL in translation is not known. Recent biochemical and structural studies indicate that the SRL is critical for triggering GTP hydrolysis on elongation factors Tu and G (EF-Tu and EF-G). To determine the functional role of the SRL in the elongation stage of protein synthesis, we analyzed mutations in the SRL that are known to abolish protein synthesis and are lethal to cells. Here, we show that the SRL is not critical for GTP hydrolysis on EF-Tu and EF-G. The SRL also is not essential for peptide bond formation. Our results, instead, suggest that the SRL is crucial for anchoring EF-G on the ribosome during mRNA-tRNA translocation.

During protein synthesis, mRNA and tRNAs are iteratively translocated by the ribosome. Precisely what molecular event is rate limiting for translocation is not known. Here we show that disruption of the interactions between the A-site codon and the ribosome accelerates translocation, suggesting that the release of the mRNA from the decoding center of the ribosome is the rate-limiting step of translocation. These results provide insight into the molecular mechanism of translocation.

The selection of aminoacyl-tRNAs by the ribosome is a fundamental step in the elongation cycle of protein synthesis. tRNA selection is a multistep process that ensures only correct aminoacyltRNAs are accepted, while incorrect aminoacyl-tRNAs are rejected. A key step in tRNA selection is the formation of base pairs between the anticodon of the aminoacyl-tRNA and the mRNA codon in the A site, called "codon recognition". Here, we report the development of a new, fluorescence-based, kinetic assay to monitor codon recognition by the ribosome. Using this assay we show that codon recognition is a second-order binding step under optimal conditions. Additionally, we show that at low Mg 2+ concentration, the polyamines spermine and spermidine stimulate codon recognition by the ribosome without loss of fidelity. Polyamines may accelerate codon recognition by altering the structure and dynamics of the anticodon arm of the aminoacyltRNA. KeywordsRibosome; elongation factor Tu; decoding; translation In all living organisms ribosomes decode the information in the mRNA to synthesize the corresponding polypeptide. During this process, the ribosome binds a ternary complex consisting of an aminoacyl-tRNA, an elongation factor (elongation factor Tu in E. coli) and a GTP to the A site. The cognate ternary complex is rapidly accommodated into the ribosome and participates in peptidyl transfer, while near-cognate and non-cognate ternary complexes fail to be accommodated and are rejected. This process is called tRNA selection and has to be sufficiently accurate in order to synthesize functional proteins. The error rate of translation in vivo has been estimated to be on the order of 10 -3 to 10 -4 (1,2). However, the difference in the free energy of binding between cognate and some near-cognate aminoacyl-tRNAs are small and does not explain how this high level of fidelity is achieved during protein synthesis.The high fidelity of protein synthesis can be explained by a network of ribosome contacts that recognize the cognate codon-anticodon complex and by a conformational change in the decoding center of the ribosome. Structural studies revealed that the universally conserved bases G530, A1492 and A1493 in 16S rRNA precisely monitor the geometry of the codon- NIH Public Access Author ManuscriptBiochemistry. Author manuscript; available in PMC 2011 August 24. NIH-PA Author ManuscriptNIH-PA Author Manuscript NIH-PA Author Manuscriptanticodon helix in the 30S subunit decoding center (3). These interactions are first established when the ternary complex binds to the ribosome (in the A/T state) and persist even after the aminoacyl-tRNA has been fully accommodated into the A site (in the A/A state) (4-6). The codon-anticodon base pairs in the A site trigger a conformational change in the ribosome called "domain closure" that is proposed to play a key role in accepting the cognate aminoacyl-tRNA (7). In contrast, near-cognate and non-cognate aminoacyl-tRNAs have mismatches between the codon and the anticodon that distorts the helix bac...

Accurate tRNA selection by the ribosome is essential for the synthesis of functional proteins. Previous structural studies indicated that the ribosome distinguishes between cognate and near-cognate tRNAs by monitoring the geometry of the codon-anticodon helix in the decoding center using the universally conserved 16S rRNA bases G530, A1492 and A1493. These bases form hydrogen bonds with the 2’-hydroxyl groups of the codon-anticodon helix, which are expected to be disrupted with a near-cognate codon-anticodon helix. However, a recent structural study showed that G530, A1492 and A1493 form hydrogen bonds in an identical manner with both cognate and near-cognate codon-anticodon helices. To understand how the ribosome discriminates between cognate and near-cognate tRNAs, we made 2’-deoxynucleotide and 2’-fluoro substituted mRNAs, which disrupt the hydrogen bonds between the A site codon and G530, A1492 and A1493. Our results show that multiple 2’-deoxynucleotide substitutions in the mRNA substantially inhibit tRNA selection, whereas multiple 2’-fluoro substitutions in the mRNA have only modest effects on tRNA selection. Furthermore, the miscoding antibiotics paromomycin and streptomycin rescue the defects in tRNA selection with the multiple 2’-deoxynucleotide substituted mRNA. These results suggest that steric complementarity in the decoding center is more important than the hydrogen bonds between the A site codon and G530, A1492 and A1493 for tRNA selection.

No abstract

No abstract

22Drug target identification is one of the most important aspects of pre-clinical development yet it is 23 also among the most complex, labor-intensive, and costly. This represents a major issue, as lack