Drug discovery is an academical and commercial process of global importance. Accurate identification of drug-target interactions (DTIs) can significantly facilitate the drug discovery process. Compared to the costly, labour-intensive and time-consuming experimental methods, machine learning (ML) plays an ever-increasingly important role in effective, efficient and high-throughput identification of DTIs. However, upstream feature extraction methods require tremendous human resources and expert insights, which limited the application of ML approaches. Inspired by the unsupervised representation learning methods like Word2vec, we here proposed SPVec, a novel way to automatically represent raw data such as SMILES strings and protein sequences into continuous, information-rich and lower-dimensional vectors, so as to avoid the sparseness and bit collisions from the cumbersomely manually extracted features. Visualization of SPVec nicely illustrated that the similar compounds or proteins occupy similar vector space, which indicated that SPVec not only encodes compound substructures or protein sequences efficiently, but also implicitly reveals some important biophysical and biochemical patterns. Compared with manually-designed features like MACCS fingerprints and amino acid composition, SPVec showed better performance with several state-of-art machine learning classifiers such as Gradient Boosting Decision Tree, Random Forest and Deep Neural Network on BindingDB. The performance and robustness of SPVec were also confirmed on independent test sets obtained from DrugBank database. Also, based on the whole DrugBank dataset, we predicted the possibilities of all unlabeled DTIs, where two of the top five predicted novel DTIs were supported by external evidences. These results indicated that SPVec can provide an effective and efficient way to discover reliable DTIs, which would be beneficial for drug reprofiling.

Mercuric reductase, MerA, is a key enzyme in bacterial mercury resistance. This homodimeric enzyme captures and reduces toxic Hg²⁺ to Hg⁰, which is relatively unreactive and can exit the cell passively. Prior to reduction, the Hg²⁺ is transferred from a pair of cysteines (C558′ and C559′ using Tn501 numbering) at the C-terminus of one monomer to another pair of cysteines (C136 and C141) in the catalytic site of the other monomer. Here, we present the X-ray structure of the C-terminal Hg²⁺ complex of the C136A/C141A double mutant of the Tn501 MerA catalytic core and explore the molecular mechanism of this Hg transfer with quantum mechanical/molecular mechanical (QM/MM) calculations. The transfer is found to be nearly thermoneutral and to pass through a stable tricoordinated intermediate that is marginally less stable than the two end states. For the overall process, Hg²⁺ is always paired with at least two thiolates and thus is present at both the C-terminal and catalytic binding sites as a neutral complex. Prior to Hg²⁺ transfer, C141 is negatively charged. As Hg²⁺ is transferred into the catalytic site, a proton is transferred from C136 to C559′ while C558′ becomes negatively charged, resulting in the net transfer of a negative charge over a distance of ∼7.5 Å. Thus, the transport of this soft divalent cation is made energetically feasible by pairing a competition between multiple Cys thiols and/or thiolates for Hg²⁺ with a competition between the Hg²⁺ and protons for the thiolates.

The human KRAS (Kirsten rat sarcoma) is an oncogene, involved in the regulation of cell growth and division. The mutations in the KRAS gene have the potential to cause normal cells to become cancerous in human lungs. In the present study, we focus on non-synonymous single nucleotide polymorphisms (nsSNPs), which are point mutations in the DNA sequence leading to the amino acid variants in the encoded protein. To begin with, we developed a pipeline to utilize a set of computational tools in order to obtain the most deleterious nsSNPs (Q22K, Q61P, and Q61R) associated with lung cancer in the human KRAS gene. Furthermore, molecular dynamics simulation and structural analyses of the 3D structures of native and mutant proteins confirmed the impact of these nsSNPs on the stability of the protein. Finally, the experimental results demonstrated that the structural stability of the mutant proteins was worse than that of the native protein. This study provides significant guidance for narrowing down the number of KRAS mutations to be screened as potential diagnostic biomarkers and to better understand the structural and functional mechanisms of the KRAS protein.