DeepReac+: deep active learning for quantitative modeling of organic chemical reactions.

Various computational methods have been developed for quantitative modeling of organic chemical reactions; however, the lack of universality as well as the requirement of large amounts of experimental data limit their broad applications. Here, we present DeepReac+, an efficient and universal computational framework for prediction of chemical reaction outcomes and identification of optimal reaction conditions based on deep active learning. Under this framework, DeepReac is designed as a graph-neural-network-based model, which directly takes 2D molecular structures as inputs and automatically adapts to different prediction tasks. In addition, carefully-designed active learning strategies are incorporated to substantially reduce the number of necessary experiments for model training. We demonstrate the universality and high efficiency of DeepReac+ by achieving the state-of-the-art results with a minimum of labeled data on three diverse chemical reaction datasets in several scenarios. Collectively, DeepReac+ has great potential and utility in the development of AI-aided chemical synthesis. DeepReac+ is freely accessible at https://github.com/bm2-lab/DeepReac.

iDMer: an integrative and mechanism-driven response system for identifying compound interventions for sudden virus outbreak.

Emerging viral infections seriously threaten human health globally. Several challenges exist in identifying effective compounds against viral infections: (1) at the initial stage of a new virus outbreak, little information, except for its genome information, may be available; (2) although the identified compounds may be effective, they may be toxic in vivo and (3) cytokine release syndrome (CRS) triggered by viral infections is the primary cause of mortality. Currently, an integrative tool that takes all those aspects into consideration for identifying effective compounds to prevent viral infections is absent. In this study, we developed iDMer, as an integrative and mechanism-driven response system for addressing these challenges during the sudden virus outbreaks. iDMer comprises three mechanism-driven compound identification modules, that is, a virus-host interaction-oriented module, an autophagy-oriented module and a CRS-oriented module. As a one-stop integrative platform, iDMer incorporates compound toxicity evaluation and compound combination identification for virus treatment with clear mechanisms. iDMer was successfully tested on five viruses, including the current severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Our results indicated that, for all five tested viruses, compounds that were reported in the literature or experimentally validated for virus treatment were enriched at the top, demonstrating the generalized effectiveness of iDMer. Finally, we demonstrated that combinations of the individual modules successfully identified combinations of compounds effective for virus intervention with clear mechanisms.

ASNEO: Identification of personalized alternative splicing based neoantigens with RNA-seq.

Cancer neoantigens have shown great potential in immunotherapy, while current software focuses on identifying neoantigens which are derived from SNVs, indels or gene fusions. Alternative splicing widely occurs in tumor samples and it has been proven to contribute to the generation of candidate neoantigens. Here we present ASNEO, which is an integrated computational pipeline for the identification of personalized Alternative Splicing based NEOantigens with RNA-seq. Our analyses showed that ASNEO could identify neopeptides which are presented by MHC I complex through mass spectrometry data validation. When ASNEO was applied to two immunotherapy-treated cohorts, we found that alternative splicing based neopeptides generally have a higher immune score than that of somatic neopeptides and alternative splicing based neopeptides could be a marker to predict patient survival pattern. Our identification of alternative splicing derived neopeptides would contribute to a more complete understanding of the tumor immune landscape. Prediction of patient-specific alternative splicing neopeptides has the potential to contribute to the development of personalized cancer vaccines.

Mechanistic Insights into Recognitions of Ubiquitin and Myosin VI by Autophagy Receptor TAX1BP1.

TAX1BP1, a ubiquitin-binding adaptor, plays critical roles in the innate immunity and selective autophagy. During autophagy, TAX1BP1 may not only function as an autophagy receptor to recruit ubiquitylated substrates for autophagic degradation, but also serve as a Myosin VI cargo adaptor protein for mediating the maturation of autophagosome. However, the mechanistic basis underlying the specific interactions of TAX1BP1 with ubiquitin and Myosin VI remains elusive. Here, using biochemical, NMR and structural analyses, we elucidate the detailed binding mechanism and uncover the key determinants for the interaction between TAX1BP1 and ubiquitin. In addition, we reveal that both tandem zinc-fingers of TAX1BP1 and the conformational rigidity between them are required for the Myosin VI binding of TAX1BP1, and ubiquitin and Myosin VI are mutually exclusive in binding to TAX1BP1. Collectively, our findings provide mechanistic insights into the dual functions of TAX1BP1 in selective autophagy.

Structural Insights into a Flavin-Dependent [4 + 2] Cyclase that Catalyzes trans-Decalin Formation in Pyrroindomycin Biosynthesis.

Here, we provide structural insights into PyrE3, a flavin-dependent [4 + 2] cyclase that catalyzes trans-decalin formation in the biosynthesis of pyrroindomycins. PyrE3 shares an architecture/domain organization head-to-tail similarity with the members of the family of para-hydroxybenzoate hydroxylase (pHBH)-fold monooxygenases, and possesses a flavin adenine dinucleotide (FAD)-binding domain, a middle domain, and a C-terminal thioredoxin-like domain. The FAD-binding domain forms a central hub of the protein structure, and binds with FAD in a "closed" conformation of pHBH-fold family monooxygenases known for their highly dynamic catalytic processes. FAD plays an essential structural role in PyrE3, where it is amenable to redox change; however, redox change has little effect on [4 + 2] cyclization activity. PyrE3 appears to selectively accommodate a tetramate-containing, linear polyene intermediate in a highly positively charged pocket, which is located at the interface between the FAD-binding domain and the middle domain, and can accelerate trans-decalin formation likely through an endo-selective [4 + 2] transition state.

Structural insights into the ubiquitin recognition by OPTN (optineurin) and its regulation by TBK1-mediated phosphorylation.

OPTN (optineurin), a ubiquitin-binding scaffold protein, functions as an important macroautophagy/autophagy receptor in selective autophagy processes. Mutations in OPTN have been linked with human neurodegenerative diseases including ALS and glaucoma. However, the mechanistic basis underlying the recognition of ubiquitin by OPTN and its regulation by TBK1-mediated phosphorylation are still elusive. Here, we demonstrate that the UBAN domain of OPTN preferentially recognizes linear ubiquitin chain and forms an asymmetric 2:1 stoichiometry complex with the linear diubiquitin. In addition, our results provide new mechanistic insights into how phosphorylation of UBAN would regulate the ubiquitin-binding ability of OPTN and how disease-associated mutations in the OPTN UBAN domain disrupt its interaction with ubiquitin. Finally, we show that defects in ubiquitin-binding may affect the recruitment of OPTN to linear ubiquitin-decorated mutant Huntington protein aggregates. Taken together, our findings clarify the interaction mode between UBAN and linear ubiquitin chain in general, and expand our knowledge of the molecular mechanism of ubiquitin-decorated substrates recognition by OPTN as well as the pathogenesis of neurodegenerative diseases caused by OPTN mutations.

Structural Insights into SHARPIN-Mediated Activation of HOIP for the Linear Ubiquitin Chain Assembly.

The linear ubiquitin chain assembly complex (LUBAC) is the sole identified E3 ligase complex that catalyzes the formation of linear ubiquitin chain, and it is composed of HOIP, HOIL-1L, and SHARPIN. The E3 activity of HOIP can be effectively activated by HOIL-1L or SHARPIN, deficiency of which leads to severe immune system disorders. However, the underlying mechanism governing the HOIP-SHARPIN interaction and the SHARPIN-mediated activation of HOIP remains elusive. Here, we biochemically and structurally demonstrate that the UBL domain of SHARPIN specifically binds to the UBA domain of HOIP and thereby associates with and activates HOIP. We further uncover that SHARPIN and HOIL-1L can separately or synergistically bind to distinct sites of HOIP UBA with induced allosteric effects and thereby facilitate the E2 loading of HOIP for its activation. Thus, our findings provide mechanistic insights into the assembly and activation of LUBAC.

Hydroxyl regioisomerization of anthracycline catalyzed by a four-enzyme cascade.

Ranking among the most effective anticancer drugs, anthracyclines represent an important family of aromatic polyketides generated by type II polyketide synthases (PKSs). After formation of polyketide cores, the post-PKS tailoring modifications endow the scaffold with various structural diversities and biological activities. Here we demonstrate an unprecedented four-enzyme-participated hydroxyl regioisomerization process involved in the biosynthesis of kosinostatin. First, KstA15 and KstA16 function together to catalyze a cryptic hydroxylation of the 4-hydroxyl-anthraquinone core, yielding a 1,4-dihydroxyl product, which undergoes a chemically challenging asymmetric reduction-dearomatization subsequently acted by KstA11; then, KstA10 catalyzes a region-specific reduction concomitant with dehydration to afford the 1-hydroxyl anthraquinone. Remarkably, the shunt product identifications of both hydroxylation and reduction-dehydration reactions, the crystal structure of KstA11 with bound substrate and cofactor, and isotope incorporation experiments reveal mechanistic insights into the redox dearomatization and rearomatization steps. These findings provide a distinguished tailoring paradigm for type II PKS engineering.

Structural insights into the interaction and disease mechanism of neurodegenerative disease-associated optineurin and TBK1 proteins.

Optineurin is an important autophagy receptor involved in several selective autophagy processes, during which its function is regulated by TBK1. Mutations of optineurin and TBK1 are both associated with neurodegenerative diseases. However, the mechanistic basis underlying the specific interaction between optineurin and TBK1 is still elusive. Here we determine the crystal structures of optineurin/TBK1 complex and the related NAP1/TBK1 complex, uncovering the detailed molecular mechanism governing the optineurin and TBK1 interaction, and revealing a general binding mode between TBK1 and its associated adaptor proteins. In addition, we demonstrate that the glaucoma-associated optineurin E50K mutation not only enhances the interaction between optineurin and TBK1 but also alters the oligomeric state of optineurin, and the ALS-related TBK1 E696K mutation specifically disrupts the optineurin/TBK1 complex formation but has little effect on the NAP1/TBK1 complex. Thus, our study provides mechanistic insights into those currently known disease-causing optineurin and TBK1 mutations found in patients.

Structural basis of FYCO1 and MAP1LC3A interaction reveals a novel binding mode for Atg8-family proteins.

FYCO1 (FYVE and coiled-coil domain containing 1) functions as an autophagy adaptor in directly linking autophagosomes with the microtubule-based kinesin motor, and plays an essential role in the microtubule plus end-directed transport of autophagic vesicles. The specific association of FYCO1 with autophagosomes is mediated by its interaction with Atg8-family proteins decorated on the outer surface of autophagosome. However, the mechanistic basis governing the interaction between FYCO1 and Atg8-family proteins is largely unknown. Here, using biochemical and structural analyses, we demonstrated that FYCO1 contains a unique LC3-interacting region (LIR), which discriminately binds to mammalian Atg8 orthologs and preferentially binds to the MAP1LC3A and MAP1LC3B. In addition to uncovering the detailed molecular mechanism underlying the FYCO1 LIR and MAP1LC3A interaction, the determined FYCO1-LIR-MAP1LC3A complex structure also reveals a unique LIR binding mode for Atg8-family proteins, and demonstrates, first, the functional relevance of adjacent sequences C-terminal to the LIR core motif for binding to Atg8-family proteins. Taken together, our findings not only provide new mechanistic insight into FYCO1-mediated transport of autophagosomes, but also expand our understanding of the interaction modes between LIR motifs and Atg8-family proteins in general.