Prediction and design of protease enzyme specificity using a structure-aware graph convolutional network.

Site-specific proteolysis by the enzymatic cleavage of small linear sequence motifs is a key posttranslational modification involved in physiology and disease. The ability to robustly and rapidly predict protease-substrate specificity would also enable targeted proteolytic cleavage by designed proteases. Current methods for predicting protease specificity are limited to sequence pattern recognition in experimentally derived cleavage data obtained for libraries of potential substrates and generated separately for each protease variant. We reasoned that a more semantically rich and robust model of protease specificity could be developed by incorporating the energetics of molecular interactions between protease and substrates into machine learning workflows. We present Protein Graph Convolutional Network (PGCN), which develops a physically grounded, structure-based molecular interaction graph representation that describes molecular topology and interaction energetics to predict enzyme specificity. We show that PGCN accurately predicts the specificity landscapes of several variants of two model proteases. Node and edge ablation tests identified key graph elements for specificity prediction, some of which are consistent with known biochemical constraints for protease:substrate recognition. We used a pretrained PGCN model to guide the design of protease libraries for cleaving two noncanonical substrates, and found good agreement with experimental cleavage results. Importantly, the model can accurately assess designs featuring diversity at positions not present in the training data. The described methodology should enable the structure-based prediction of specificity landscapes of a wide variety of proteases and the construction of tailor-made protease editors for site-selectively and irreversibly modifying chosen target proteins.

Prediction and Design of Protease Enzyme Specificity Using a Structure-Aware Graph Convolutional Network.

Site-specific proteolysis by the enzymatic cleavage of small linear sequence motifs is a key post-translational modification involved in physiology and disease. The ability to robustly and rapidly predict protease substrate specificity would also enable targeted proteolytic cleavage - editing - of a target protein by designed proteases. Current methods for predicting protease specificity are limited to sequence pattern recognition in experimentally-derived cleavage data obtained for libraries of potential substrates and generated separately for each protease variant. We reasoned that a more semantically rich and robust model of protease specificity could be developed by incorporating the three-dimensional structure and energetics of molecular interactions between protease and substrates into machine learning workflows. We present Protein Graph Convolutional Network (PGCN), which develops a physically-grounded, structure-based molecular interaction graph representation that describes molecular topology and interaction energetics to predict enzyme specificity. We show that PGCN accurately predicts the specificity landscapes of several variants of two model proteases: the NS3/4 protease from the Hepatitis C virus (HCV) and the Tobacco Etch Virus (TEV) proteases. Node and edge ablation tests identified key graph elements for specificity prediction, some of which are consistent with known biochemical constraints for protease:substrate recognition. We used a pre-trained PGCN model to guide the design of TEV protease libraries for cleaving two non-canonical substrates, and found good agreement with experimental cleavage results. Importantly, the model can accurately assess designs featuring diversity at positions not present in the training data. The described methodology should enable the structure-based prediction of specificity landscapes of a wide variety of proteases and the construction of tailor-made protease editors for site-selectively and irreversibly modifying chosen target proteins.

Integrated cooling (i-Cool) textile of heat conduction and sweat transportation for personal perspiration management.

Perspiration evaporation plays an indispensable role in human body heat dissipation. However, conventional textiles tend to focus on sweat removal and pay little attention to the basic thermoregulation function of sweat, showing limited evaporation ability and cooling efficiency in moderate/profuse perspiration scenarios. Here, we propose an integrated cooling (i-Cool) textile with unique functional structure design for personal perspiration management. By integrating heat conductive pathways and water transport channels decently, i-Cool exhibits enhanced evaporation ability and high sweat evaporative cooling efficiency, not merely liquid sweat wicking function. In the steady-state evaporation test, compared to cotton, up to over 100% reduction in water mass gain ratio, and 3 times higher skin power density increment for every unit of sweat evaporation are demonstrated. Besides, i-Cool shows about 3 °C cooling effect with greatly reduced sweat consumption than cotton in the artificial sweating skin test. The practical application feasibility of i-Cool design principles is well validated based on commercial fabrics. Owing to its exceptional personal perspiration management performance, we expect the i-Cool concept can provide promising design guidelines for next-generation perspiration management textiles.

A new mitochondria-targeting fluorescent probe for ratiometric detection of H2O2 in live cells.

With this research we presented a ratiometric and mitochondria-target fluorescent probe (Mito-HT) for detection of H2O2 both in vitro and in live cells. Mito-HT was constructed by direct conjugation of aryl boronate to fluorophore with three synthetic steps. The borate group is cleaved from Mito-HT in the presence of H2O2, resulting in the exposure of the hydroxyl group of the electron donating group. Then the ICT mechanism was turned on, and the fluorescence emission of Mito-HT at 493 nm was red-shifted to 562 nm, thereby achieving radiometric detection of H2O2. Mito-HT exhibited a highly selectivity towards H2O2, and this interaction can be completed within 40 min. Mito-HT could be used for quantitative detection of H2O2 (0-200 µM) through ratiometric fluorescence signal readout. And limit of detection (LOD) is approximately 0.33 µM. The relatively high stability and medium fluorescence quantum yield of Mito-HT (0.39) and Mito-HT-OH (0.43) enable clear mitochondria localization and dual-channel fluorescence imaging of H2O2 in live cells with confocal microscopy.

A new long-wavelength fluorescent probe for tracking peroxynitrite in live cells and inflammatory sites of zebrafish.

Peroxynitrite (OONO-), as a reactive oxygen species (ROS), would be mostly profoundly implicated in diseases such as inflammation in organisms. However, bioimaging of ONOO- still faces difficulties owing to the shortage of bioimaging and real-time dynamic tracking distribution of ROS in inflammation. To address this challenge, we designed and synthesized a long-wavelength fluorescent probe based on tricyanofuran (ACDM-BE), which exhibits a fast response (response time is 40 s), high selectivity and great sensitivity (LOD is approximately 21 nM) towards ONOO-. ACDM-BE was shown to be capable of detecting ONOO- in living cells and monitor the changes in ONOO- levels under the stimulus of various concentrations of SIN-1 (from 100 to 700 µM), which was successfully tracked by the fluorescence changes in live cells. It is worth noting that ACDM-BE further demonstrated its ability to track the dynamic changes of the level of ONOO- in the inflammatory sites of larval zebrafish. Thus, ACDM-BE could be employed as an efficient tool for exploiting the role of ONOO- in inflammation in living biosystems.

Two-isophorone fluorophore-based design of a ratiometric fluorescent probe and its application in the sensing of biothiols.

2,4-Dinitrobenzenesulfonyl (DNBS) has been widely used for the design of small fluorescent probes for biothiols due to its high reactivity. However, most DNBS-based fluorescent probes exhibit "off-on" fluorescence response towards biothiols due to the strong quenching effects of DBNS on the fluorophores. Herein, we present an alternative design of a ratiometric fluorescent probe based on DNBS for biothiols. A new fluorophore bearing two isophorone malononitrile structures was conjugated with DNBS to provide a target probe (CHT), which exhibited a ratiometric sensing behavior towards biothiols. The sensing process is rapid and highly selective. Most importantly, CHT has high stability in the quantitative detection of Cys compared to the control probe CHM, which performed an "off-on" sensing for biothiols. Endogenous biothiols were successfully monitored with CHT in live cells through the ratiometric fluorescence signal. This new fluorophore bearing two isophorone malononitrile moieties will pave a new avenue to design ratiometric fluorescent probes for imaging and quantitative detection.

Extraction of ferulic acid and vanilla acid by hydrophobic ionic liquid 1-hexyl-3-methylimidazolium hexafluorophosphate.

The study was carried out to evaluate the extraction efficiency of ferulic acid (FA) and vanilla acid (VA) from aqueous phase into IL phase. To achieve the highest extraction efficiency, the influence of varying key parameters was evaluated and optimized by response surface methodology based on Box-Behnken design, including phase volume ratio, extraction temperature and extraction time. FA (or VA) extraction under the optimal conditions were: phase volume ratio of 1.38 (1.28), extraction temperature of 66.34 °C (49.28 °C) and extraction time of 33.83 min (36.64 min) under optimum conditions an average extraction efficiency of 97.11 ± 1.05% for FA was achieved, while VA was 85.43 ± 1.62%. This was very close to the predicted value from the model, 98.05% (86.16%). Additionally, recycling and utilization of ILs were performed well with the recovery ratio for 81.0%. Based on thermodynamic analysis, FTIR and 1H NMR analysis, the combination of hydrophobic interaction and hydrogen-bond interaction resulted in the real extraction result above. It is desirable to provide a useful reference for the separation and purification of FA, VA, and extend the potential application of ionic liquid in the separation of natural active compounds with great prospects.