Enhancing Generic Reaction Yield Prediction through Reaction Condition-Based Contrastive Learning.

Deep learning (DL)-driven efficient synthesis planning may profoundly transform the paradigm for designing novel pharmaceuticals and materials. However, the progress of many DL-assisted synthesis planning (DASP) algorithms has suffered from the lack of reliable automated pathway evaluation tools. As a critical metric for evaluating chemical reactions, accurate prediction of reaction yields helps improve the practicality of DASP algorithms in the real-world scenarios. Currently, accurately predicting yields of interesting reactions still faces numerous challenges, mainly including the absence of high-quality generic reaction yield datasets and robust generic yield predictors. To compensate for the limitations of high-throughput yield datasets, we curated a generic reaction yield dataset containing 12 reaction categories and rich reaction condition information. Subsequently, by utilizing 2 pretraining tasks based on chemical reaction masked language modeling and contrastive learning, we proposed a powerful bidirectional encoder representations from transformers (BERT)-based reaction yield predictor named Egret. It achieved comparable or even superior performance to the best previous models on 4 benchmark datasets and established state-of-the-art performance on the newly curated dataset. We found that reaction-condition-based contrastive learning enhances the model's sensitivity to reaction conditions, and Egret is capable of capturing subtle differences between reactions involving identical reactants and products but different reaction conditions. Furthermore, we proposed a new scoring function that incorporated Egret into the evaluation of multistep synthesis routes. Test results showed that yield-incorporated scoring facilitated the prioritization of literature-supported high-yield reaction pathways for target molecules. In addition, through meta-learning strategy, we further improved the reliability of the model's prediction for reaction types with limited data and lower data quality. Our results suggest that Egret holds the potential to become an essential component of the next-generation DASP tools.

Design and synthesis of isatin derivatives as effective SARS-CoV-2 3CL protease inhibitors.

SARS-CoV-2 chymotrypsin-like cysteine protease (3CLpro ) is one of the most widely developed drug targets for COVID-19. This study aimed to design and synthesize isatin derivatives to target SARS-CoV-2 3CLpro in a covalent binding manner. Through the process, a potent 3CLpro inhibitor (5g) was discovered with an IC50 value of 0.43 ± 0.17 µM. To understand the binding affinity and specificity of 5g as a candidate inhibitor of SARS-CoV-2 3CLpro , several assays were conducted, including FRET enzyme activity assays, thermodynamic-based and kinetic-based validation of inhibitor-target interactions, and cell-based FlipGFP assays. The interaction mechanism between 3CLpro -5g was characterized by docking. Overall, these findings suggest that 5g is a new potent SARS-CoV-2 3CLpro inhibitor for the treatment of COVID-19.

In silico screening of off-line comprehensive two-dimensional counter-current chromatography with liquid chromatography for four saponins isolation.

Being restrained by the limited peak capacity, one-dimensional chromatography usually leads to an unsatisfactory separation with low purity of compounds in a complex mixture. To obtain more highly pure targets for standard reference and to discover new substances for structural elucidation, two-dimensional chromatography is more and more prevalent in many fields. As few metrics on assessment of the preparative capability of two-dimensional chromatographic separations are reported, a methodology of in silico screening of various two-dimensional chromatographic separations with a minimal number of experiments was demonstrated in this work, which was based on three descriptors including the occupation rate of peaks and system homogeneity of a two-dimensional separation space, and the minimal distance of all nearest-neighbor distances of peaks. Combining the advantages of counter-current chromatography and liquid chromatography, we elaborated the methodology by employing off-line comprehensive two-dimensional counter-current chromatography with liquid chromatography to be in silico screened for separation of four saponins from Panax notoginseng at an analytical scale to simulate the case of preparative scale transfer. The predictive results were presented by two-dimensional contour plots and verified by experiments. The result showed that the experimental results were in general accord with the predictive results.

Liquid chromatographic study of two structural isomeric pentacyclic triterpenes on reversed-phase stationary phase with hydroxypropyl-ß-cyclodextrin as mobile phase additive.

Retention behavior of two structural isomeric pentacyclic triterpenic acids, maslinic acid and corosolic acid, was investigated by reverse phase high performance liquid chromatography (HPLC) with hydroxypropyl-ß-cyclodextrin (HP-ß-CD) as mobile phase additive. Inclusion complexation of maslinic acid, corosolic acid with hydroxypropyl-ß-cyclodextrin was evaluated under different concentration of hydroxypropyl-ß-cyclodextrin. Apparent formation constant (Km) between methanol and hydroxypropyl-ß-cyclodextrin was determined to be 13.82 L mol-1 under 25 °C using UV-spectrophotometry. Two retention models were employed individually for evaluation of inclusion complexation between the two pentacyclic triterpenic acids and hydroxypropyl-ß-cyclodextrin. It was found that a higher apparent formation constant (Kf) for corosolic acid and hydroxypropyl-ß-cyclodextrin was obtained, 19115 L mol-1, indicating that a greater affinity of hydroxypropyl-ß-cyclodextrin with corosolic acid was produced compared with that of maslinic acid, 11775 L mol-1, in the selected mobile phase, and stoichiometric ratio for both of inclusion complex was found to be 1:1. Thermodynamic analysis showed that a negative standard enthalpy change (ΔH) and an entropy change (ΔS*) for analyte transfer were obtained, where ΔH of maslinic acid and corosolic acid was found to be -10.188 kJ mol-1 and -10.650 kJ mol-1, ΔS* of two compounds was -2.092 and -2.180, respectively, indicating that transfer of structural isomers from mobile phase to stationary phase was enthalpically driven. Meanwhile, positive values were obtained for standard enthalpy change and standard entropy change, 136 kJ mol-1 and 274 kJ mol-1 and 536 J mol-1 K-1and 1004 J mol-1 K-1, for inclusion complexation between maslinic acid, corosolic acid and hydroxypropyl-ß-cyclodextrin, while negative values were obtained for Gibbs free energy during formation of inclusion complex, -160 kJ mol-1 and -299 kJ mol-1, indicating a spontaneous inclusion reaction happened.

Preparative separation of structural isomeric pentacyclic triterpenes from Eriobotrya japonica (Thunb.) leaves by high speed countercurrent chromatography with hydroxypropyl-ß-cyclodextrin as additive.

Maslinic acid and corosolic acid with high purity were successfully separated from Eriobotrya japonica (Thunb.) leaves by two-step countercurrent chromatographic separation. Two biphasic solvent systems composed of petroleum ether-ethyl acetate-ethanol-water (6:4:5:5, v/v) and petroleum ether-ethyl acetate-ethanol-0.10 mol/L of hydroxypropyl-ß-cyclodextrin with pH 7.0 (8:2:3.5:6.5, v/v) were selected according to the partition performance of the main structural isomeric pentacyclic triterpenes. The influences of pH value and concentration of hydroxypropyl-ß-cyclodextrin in separation of two isomers were investigated. In first step countercurrent chromatographic separation, a mixture of two target structural isomers (14.12 mg of sample I) was separated from 40.00 mg of a partially purified sample. In second step countercurrent chromatographic separation, maslinic acid and corosolic acid were completely isolated from 12.00 mg of sample I with hydroxypropyl-ß-cyclodextrin as aqueous phase additive. The recoveries of the two isomers were over 90%, yielding 5.18 mg of maslinic acid and 5.47 mg of corosolic acid, respectively.

Enantioseparation of 2-(4-chlorophenyl)succinic acid by countercurrent chromatography and investigation of injection volume on resolution.

2-(4-Chlorophenyl)succinic acid was successfully enantioseparated by countercurrent chromatography using hydroxypropyl-ß-cyclodextrin as chiral selector. A two-phase solvent system composed of n-hexane-ethyl acetate-0.1 mol/L phosphate buffer with pH 2.65 (5:5:10, v/v) was selected. Enantioselective liquid-liquid extraction was used to optimize the enantioseparation conditions. Meanwhile, the influence of injection volume on resolution in countercurrent chromatography was investigated and a linear relationship between the inflection point of injection volume and sample loading was tentatively obtained. The peak resolution will decrease significantly when the injection volume over the inflection point was used. In addition, it could be found that the smaller amount of sample loading, the larger impact of injection volume on resolution could be observed, which might serve as a good reference for the selection of sample volume in enantioseparations by countercurrent chromatography. Under optimized conditions, 20 mg of 2-(4-chlorophenyl)succinic acid racemate dissolved in 10 mL of aqueous phase was successfully enantioseparated by countercurrent chromatography. The recovery for both of the enantiomer of (±)-2-(4-chlorophenyl)succinic acid reached within 70-75% with a purity of 99.0%.

Enantioseparation of three constitutional isomeric 2-(methylphenyl)propanoic acids by countercurrent chromatography.

2-(3-Methylphenyl)propanic acid and 2-(4-methylphenyl)propanoic acid were successfully enantioseparated by countercurrent chromatography using hydroxypropyl-ß-cyclodextrin (HP-ß-CD) as a chiral selector. 2-(2-Methylphenyl)propanoic acid was also studied to compare the enantioseparation ability of three isomeric 2-(methylphenyl)propanoic acids. Totally 20 mg of 2-(3-methylphenyl)propanic acid and 20mg of 2-(4-methylphenyl)propanic acid were enantioseparated individually by countercurrent chromatography. Recovery for the (±)-2-(3-methylphenyl)propanic acid enantiomer was in the range of 85%-90% with 98.0%-98.8% purity and recovery for the (±)-2-(4-methylphenyl)propanic acid enantiomer was in the range of 80%-83% with 97.0%-98.0% purity. The enantioseparation factor in countercurrent chromatography for 2-(4-methylphenyl)propanic acid and 2-(3-methylphenyl)propanic acid were 1.31 and 1.26, and the peak resolution in HPLC reached 2.2 and 1.4. However, no enantioseparation could be found for 2-(2-methylphenyl)propanic acid. In addition, the inclusion complexes were investigated by UV spectrophotometer. The inclusion formation constant of inclusion complex between 2-(4-methylphenyl)propanic acid, 2-(3-methylphenyl)propanic acid, 2-(2-methylphenyl)propanic acid and HP-ß-CD were determined as 121.73 mol/L, 78.12 mol/L and 53.18 mol/L, respectively. The present results showed that enantiorecognition was greatly affected by substituted positions of methyl group on the benzene ring. Combined with our previous results, the steric hindrance had a significant effect on inclusion interaction between HP-ß-CD and racemic 2-(substitutedphenyl)propanoic acids. No enantiorecognition could be achieved for 2-(substitutedphenyl)propanoic acids with ortho-substituent group on benzene, while the influence of meta- and para- group on enantiorecognition varies with different substituent groups on benzene ring.

Enantioseparation of ondansetron by countercurrent chromatography using sulfobutyl ether-ß-cyclodextrin as chiral selector.

Ondansetron, a highly selective 5-hydroxytryptamine 3 receptor antagonist, was successfully enantioseparated by recycling countercurrent chromatography using sulfobutyl ether-ß-cyclodextrin as chiral selector. Important factors for the enantioseparation were optimized, including different organic solvent, type of substituted ß-cyclodextrin, pH of aqueous phase, concentration of chiral selector, and separation temperature. A biphasic solvent system composed of n-hexane: n-butyl acetate: 0.1 mol/L phosphate buffer solution pH 9.2 with 50 mmol/L of sulfobutyl ether-ß-cyclodextrin (2.5:7.5:10, v/v/v) was selected. Under optimized separation conditions, 5 mg of ondansetron was enantioseparated using recycling countercurrent chromatography, yielding 1.2 and 1.5 mg of ondansetron enantiomers with 97.5 and 95.8% purity and the recovery reached 48-60%.