Protein engineering using directed evolution and (semi)rational design has emerged as a powerful strategy for optimizing and enhancing enzymes or proteins with desired properties. Integrating artificial intelligence methods has further enhanced and accelerated protein engineering through predictive models developed in data-driven strategies. However, the lack of explainability and interpretability in these models poses challenges. Explainable Artificial Intelligence addresses the interpretability and explainability of machine learning models, providing transparency and insights into predictive processes. Nonetheless, there is a growing need to incorporate explainable techniques in predicting protein properties in machine learning-assisted protein engineering. This work explores incorporating explainable artificial intelligence in predicting protein properties, emphasizing its role in trustworthiness and interpretability. It assesses different machine learning approaches, introduces diverse explainable methodologies, and proposes strategies for seamless integration, improving trust-worthiness. Practical cases demonstrate the explainable model’s effectiveness in identifying DNA binding proteins and optimizing Green Fluorescent Protein brightness. The study highlights the utility of explainable artificial intelligence in advancing computationally assisted protein design, fostering confidence in model reliability.

A bottleneck in the development of new anti-cancer drugs is the recognition of their mode of action (MoA). We combined metabolomics and machine learning to predict MoAs of novel anti-proliferative drug candidates, focusing on human prostate cancer cells (PC-3). As proof of concept, we studied 38 drugs with known effects on 16 key processes of cancer metabolism, profiling low molecular weight intermediates of the central carbon and cellular energy metabolism (CCEM) by LC-MS/MS. These metabolic patterns unveiled distinct MoAs, enabling accurate MoA predictions for novel agents by machine learning. We validate the transferability of MoA predictions from PC-3 to two other cancer cell models and show that correct predictions are still possible, but at the expense of prediction quality. Furthermore, metabolic profiles of treated cells yield insights into intracellular processes, exemplified for drugs inducing different types of mitochondrial dysfunction. Specifically, we predict that pentacyclic triterpenes inhibit oxidative phosphorylation and affect phospholipid biosynthesis, as supported by respiration parameters, lipidomics, and molecular docking. Using biochemical insights from individual drug treatments, our approach offers new opportunities, including the optimization of combinatorial drug applications.

Protein engineering using directed evolution and (semi)rational design has emerged as a powerful strategy for optimizing and enhancing enzymes or proteins with desired properties. Integrating artificial intelligence methods has further enhanced and accelerated protein engineering through predictive models developed in data-driven strategies. However, the lack of explainability and interpretability in these models poses challenges. Explainable Artificial Intelligence addresses the interpretability and explainability of machine learning models, providing transparency and insights into predictive processes. Nonetheless, there is a growing need to incorporate explainable techniques in predicting protein properties in machine learning-assisted protein engineering. This work explores incorporating explainable artificial intelligence in predicting protein properties, emphasizing its role in trustworthiness and interpretability. It assesses different machine learning approaches, introduces diverse explainable methodologies, and proposes strategies for seamless integration, improving trust-worthiness. Practical cases demonstrate the explainable model’s effectiveness in identifying DNA binding proteins and optimizing Green Fluorescent Protein brightness. The study highlights the utility of explainable artificial intelligence in advancing computationally assisted protein design, fostering confidence in model reliability.

A bottleneck in the development of new anti-cancer drugs is the recognition of their mode of action (MoA). We combined metabolomics and machine learning to predict MoAs of novel anti-proliferative drug candidates, focusing on human prostate cancer cells (PC-3). As proof of concept, we studied 38 drugs with known effects on 16 key processes of cancer metabolism, profiling low molecular weight intermediates of the central carbon and cellular energy metabolism (CCEM) by LC-MS/MS. These metabolic patterns unveiled distinct MoAs, enabling accurate MoA predictions for novel agents by machine learning. We validate the transferability of MoA predictions from PC-3 to two other cancer cell models and show that correct predictions are still possible, but at the expense of prediction quality. Furthermore, metabolic profiles of treated cells yield insights into intracellular processes, exemplified for drugs inducing different types of mitochondrial dysfunction. Specifically, we predict that pentacyclic triterpenes inhibit oxidative phosphorylation and affect phospholipid biosynthesis, as supported by respiration parameters, lipidomics, and molecular docking. Using biochemical insights from individual drug treatments, our approach offers new opportunities, including the optimization of combinatorial drug applications.

The new farnesyl pyrophosphate (FPP) derivative with a shifted olefinic double bond from C6‐C7 to C7‐C8 is accepted and converted by the sesquiterpene cyclases protoilludene synthase (Omp7) as well as viridiflorene synthase (Tps32). In both cases, a so far unknown germacrene derivative was found to be formed, which we name “germacrene F”. Both cases are examples in which a modification around the central olefinic double bond in FPP leads to a change in the mode of initial cyclization (from 1→11 to 1→10). For Omp7 a rationale for this behaviour was found by carrying out molecular docking studies. Temperature‐dependent NMR experiments, accompanied by NOE studies, show that germacrene F adopts a preferred mirror‐symmetric conformation with both methyl groups oriented in the same directions in the cyclodecane ring.

A bottleneck in the development of new anti‐cancer drugs is the recognition of their mode of action (MoA). Metabolomics combined with machine learning allowed to predict MoAs of novel anti‐proliferative drug candidates, focusing on human prostate cancer cells (PC‐3). As proof of concept, 38 drugs are studied with known effects on 16 key processes of cancer metabolism, profiling low molecular weight intermediates of the central carbon and cellular energy metabolism (CCEM) by LC‐MS/MS. These metabolic patterns unveiled distinct MoAs, enabling accurate MoA predictions for novel agents by machine learning. The transferability of MoA predictions based on PC‐3 cell treatments is validated with two other cancer cell models, i.e., breast cancer and Ewing\'s sarcoma, and show that correct MoA predictions for alternative cancer cells are possible, but still at some expense of prediction quality. Furthermore, metabolic profiles of treated cells yield insights into intracellular processes, exemplified for drugs inducing different types of mitochondrial dysfunction. Specifically, it is predicted that pentacyclic triterpenes inhibit oxidative phosphorylation and affect phospholipid biosynthesis, as confirmed by respiration parameters, lipidomics, and molecular docking. Using biochemical insights from individual drug treatments, this approach offers new opportunities, including the optimization of combinatorial drug applications.

Histamine is a biogenic amine that can cause food poisoning in an increasing fraction of the population. Histamine detection and quantification are crucial for evaluating the freshness of food products and informing histamine-sensitive consumers regarding histamine concentration in fermented or processed food products. Several analytical methods exist for quantifying histamine from food samples, most based on chromatographic analysis. This review summarizes the current knowledge of analytical methodologies for detecting and quantifying histamine. We highlight the importance of using timely detection tools for biogenic amines to indicate the degree of freshness or deterioration of food. A multidisciplinary approach based on molecular and enzymatic methods for detecting and quantifying histamine and other biogenic amines is presented, where histamine dehydrogenase and histamine oxidase enzymes from microbial sources stand out as potential molecular tools for histamine detection, and with which rapid, scalable, and user-friendly assay platforms can be used. In addition to typical enzyme technology concerns, the enzymatic detection of histamine faces substrate specificity and substrate inhibition challenges that affect the specific identification of histamine and the detection limit of the enzymatic assay. These challenges can be overcome by enzyme engineering, immobilization, or their simultaneous integration to obtain biocatalysts with increased histamine detection, quantification, or performance.

Presilphiperfolan-8β-ol synthase (BcBOT2), a substrate-promiscuous sesquiterpene cyclase (STC) of fungal origin, is capable of converting two new farnesyl pyrophosphate (FPP) derivatives modified at C7 of farnesyl pyrophosphate (FPP) bearing either a hydroxymethyl group or a methoxymethyl group. These substrates were chosen based on a computationally generated model. Biotransformations yielded five new oxygenated terpenoids. Remarkably, the formation of one of these tricyclic products can only be explained by a cationically induced migration of the methoxy group, presumably via a Meerwein-salt intermediate, unprecedented in synthetic chemistry and biosynthesis. The results show the great principle and general potential of terpene cyclases for mechanistic studies of unusual cation chemistry and for the creation of new terpene skeletons.

The development of potent adjuvants is an important step for improving the performance of subunit vaccines. CD1d agonists, such as the prototypical α‐galactosyl ceramide (α‐GalCer), are of special interest due to their ability to activate iNKT cells and trigger rapid dendritic cell maturation and B‐cell activation. Herein, we introduce a novel derivatization hotspot at the α‐GalCer skeleton, namely the N‐substituent at the amide bond. The multicomponent diversification of this previously unexplored glycolipid chemotype space permitted the introduction of a variety of extra functionalities that can either potentiate the adjuvant properties or serve as handles for further conjugation to antigens toward the development of self‐adjuvanting vaccines. This strategy led to the discovery of compounds eliciting enhanced antigen‐specific T cell stimulation and a higher antibody response when delivered by either the parenteral or the mucosal route, as compared to a known potent CD1d agonist. Notably, various functionalized α‐GalCer analogues showed a more potent adjuvant effect after intranasal immunization than a PEGylated α‐GalCer analogue previously optimized for this purpose. Ultimately, this work could open multiple avenues of opportunity for the use of mucosal vaccines against microbial infections.

The new species Cortinarius steglichii is described from Chilean Nothofagus forest based on morphological and microscopical attributes, molecular phylogeny, and chemical analysis of secondary metabolites. C. steglichii is characterized by abundant, long, ramified cystidia on the lamellar edges and stipe apex, further by a deep violet color reaction after treatment with KOH. As responsible secondary metabolite for the cytoplasmatic color reaction of cystidia and some hyphae, the new diterpenoid steglichon (1) could be recognized, showing also remarkable antibacterial and anticancer activity. Phylogenetic analyses (ITS, LSU, RPB1) confirm the close relationship to species of the Cortinarius dulciolens group.