Natural product structure and fragment-based compound development inspire pseudo-natural product design through different combinations of a given natural product fragment set to compound classes expected to be chemically and biologically diverse. We describe the synthetic combination of the fragment-sized natural products quinine, quinidine, sinomenine, and griseofulvin with chromanone or indole-containing fragments to provide a 244-member pseudo-natural product collection. Cheminformatic analyses reveal that the resulting eight pseudo-natural product classes are chemically diverse and share both drug- and natural product-like properties. Unbiased biological evaluation by cell painting demonstrates that bioactivity of pseudo-natural products, guiding natural products, and fragments differ and that combination of different fragments dominates establishment of unique bioactivity. Identification of phenotypic fragment dominance enables design of compound classes with correctly predicted bioactivity. The results demonstrate that fusion of natural product fragments in different combinations and arrangements can provide chemically and biologically diverse pseudo-natural product classes for wider exploration of biologically relevant chemical space.

No abstract

Unbiased morphological profiling of bioactivity, for example, in the cell painting assay (CPA), enables the identification of a small molecule's mode of action based on its similarity to the bioactivity of reference compounds, irrespective of the biological target or chemical similarity. This is particularly important for small molecules with nonprotein targets as these are rather difficult to identify with widely employed target-identification methods. We employed morphological profiling using the CPA to identify compounds that are biosimilar to the iron chelator deferoxamine. Structurally different compounds with different annotated cellular targets provoked a shared physiological response, thereby defining a cluster based on their morphological fingerprints. This cluster is based on a shared mode of action and not on a shared target, that is, cell-cycle modulation in the S or G2 phase. Hierarchical clustering of morphological fingerprints revealed subclusters that are based on the mechanism of action and could be used to predict target-related bioactivity.

We have evaluated the commercially available Burgess catalyst in hydrogen isotope exchange reactions with several substrates bearing different directing group functionalities and have obtained moderate to high (50%-97%D) deuterium incorporations. The broad applicability in hydrogen isotope exchange reactions makes the Burgess catalyst a possible alternative compared to other commercially available iridium(I)-catalysts.