We have developed a robust technology platform that is now used in routine for all our natural product drug discovery projects. This platform integrates several advanced approaches, including screening of a pre-formatted extract library in 96- and 384-well format, miniaturized tracking of activity of compounds in extracts through HPLC-based activity profiling, UHPLC-HRMS/MS and molecular networking for advanced annotation of peaks within complex mixtures, structure elucidation using microprobes NMR, and the evaluation of biological activity for pure compounds with both in vitro as well as in vivo methods. This workflow is described in detail below.

Our in-house extract library serves as an invaluable source of natural product structures. It includes extracts from plants collected worldwide through collaborations (mainly Panama) and commercial sources (i.e. herbal drugs from China and Europe). Currently, the library includes more than 2,500 plant extracts derived from 700 plant species across 140 plant families, and is continually expanded by incorporating extracts also from fungi and bacteria. Importantly, all extracts are fully compliant with the Nagoya protocol, which assures shared benefit with the country of origin. Extracts are generated automatically using an Accelerated Solvent Extraction (ASE) system and are stored at a concentration of 10 mg/ml in DMSO. Plates are easily transferable to 96- and 384- assay formats with a pipetting robot to prepare screening plates that can be directly used in our assays.

The extract library is subsequently screened on various biological assays as illustrated by the following ongoing projects:

• Melanoma: In this project, we aim to identify natural inhibitors targeting aberrant MAPK/ERK and/or PI3K/AKT signaling pathways in melanoma. To achieve this, we have developed an innovative high-content screening assay that quantifies ERK and AKT at the single-cell level. This assay was initially used to screen our crude extract library and has now been scaled-up to 384-well format to screen large libraries of pure compounds of both natural and synthetic origins. Since the assay measures downstream protein activity (i.e. ERK and AKT), our current focus is to find the specific targets of the compound hits we have identified in these two screening campaigns.
   
• Forgetting: Memory maintenance and forgetting are fundamental processes in our lives. However, the mechanisms underlaying forgetting are still poorly understood. Recently, a protein called Musashi (MSI) was found to play a crucial role in the process of forgetting in C. elegans, which prompted us to look for natural MSI inhibitors in a collaborative project with Dr. Attila Stetak and Prof. Andreas Papassitiropoulos (division molecular neuroscience, UniBasel). Using a biochemical assay, we identified several highly potent compounds, with activities in the nanomolar range. Selected compounds are currently being tested in vivo in C. elegans. Ongoing studies also aim to further investigate this activity and its mechanisms.
   
• Glioblastoma: Glioblastoma is an aggressive and devastating cancer with very poor long-term prognosis and the limited available treatment options drive our search for natural compounds with activity against glioblastoma cells. We focus especially on mitochondria-targeting agents, or mitocans, to advance potential therapies in a collaboration with the group of Prof. Anne Eckert (UPK, UniBasel).
   
• Amanitin toxicity: Amanita phalloides, also known as death cap, is considered the most poisonous mushroom and responsible for more than 95 % fatalities caused by mushroom poisoning worldwide. Its main toxin is α-amanitin that disrupts protein biosynthesis, ultimately leading to liver and kidney failure. Unfortunately, no antidote exists, and in case of severe poisoning, liver transplant remains the only viable option. Other treatments include the use of silybinin, a natural product derived from Silybum marianum, which appears to compete with a liver transporter (OATP_1B3) that α-amanitin uses to enter the hepatocytes. The objective of this project is to identify natural products binding to the OATP_1B3 transporter, thereby inhibiting the hepatic uptake of amatoxins, as well as to investigate the interaction of α-amanitin with OATP transporters in more details. This project is a collaboration with the group of Prof. Henriette Meyer zu Schwabedissen (division biopharmacy, UniBasel).

The HPLC-based activity profiling was developed to correlate the biological activity of an extract with individual peaks (i.e. components) in its chromatogram. In this approach, the extract is loaded onto an analytical HPLC-DAD-CAD-MS system in two portions. The first is used to generate analytical data (HPLC-traces), while the second is collected into microfractions, which are then dried, redissolved in DMSO, and tested for biological activity. The HPLC-based activity profiles are generated by combining the HPLC traces with the activity values of each fraction (in the R language). This MS and UV data associated with peaks eluting in active windows are used to annotate these by comparison with in-house and online databases. These results are thus used for selecting the most promising samples for scale-up extraction and targeted isolation of the active constituents. Recently, this approach has been combined with UHPLC-HRMS/MS and molecular networking to further improve the annotation of specific peaks and thus streamline the selection process.

Extracts of interest are scaled up for targeted isolation using different chromatography techniques, e.g. open column chromatography, size exclusion chromatography, centrifugal partition chromatography (CPC), as well as preparative HPLC chromatography. The structure of the isolated compounds is then elucidated via microprobe NMR and HRMS analysis. When necessary, absolute configurations are determined using X-ray crystallography, ECD spectroscopy or chemical derivatization (e.g. mosher ester analysis).

The biological activity of all isolated compounds is finally assessed. In most cases, this is done using the assay employed for the primary screening. Ideally, a concentration-dependent curve is established (i.e. IC₅₀ or ED₅₀). When a series of analogues are isolated their activity values can provide a structure-activity relationship (SAR) that can inform further investigations, including testing on additional cell lines, or in in vivo models. In the example mentioned here, a series of thymol derivatives were identified in the melanoma project. Testing all these derivatives revealed that only the compounds with an epoxide displayed an IC₅₀ < 20 μM on the ERK pathway and were considered as active, while all the others had an IC₅₀ > 30 μM. 

Depending on the need of the project, additional investigations are performed including target identification, in vivo testing and/or structure optimization.