Plants and their rich biodiversity have been used by humans to treat diseases since the dawn of time. Synthetic chemists still struggle to reproduce the complexity of certain chemical reactions that nature has developed throughout millions of years of evolution, which makes the world of flora a great source of inspiration for new bioactive chemicals to be discovered. Yet one of the challenges in Natural Products research is to disentangle the hundreds of molecules contained within each plant extract to highlight those of biological interest. 

 

Additionally, Natural products have a proven track record in the development of anti-bacterials, while antibiotic resistances are on the rise. This is particularly problematic in the case of the world’s deadliest infectious disease: tuberculosis. This disease is caused by mycobacteria, which are notoriously difficult to treat with only few antibiotics being at hand. 

 

In a collaborative effort with the company Pierre Fabre, our team was able to take advantage of their historical library of 18 000 plant samples that could be used as a source of inspiration, for the search of new anti-mycobacterial molecules. In the frame of our study, we proposed an original approach in which 1600 extracts representing approximately 10% of the entire library of plant extracts were analysed with advanced Mass Spectrometry (MS). This method allowed for the identification/detection of hundreds of chemical constituents in each plant extract. 

 

Among these chemicals, for those that have been previously reported in the literature, we could then associate a possible structure also referred to as an “annotation”. 

At the end of this process stood an atlas with over 37 000 predicted structures, that revealed which molecules were likely to be found in which plant. We effectively transformed the collection of extracts into a virtual chemical library of natural compounds. 

 

One way to explore this atlas was to look for natural analogs of commercial molecules previously screened in our advanced biological model. This biological model is in fact an anti-mycobacterial assay carried out using a mycobacterium, hosted in an amoeba, which mimics the behavior of some of our immune cells in the human body when attacking mycobacteria. 

 

In fact, one active molecule from this previous screen (“molecule D”) was targeted using a computational tool called “DataWarrior” that allowed us to search for analog molecules by structural similarity. Within all 37 000 annotations, we found 4 structural analogs of molecule D that were detected in a single plant: Cananga brandisiana. This plant was selected and isolation of said molecules from it was carried out to see if our predicted annotations made from MS data would be correct. Indeed, they were, and these natural analogs were then isolated and subsequently tested against the disease model that was used to identify the previously mentioned template molecule. 

 

Eventually, we managed to isolate a total of 8 analog molecules, of which one in particular (Onychine) showed a promising activity profile when compared to the initial molecule. While in absolute terms its anti-mycobacterial activity was less strong than that of molecule D, it did show a reduced toxicity towards the cells hosting the bacteria in the experiment (amoeba). This information will help us to make links between activity and molecular structure for this type of molecules to improve their bioactivity. 

 

In conclusion, we successfully demonstrated a novel approach for the targeted isolation of bioactive natural products from a diverse collection of plant extracts. Establishing a chemical atlas for our plant library allowed us to avoid cumbersome isolation procedures and instead to efficiently target the molecules of interest. This also provided a large database of over 37 000 predicted molecules that can be explored in different contexts. The combination of computational analysis, mass spectrometry profiling, and biological testing allowed for the isolation of promising molecules with potential for combatting antibiotic resistance in diseases like tuberculosis.