Pharmacoinformatic Investigation of Medicinal Plants from East Africa

Abstract

Medicinal plants have widely been used in the traditional treatment of ailments and have been proven effective. Their contribution still holds an important place in modern drug discovery due to their chemical, and biological diversities. However, the poor documentation of traditional medicine, in developing African countries for instance, can lead to the loss of knowledge related to such practices. In this study, we present the Eastern Africa Natural Products Database (EANPDB) containing the structural and bioactivity information of 1870 unique molecules isolated from about 300 source species from the Eastern African region. This represents the largest collection of natural products (NPs) from this geographical region, covering literature data of the period from 1962 to 2019. The computed physicochemical properties and toxicity profiles of each compound have been included. A comparative analysis of some physico-chemical properties like molecular weight, H-bond donor/acceptor, logP_o/w , etc. as well scaffold diversity analysis has been carried out with other published NP databases. EANPDB was combined with the previously published Northern African Natural Products Database (NANPDB), to form a merger African Natural Products Database (ANPDB), containing ∼6500 unique molecules isolated from about 1000 source species (freely available at http://african-compounds.org). As a case study, latrunculins A and B isolated from the sponge Negombata magnifica (Podospongiidae) with previously reported antitumour activities, were identified via substructure searching as molecules to be explored as putative binders of histone deacetylases (HDACs).

Keywords: Eastern Africa; database; drug discovery; medicinal plants; natural products (NPs).

Figure 1

Bar chart showing the percentage contribution by family.

Figure 2

Distribution of main compound classes.

Figure 3

Distribution of reported biological activities within the EANPDB.

Figure 4

Molecules from the EANPDB with reported mechanisms of action (MOA).

Figure 5

Chemical space analysis using the best 3 PCAs (PCA1, PCA2 and PCA3). A) Distribution of molecules in EANPDB, B) Distribution of molecules in NANPDB, C) Distribution of molecules in NuBBE and D) Overlay of all three datasets to appreciate their spatial distribution. Blue, Green and black balls represent EANPDB, NANPDB and NuBBE respectively.

Figure 6

Distribution of A) molecular weight and B) logP (octanol‐water) coefficient for EANPDB, DrugBank, StreptomeDB, NANPDB and NuBBE datasets.

Figure 7

Distribution of A) H‐bond donor and B) H‐bond acceptor for EANPDB, DrugBank, StreptomeDB, NANPDB and NuBBE datasets.

Figure 8

Percentile violation of A) drug‐likeness (Lipinski's rule of five violation) B) lead‐likeness (Jorgensen's Rule of Three violation).

Figure 9

Distribution of scaffold similarity between EANPDB and NANPDB using A) CSR and B) SSE30.

Figure 10

Display of most frequent cyclic scaffolds.

Figure 11

Classical pharmacophore of HDAC inhibitors and the Chemical structure of romidepsin, largazole and the newly proposed latrunculins A and B from the African NP database.

Figure 12

Partial “CS” substructure search output result.