Cyclodipeptide synthases (CDPS) divert aminoacyl-tRNAs to produce cyclodipeptides and complex derivatives, such as diketopiperazines, which constitute a broad class of natural products synthesized by microorganisms and possessing pharmacological properties. Most CDPS have a relaxed specificity and often produce several cyclopeptides. This hinders the identification of specificity determinants. To overcome this problem, we selected Nbra-CDPS, the CDPS enzyme from Nocardia Brasiliensis. Nbra-CDPS has the advantage of using different substrates, Ala and Glu tRNA that target the first and second CDPS pockets respectively, thus, synthesizing cyclo-Ala-Glu (cAE) as the main product. A recent study of Gondry's team (Canu et al., 2020) shows that CDPS interact mainly with the acceptor arms of tRNAs (also termed miHx). A major objective of our study is to identify the CDPS amino acid residues responsible for substrate specificity. The substitution of these residues will allow to generate enzymes that can use non canonical amino acids. Ultimately, the intent of the project is to produce by an ecological biosynthesis process various diketopiperazines with high therapeutic potential. Firstly, we performed biophysical interaction measurements to evaluate and improve the affinity of Nbra-CDPS for several miHx by BioLayer Interferometry (BLI) and electrophoretic mobility shift assay (EMSA). The optimized conditions allowed us to stabilize the complex with a KD for non acylated miHxAla in the nanomolar range by magnesium removal and reduction of salt concentration to 100 mM. Furthermore, biophysical studies provided guidance to identify conditions that enhance the complexed state. Thereby, we have solved the structure of the complex between Nbra-CDPS and a non-acetylated miHxAla at 3.4 Å by X-ray crystallography and we identified some obstacles that may hinder the study of the complex with the acylated form of the substrate. In particular, the substrate Ala-miHxAla is not stable and deacylates significantly. Thus, we recently implemented the use of more stable analogues of the acylated substrate through amide bond formation (Ala-miHx-NHAla) for structural and biophysical studies. Indeed, we initiated a collaboration with H. Suga's team who documented the synthesis of miHx-NH (Katoh and Suga, 2019). The aminoacylation of these substrates with flexizyme has been set up to provide stable Ala-miHx-NHAla. The crystallization screening using these stable derivatives is in progress. In addition, NMR studies were performed on isotopically labeled Nbra-CDPS and highlighted its interaction with a non-acylated miHxAla substrate by 15N HSQC spectra. The preliminary NMR assignment provided first molecular information of residues involved in the recognition with miHx. We also intend to study the structure of the miHxAla by NMR to identify the effect of the interaction of Nbra-CDPS/miHxAla on the RNA acceptor arm folding. In conclusion, the combination of all these structural and biophysical approaches will allow us to decipher the molecular mechanism of recognition between CDPS and their substrates and thus guide the engineering of these enzymes.