Incorporation of natural and non-natural amino acids in human cells for advanced in-cell NMR labelling schemes

This thesis presents significant advancements in developing selective isotopic labelling strategies for in-cell nuclear magnetic resonance (NMR) spectroscopy within human cells, aiming to overcome limitations related to cost, sensitivity, specificity, applicability to complex biological systems and physiological models. We first established a cost-effective and innovative methodology for selective side-chain isotope labelling in mammalian cells. This method exploits the activity of human cellular transaminases to convert amino acid late precursors (α-keto acids) into their cognate L-amino acids, achieving efficient and residue-specific incorporation of 13C labelled aromatic and aliphatic residues with minimal scrambling. High-quality in-cell and lysate NMR spectra were obtained, and the functional utility was validated by monitoring ligand-induced chemical shift perturbations in live cells expressing carbonic anhydrase II (CA II) treated with different inhibitors. We also showed the incorporation of a non-natural amino acid, 5-fluorotryptophan selectively labelled with 13C at the C5 position, in proteins expressed in human cells (CA II and superoxide dismutase 1). This strategy allows recording 2D 13C, 19F NMR spectra in which optimal peak dispersion is achieved by exploiting the TROSY effect. Furthermore, the thesis focused on designing selective labelling schemes for intrinsically disordered regions (IDRs) of proteins, applied to the challenging case of BRCA1. The expression of four fragments was achieved in human cells, including a very long fragment (219-1357) spanning almost the entire IDR of BRCA1. Optimized isotopic labelling schemes were then implemented, utilizing either 13C, 15N uniformly labelled amino acids (L-histidine and L-cysteine) or 13Cα or 13C’ labelled α-ketoacid precursors (which are converted to L-leucine and L-phenylalanine), enabling residue-specific observation and greatly reducing spectral crowding in the in-cell and lysate NMR spectra of BRCA1. To facilitate future interaction studies with physiological partners, a co-expression protocol was established, and a monoclonal inducible BRCA1 cell line was developed, enabling the isotopic labelling of only one interacting protein to simplify spectral interpretation. Finally, the in-cell NMR approach was extended to investigate ligand-target interactions under increasing cell culture complexity. By utilizing a flow NMR bioreactor and time-resolved 19F NMR, the displacement of a fluorinated CA II inhibitor by competition with methazolamide as a reference compound was monitored across four different cellular models: transiently transfected HEK293T cells, stably transfected inducible monoclonal cells expressing CA II (HEK293-TRexTR-hCAII/Mono), HEK293-TRexTR-hCAII/Mono spheroids, and mixed spheroids (HEK293-TRexTR-hCAII/Mono mixed with MCF-7 cells). These experiments aimed to detect differences in relative binding affinity amongst the different models.