S T R U C T U R A L B I O L O G Y
S C I E N T I F I C H I G H L I G H T S
4 6 H I G H L I G H T S 2 0 2 2 I
Over the past decade, the bacterial enzymes have been studied with the aim to develop new antibiotics. An original approach led to promising NADK inhibitors  that have recently been optimised to submicromolar compounds . Since then, biological characterisation of eukaryotic NADKs is gaining momentum, with recent work showing their involvement in cancer development and especially metastasis for the cytoplasmic enzyme , while the mitochondrial counterpart was recently shown to be essential for proline synthesis to support cancer cell growth and proliferation .
In this context, there was interest in studying the mitochondrial NAD kinase (NADK2) from humans, which appeared highly divergent. Cloning, over-expression and purification using standard procedures rapidly provided a stable sample. Due to its low enzymatic activity, the quality of the protein conformation was questioned. However, small-angle X-ray spectroscopy (SAXS) experiments at beamline BM29 demonstrated a folded protein at various concentrations with no visible aggregates. In agreement, crystals were rapidly obtained and improved to diffracting qualities in a matter of weeks. Data collections recorded at beamline ID30B, at resolutions ranging from 2.25 to 3.33 Å, provided useful datasets. While collecting other datasets remotely, a clear solution of molecular replacement was obtained using a model built using a newly released version of AlphaFold .
Structure refinement revealed an impressively low R.M.S.D. of 0.8 Å (over ~170 common residues), while the best models derived from the closest crystal structures differ by 1.5 to 2.0 Å. Not only the common core was well predicted, but also one of the three original insertions (ranging from 20 to 40 residues) with no equivalent in any known structure (Figure 37a). Interestingly, another 30-long insertion, predicted with low confidence, had no electron density visible in the crystals. As a follow-up study, new crystals were grown in distinct conditions. Surprisingly, another poorly defined insertion became more structured, and its new organisation better matched the AlphaFold model (unpublished results).
Hence, last year, experimental structural biology entered a new area where structure modelling can guide any step, from cloning (domain boundaries) to structure refinement (straightforward molecular replacement), while X-ray crystallography remains unrivalled for high-resolution structures and especially the placement of relevant ligands and water molecules. Here, the new crystal structures validated a model deduced from crystal structures of a bacterial enzyme solved 15 years ago . Indeed, the putative chelation of the catalytic metal by the ligands and not the protein, had never been observed until now (Figure 37b).
Fig. 37: Structure of human mitochondrial NADK2. a) Crystal structures of NAD kinase (blue) superimposed onto its AlphaFold model (red). b) Zoom on the metal bound to the product of the enzymatic reaction.
PRINCIPAL PUBLICATION AND AUTHORS
Crystal structure of human NADK2 reveals a dimeric organization and active site occlusion by lysine acetylation, C. Mary (a), M. Hoseini Soflaee (b), R. Kesavan (b), M. Gelin (a), H. Brown (b), L.G. Zacharias (b), T.P. Mathews (b), A. Lemoff (b), C. Lionne (a), G. Labesse (a), G. Hoxhaj (b), Mol. Cell. 82, 3299-3311 (2022); https:/doi.org/10.1016/j.molcel.2022.06.026 (a) Centre de Biologie Structurale, CNRS, INSERM, Univ. Montpellier (France) (b) Children s Medical Center Research Institute, University of Texas Southwestern Medical Center, Dallas (USA)
 G. Labesse et al., Trends Biochem. Sci. 27, 273-275 (2002).  M. Gelin et al., Structure 20, 1107-1117 (2012).  M. Gelin et al., ACS Infect. Dis. 6, 422-435 (2020).  G. Hoxhaj et al., Science 363, 1088-1092 (2019).  D.H. Tran et al., Nat. Metab. 3, 571-585 (2021).  J. Jumper et al., Nature 596, 583-589 (2021).  G. Poncet-Montange et al., J. Biol. Chem. 282, 33925- 33934 (2007).