The integration of AlphaFold-predicted and crystal structures of human trans-3-hydroxy-L-proline dehydratase reveals a regulatory catalytic mechanism br

Computational methods for protein structure prediction have made significant strides forward, as evi-denced by the last development of the neural network AlphaFold, which outperformed the CASP14 com-petitors by consistently predicting the structure of target proteins. Here we show an integrated structural investigation that combines the AlphaFold and crystal structures of human trans-3-Hydroxy-L-proline dehydratase, an enzyme involved in hydroxyproline catabolism and whose structure had never been reported before, identifying a structural element, absent in the AlphaFold model but present in the crystal structure, that was subsequently proved to be functionally relevant. Although the AlphaFold model lacked information on protein oligomerization, the native dimer was reconstructed using template-based and ab initio computational approaches. Moreover, molecular phasing of the diffraction data using the AlphaFold model resulted in dimer reconstruction and straightforward structure solution. Our work adds to the integration of AlphaFold with experimental structural and functional data for protein analy-sis, crystallographic phasing and structure solution. (C) 2022 The Authors. Published by Elsevier B.V. on behalf of Research Network of Computational and Structural Biotechnology.