Tetrameric Structures of Inorganic CBS-Pyrophosphatases from Various Bacterial Species Revealed by Small-Angle X-ray Scattering in Solution

The Structure and Nucleotide-Binding Characteristics of Regulated Cystathionine β-Synthase Domain-Containing Pyrophosphatase without One Catalytic Domain

Regulatory adenine nucleotide-binding cystathionine β-synthase (CBS) domains are widespread in proteins; however, information on the mechanism of their modulating effects on protein function is scarce. The difficulty in obtaining structural data for such proteins is ascribed to their unusual flexibility and propensity to form higher-order oligomeric structures. In this study, we deleted the most movable domain from the catalytic part of a CBS domain-containing bacterial inorganic pyrophosphatase (CBS-PPase) and characterized the deletion variant both structurally and functionally. The truncated CBS-PPase was inactive but retained the homotetrameric structure of the full-size enzyme and its ability to bind a fluorescent AMP analog (inhibitor) and diadenosine tetraphosphate (activator) with the same or greater affinity. The deletion stabilized the protein structure against thermal unfolding, suggesting that the deleted domain destabilizes the structure in the full-size protein. A "linear" 3D structure with an unusual type of domain swapping predicted for the truncated CBS-PPase by Alphafold2 was confirmed by single-particle electron microscopy. The results suggest a dual role for the CBS domains in CBS-PPase regulation: they allow for enzyme tetramerization, which impedes the motion of one catalytic domain, and bind adenine nucleotides to mitigate or aggravate this effect.

Ab initio determination of the shape of membrane proteins in a nanodisc

New software, called Marbles, is introduced that employs SAXS intensities to predict the shape of membrane proteins embedded into membrane nanodiscs. To gain computational speed and efficient convergence, the strategy is based on a hybrid approach that allows one to account for the contribution of the nanodisc to the SAXS intensity through a semi-analytical model, while the embedded membrane protein is treated as a set of beads, similarly to as in well known ab initio methods. The reliability and flexibility of this approach is proved by benchmarking the code, implemented in C++ with a Python interface, on a toy model and two proteins with very different geometry and size.