Abstract
Cyclic dinucleotides (CDNs) act as intracellular messengers and modulate many cellular activities including innate immune activation, virulence factor production and biofilm formation in bacterial pathogens. The cellular levels of CDNs are fine-tuned by cyclases (involved in their synthesis) and phosphodiesterases (PDEs) (involved in their degradation). HD-GYP proteins represent a relative novel addition to the PDE functional superfamily, and belong to a large superfamily of proteins annotated as HD-domain, designated by the tandem histidine-aspartate as a core structural motif. In this study, we aimed to determine the metal cofactor specificity of HD-GYP PDEs and catalytic proficiencies regarding different substrates and their abilities to perform one-step (to 5’-pNpN) or two-step hydrolysis (to two molecules of NMP). We initially focused on the HD-GYP protein from Vibrio cholera (VCA0681), for which no three-dimensional structure has been resolved and which is the key PDE for regulating canonical c-di-GMP and the hybrid c-GAMP. VCA0681 exhibits a somewhat stringent Fe-dependent activity and contains binding motifs for the ligation of two diiron cofactors, with seven conserved protein residues typified as HD-[HD-GYP] (the first dimetal domain lacks the GYP triad, suggested to be important for c-di-GMP hydrolysis). It was the first PDE reported to degrade the novel cyclic dinucleotide c-GAMP, associated with innate immune response. To better understand the substrate and metal specificity of this and other homologous HD-GYP PDEs, apart from exhaustive biochemical and biophysical studies, the three-dimensional structure is much needed. Though VCA0681 is considered as a prototypical PDE, it can only be isolated as a linear fusion with the maltose binding protein (MBP), removal of which leads to protein instability and precipitation, precluding any biophysical and structural investigations on the native protein. We thus carried out a bioinformatics analysis to identify a structurally and functionally homologous HD-[HD-GYP] domain protein. SO3491 from the model -proteobacterium Shewanella oneidensis was selected as the best candidate both due to its sequence similarity (60%) and identity (43%) to VCA0681 as well as its predictability for crystallization. I characterized the kinetic and specificity profile for SO3491 including its metal dependence and substrate specificity and demonstrated that it is a true structural and functional homolog of VCA0681. SO3491 can also hydrolyze the hybrid substrate c-GAMP. Our studies redefined the role of the second metal-binding domain, which was previously unknown. We confirmed that the C-terminal HD-GYP domain is pivotal for PDE activity. Intriguingly, the N-terminal HD-domain is essential for stopping cyclic dinucleotide degradation after the first step hydrolysis, especially in the case of the hybrid c-GAMP. To gain better insight into the HD-GYP protein superfamily, we performed a phylogenetic analysis to identify whether we could establish a correlation between metal content, substrate specificity and mode of action for these HD-GYPs, with the ultimate aim to delineate their multiple occurrences and specific roles within the cell. This analysis reveals a biochemical and structural diversion of these proteins and provides a good predictive map for assigning metal-dependence, level of activity and reaction outcome in different HD-GYPs, allowing for constructing an evolutional diversification profile.