Search

Summary The structural modeling of protein interactions in the absence of close homologous templates is a challenging task. Recently, template-based docking methods have emerged to exploit local structural similarities to help ab-initio protocols provide reliable 3D models for protein interactions. In this work, we critically assess the performance of template-based docking in the twilight zone. Our results show that, while it is possible to find templates for nearly all known interactions, the quality of the obtained models is rather limited. We can increase the precision of the models at expenses of coverage, but it drastically reduces the potential applicability of the method, as illustrated by the whole-interactome modeling of nine organisms. Template-based docking is likely to play an important role in the structural characterization of the interaction space, but we still need to improve the repertoire of structural templates onto which we can reliably model protein complexes.

Azole antifungals, known as demethylase inhibitors (DMIs), target sterol 14α-demethylase (CYP51) in the ergosterol biosynthetic pathway of fungal pathogens of both plants and humans. DMIs remain the treatment of choice in crop protection against a wide range of fungal phytopathogens that have the potential to reduce crop yields and threaten food security. We used a yeast membrane protein expression system to overexpress recombinant hexahistidine-tagged S. cerevisiae lanosterol 14α-demethylase and the Y140F or Y140H mutants of this enzyme as surrogates in order characterize interactions with DMIs. The whole-cell antifungal activity (MIC50 values) of both the R- and S-enantiomers of tebuconazole, prothioconazole (PTZ), prothioconazole-desthio, and oxo-prothioconazole (oxo-PTZ) as well as for fluquinconazole, prochloraz and a racemic mixture of difenoconazole were determined. In vitro binding studies with the affinity purified enzyme were used to show tight type II binding to the yeast enzyme for all compounds tested except PTZ and oxo-PTZ. High resolution X-ray crystal structures of ScErg11p6×His in complex with seven DMIs, including four enantiomers, reveal triazole-mediated coordination of all compounds and the specific orientation of compounds within the relatively hydrophobic binding site. Comparison with CYP51 structures from fungal pathogens including Candida albicans, Candida glabrata and Aspergillus fumigatus provides strong evidence for a highly conserved CYP51 structure including the drug binding site. The structures obtained using S. cerevisiae lanosterol 14α-demethylase in complex with these agrochemicals provide the basis for understanding the impact of mutations on azole susceptibility and a platform for the structure-directed design of the next-generation of DMIs.

Although detailed, focused, and mechanistic analyses of associations among mitochondrial proteins (MPs) have identified their importance in varied biological processes, a systematic understanding of how MPs function in concert both with one another and with extra-mitochondrial proteins remains incomplete. Consequently, many questions regarding the role of mitochondrial dysfunction in the development of human disease remain unanswered. To address this, we compiled all existing mitochondrial physical interaction data for over 1200 experimentally defined yeast MPs and, through bioinformatic analysis, identified hundreds of heteromeric MP complexes having extensive associations both within and outside the mitochondria. We provide support for these complexes through structure prediction analysis, morphological comparisons of deletion strains, and protein co-immunoprecipitation. The integration of these MP complexes with reported genetic interaction data reveals substantial crosstalk between MPs and non-MPs and identifies novel factors in endoplasmic reticulum-mitochondrial organization, membrane structure, and mitochondrial lipid homeostasis. More than one-third of these MP complexes are conserved in humans, with many containing members linked to clinical pathologies, enabling us to identify genes with putative disease function through guilt-by-association. Although still remaining incomplete, existing mitochondrial interaction data suggests that the relevant molecular machinery is modular, yet highly integrated with non-mitochondrial processes.