J

J. prediction contradicts the current models, suggesting that both animal and fungal proteins share one topology. This newly predicted topology of Mfn1 and Mfn2 was exhibited biochemically, confirming that this C-terminal, redox-sensitive cysteine residues reside within the intermembrane space (IMS). Functional experiments established that redox-mediated disulfide modifications within the IMS domain name are key modulators of reversible Mfn oligomerization that drives fusion. Together, these results lead to a revised understanding of Mfns as single-spanning outer membrane proteins with an NoutCCin orientation, providing functional insight into the IMS contribution to redox-regulated fusion events. Introduction In animals, mitochondrial Flopropione fusion is usually activated during cellular Flopropione stress and starvation, events that render the cells refractive to apoptosis and protect the mitochondrial reticulum from autophagy (Shutt and McBride, 2013). However, the precise molecular mechanisms that drive mitochondrial fusion have not been established. Although mitochondrial fusion has been observed in diverse eukaryotes, the machinery driving fusion has been identified only in opisthokonts, the lineage comprising animals and fungi. Two dynamin domainCcontaining proteins are required for mitochondrial fusion in mammals: Mfn1/Mfn2 in the outer mitochondrial membrane and Opa1 in the inner mitochondrial membrane (Labb et al., 2014). Even though orthology of fungal Fzo1 to Mfns and fungal Mgm1 to Opa1 has been assumed based on domain name business and similarity of function, sequence-based orthology detection methods fail to retrieve these orthologue units (Mu?oz-Gmez et al., 2015; Purkanti and Thattai, 2015). Loss of any of these genes prospects to mitochondrial fragmentation, and mouse models lacking Mfn1, Mfn2, or Opa1 are embryonic lethal (Chen et al., 2003; Davies et al., 2007; Pareyson et al., 2015). Cell-free fusion assays developed in both and mammalian systems have identified core principles of mitochondrial fusion, namely that the process depends on GTP hydrolysis, that these GTPases are each essential for fusion, and that cytosolic factors can promote fusion (Meeusen et al., 2004, 2006; Schauss et al., 2010; Hoppins et al., 2011; Shutt et al., 2012; Mishra et al., 2014). In the mammalian system, mitochondrial fusion is usually activated upon cellular stress, where oxidized glutathione disulfide (GSSG) promotes the assembly of higher-order Mfn complexes mediated by reversible disulfide bonds (Shutt et al., 2012). The cysteine residues responsible for these dynamic oligomers were located within the C-terminal domain name, assumed to be a cytosol-exposed region. Oligomerization occurred in cis, before mitochondrial docking, suggesting that the generation of disulfide-induced Mfn2 oligomers may take action to prime them to bind in trans and Akt1 drive mitochondrial fusion (Ryan and Stojanovski, 2012; Shutt et al., 2012). A confounding element of this previous work is that the cytosol is generally considered a reducing environment, making it hard to envision how GSSG-induced redox switching may occur within a cytosolic domain name. In this study, we revisit the topology of Mfn1 and Mfn2 by validating bioinformatic predictions with biochemical and functional experiments. The data lead to a revised model of holozoan mitofusins as NoutCCin proteins, consisting of a single membrane-spanning domain name (transmembrane domain name [TMD]) with the 110 residue made up of heptad repeat 2 (HR2) and metazoan-specific disulfide-modifiable cysteine residues residing within the intermembrane space (IMS). Results and conversation Holozoan Mfns and fungal Fzo1 are divergent dynamin domainCcontaining proteins with distinct domain name architecture To better map the conserved functional domains within Mfns and Fzo1, we collected homologous protein sequences from diverse animal and fungal genomes (Table S1). Surprisingly, initial searches using Basic Local Alignment Search Tool (BLAST) suggested that certain bacterial dynamin-like proteins (BDLPs) might be more closely related to Mfn or Fzo than Mfn and Fzo are to each other or to other eukaryotic dynamin-like proteins. Phylogenetic reconstruction of the GTPase domains from Fzos, Mfns, and BDLPs provided no strong support for the relationship of Mfn or Fzo Flopropione to particular bacterial sequences or to one another (Fig. 1 A and Fig. S1, A and B). These data suggest either that (a) Mfns are as related to Fzos as either are to BDLPs (e.g., Mfn and Fzo arose from individual horizontal gene transfers), or (b) more likely, in the time since the divergence of animals and fungi, the sequences of Mfn and Fzo have diverged from a common ancestral protein so much that this evolutionary history of these proteins cannot be properly assessed by current phylogenetic methods. Open in a separate window Physique 1. Bioinformatic analysis of Fzo and Mfn. (A) Phylogenetic reconstruction of dynamin domainCcontaining proteins from bacteria, fungi,.