Exploring the subcellular role of human AIFM3

Human AIFM3 is a poorly studied protein which has been described to localize to the inner mitochondrial membrane, contain an oxidoreductase domain which potentially binds to NAD/FAD, and execute apoptosis through a caspase-dependent pathway upon overexpression. Other studies reported upregulation of AIFM3 in several cancer types including cholangiocarcinoma and breast cancer and suggested that AIFM3 could be a prognostic marker in these cancer types due to its elevated presence in the sera of patients. Although, the direct consequence or cause of AIFM3 upregulation in cancer tissues is not known yet, correlational analysis shows a positive trend between higher AIFM3 expression in late stage of cancer and higher risk of metastasis. Therefore, we decided to study the subcellular role of AIFM3 regarding its localization, structure-function relationship, and role in cell physiology. After bioinformatic analysis on the structure of AIFM3, we found that it contains the following domains: Rieske, NADB, PYR, Reductase C. Based on these findings, we designed 13 constructs each carrying a different mutation and targeting a key part of AIFM3 in one of the following categories: localization, splicing variants, protein-substrate interaction, and function. Overexpression of this minilibrary in HEK293T demonstrated an oligomerization pattern above 185kDa. In addition, the construct lacking NADB domain was unable to oligomerize suggesting that NADB domain is necessary for AIFM3 oligomerization. Since NADH is known to bind to the NADB domain and facilitate protein oligomerisation, it is speculated that NADH binding to this domain may strike AIFM3 oligomerisation. Moreover, quantitative PCR showed that AIFM3 is not expressed in HEK293T cells in the basal conditions, therefore suggesting that previous studies on AIFM3 localization require further investigation. We found that commercially available antibodies against AIFM3 detect off-targets both in western blot and immunofluorescence using wild-type or AIFM3 KO cells. Therefore, we generated a plasmid carrying a cDNA encoding for AIFM3-Flag, transiently transfected HEK293T AIFM3-KO cells to enable AIFM3 indirect detection using anti-Flag antibodies. Confocal imaging of AIFM3-Flag expressing cells showed partial localization to the plasma membrane and actin ruffles. Our results show that AIFM3 is likely directed to plasma membrane/actin ruffles via N-myristoylation at Glycine2. As an alternative plan, we overexpressed wild type AIFM3 in HEK293T cells and performed a RNAsequencing analyses to identify the associated biological pathways. Among the significantly enriched GO terms, we observed pathways related to ciliogenesis and cilia-assembly. Specifically, analysis of Differentially Expressed Genes (DEGs) that were involved in the selected pathways showed that ciliogenesis related genes are suppressed. Among the few upregulated genes in the selected pathways, we found BAG Cochaperone 3 (BAG3), which is reported to negatively regulate the formation of cilia in glioblastoma and breast cancer. Therefore, we speculate that AIFM3 overexpression negatively regulates the biogenesis of cilia by attenuation of key regulatory genes related to cilia homeostasis and upregulation of negative regulators of cilia formation. loss of cilia was reported to promote epithelial–mesenchymal transition (EMT), a process where epithelial cells lose their differentiated, adhesive characteristics and gain migratory, invasive properties. Therefore, it is likely that AIFM3 upregulation in late stages of cancer leads to cilia loss and consequently to the EMT activation and metastasis, a speculation that needs further validation. In conclusion, here we provide a new set of data that allow to better characterize the structure, localization, and the role of AIFM3, suggesting its potential function as a ciliogenesis inhibitory factor. Further research may help to shed light on AIFM3 and the mechanisms it uses.