Christian Koch

• The vast majority of all mitochondrial proteins are synthesized in the cytosol. These proteins carry characteristic targeting motifs within their sequence, which allows for the binding of chaperones, that in turn usher precursors to the mitochondrial surface for import and assembly. Though, our understanding of these early reactions is still lacking, recent efforts have shown that the ER surface can facilitate the import of mitochondrial proteins (ER-SURF) with the help of the J-protein Djp1. Close cooperation of organelles in form of membrane contact sites is crucial for cellular function. The aim of my work was to investigate whether ER-mitochondria contact sites are critical for the transfer of proteins from the ER to mitochondria. Several contact sites have been characterized between ER and mitochondria in S. cerevisiae. One contact site is called the ER mitochondria encounter structure (ERMES) and another is partly formed by Tom70. Owing to the high propensity of suppressor mutations in ERMES, I employed a knockdown approach to deplete this contact site. Using an inducible CRISPR interference (CRISPRi) system, I could rapidly and efficiently deplete Mdm34, which is a part of ERMES. I could show that depletion of Mdm34 had a synthetic negative effect in combination with a deletion of TOM70. Loss of both contact sites led to a strong decrease of many mitochondrial proteins in the whole cell proteome. Using affinity purification of ER and mitochondria in conjunction with mass spectrometry I could demonstrate that a specific set of mitochondrial proteins are enriched on the ER upon loss of Mdm34 and Tom70, which mainly were proteins of the inner membrane e.g., Oxa1 and Cox5A. Moreover, I was able to validate that the import of these proteins was hampered upon loss of both contact sites. Also, in vivo the biogenesis of Oxa1 was impeded upon single loss of Mdm34 or Tom70 and strongly impaired if both were lost. Analysis of the maximum hydrophobicity of inner membrane proteins in the ER-SURF set revealed on average a significantly higher peak compared to other inner membrane proteins. I could show that deleting or swapping the transmembrane domain of Cox5A would make it contact site independent or reliant on contact sites respectively, as revealed by an in vitro import assay. In this study I was able to demonstrate the involvement of membrane contact sites in ER-SURF and identify a list of putative clients. Furthermore, I could show that hydrophobicity of the transmembrane segment of inner membrane proteins is one determinant for ER-SURF dependence.