Mariko Ohsumi

A ubiquitin-like system mediates protein lipidation

Autophagy is a dynamic membrane phenomenon for bulk protein degradation in the lysosome/vacuole. Apg8/Aut7 is an essential factor for autophagy in yeast. We previously found that the carboxy-terminal arginine of nascent Apg8 is removed by Apg4/Aut2 protease, leaving a glycine residue at the C terminus. Apg8 is then converted to a form (Apg8-X) that is tightly bound to the membrane. Here we report a new mode of protein lipidation. Apg8 is covalently conjugated to phosphatidylethanolamine through an amide bond between the C-terminal glycine and the amino group of phosphatidylethanolamine. This lipidation is mediated by a ubiquitination-like system. Apg8 is a ubiquitin-like protein that is activated by an E1 protein, Apg7 (refs 7, 8), and is transferred subsequently to the E2 enzymes Apg3/Aut1 (ref. 9). Apg7 activates two different ubiquitin-like proteins, Apg12 (ref. 10) and Apg8, and assigns them to specific E2 enzymes, Apg10 (ref. 11) and Apg3, respectively. These reactions are necessary for the formation of Apg8-phosphatidylethanolamine. This lipidation has an essential role in membrane dynamics during autophagy.

The mouse APG10 homologue, an E2-like enzyme for Apg12p conjugation, facilitates MAP-LC3 modification.

Autophagy is a process for the bulk degradation of cytosolic compartments by lysosomes/vacuoles. The formation of autophagosomes involves a dynamic rearrangement of the membrane for which two ubiquitin-like modifications (the conjugation of Apg12p and the modification of a soluble form of MAP-LC3 to a membrane-bound form) are essential. In yeast, Apg10p is an E2-like enzyme essential for Apg12p conjugation. The isolated mouse APG10 gene product interacts with mammalian Apg12p dependent on mammalian Apg7p (E1-like enzyme), and facilitates Apg12p conjugation. The interaction of Apg10p with Apg12p is dependent on the carboxyl-terminal glycine of Apg12p. Mutational analysis of the predicted active site cysteine (Cys161) within mouse Apg10p shows that mutant Apg10pC161S, which can form a stable intermediate with Apg12p, inhibits Apg12p conjugation even in the presence of Apg7p, while overexpression of Apg7p facilitates formation of an Apg12p-Apg5p conjugate. Furthermore, the coexpression of Apg10p with Apg7p facilitates the modification of a soluble form of MAP-LC3 to a membrane-bound form, a second modification essential for autophagy. Mouse Apg10p interacts with MAP-LC3 in HEK293 cells, while no mutant Apg10pC161S forms any intermediate with MAP-LC3. Direct interaction between Apg10p and MAP-LC3 is also demonstrated by yeast two-hybrid analysis. The inability of mutant Apg10pC161S to form any intermediate with MAP-LC3 has ruled out the possibility that MAP-LC3 interacts with Apg10p as a substrate.

Localization of Cytosolic NADP-dependent Isocitrate Dehydrogenase in the Peroxisomes of Rat Liver Cells: Biochemical and Immunocytochemical Studies

Two types of NADP-dependent isocitrate dehydrogenases (ICDs) have been reported: mitochondrial (ICD1) and cytosolic (ICD2). The C-terminal amino acid sequence of ICD2 has a tripeptide peroxisome targeting signal 1 sequence (PTS1). After differential centrifugation of the postnuclear fraction of rat liver homogenate, approximately 75% of ICD activity was found in the cytosolic fraction. To elucidate the true localization of ICD2 in rat hepatocytes, we analyzed the distribution of ICD activity and immunoreactivity in fractions isolated by Nycodenz gradient centrifugation and immunocytochemical localization of ICD2 antigenic sites in the cells. On Nycodenz gradient centrifugation of the light mitochondrial fraction, ICD2 activity was distributed in the fractions in which activity of catalase, a peroxisomal marker, was also detected, but a low level of activity was also detected in the fractions containing activity for succinate cytochrome C reductase (a mitochondrial marker) and acid phosphatase (a lysosomal marker). We have purified ICD2 from rat liver homogenate and raised a specific antibody to the enzyme. On SDS-PAGE, a single band with a molecular mass of 47 kD was observed, and on immunoblotting analysis of rat liver homogenate a single signal was detected. Double staining of catalase and ICD2 in rat liver revealed co-localization of both enzymes in the same cytoplasmic granules. Immunoelectron microscopy revealed gold particles with antigenic sites of ICD2 present mainly in peroxisomes. The results clearly indicated that ICD2 is a peroxisomal enzyme in rat hepatocytes. ICD2 has been regarded as a cytosolic enzyme, probably because the enzyme easily leaks out of peroxisomes during homogenization. (J Histochem Cytochem 49:1123–1131, 2001)

Amino-terminal domain of ATRIP contributes to intranuclear relocation of the ATR-ATRIP complex following DNA damage

ATM and rad3‐related protein kinase (ATR), a member of the phosphoinositide kinase‐like protein kinase family, plays a critical role in cellular responses to DNA structural abnormalities in conjunction with its interacting protein, ATRIP. Here, we show that the amino‐terminal portion of ATRIP is relocalized to DNA damage‐induced nuclear foci in an RPA‐dependent manner, despite its lack of ability to associate with ATR. In addition, ATR‐free ATRIP protein can be recruited to the nuclear foci. Our results suggest that the N‐terminal domain of the ATRIP protein contributes to the cell cycle checkpoint by regulating the intranuclear localization of ATR.

apg15-1, a UGA mutant allele in the Saccharomyces cerevisiae APG16 gene, and its suppression by a cytoplasmic factor.

Autophagy is a complex cellular process by which starving cells utilize cytoplasmic macromolecules as nutritional resources. In Saccharomyces cerevisiae, more than 15 genes are involved in this process and most of them have been cloned and characterized by now. But there remains a complementation group represented by a single mutation, apg15-1, unclear as to its molecular nature. We obtained DNA fragments that functionally complemented apg15-1 and found that the responsible ORF, YMR159C, was already assigned as APG16. It was further found that apg15-1 was a UGA allele in which the 243rd base of the 450 bp coding region of APG16 was converted from C to T, and that the previously observed complementation between apg15-1 and apg16D was attributable to the action of a cytoplasmic omnipotent suppressor. This suppressor was readily cured by guanidine-HCl and also by overexpression or disruption of HSP104, indicating its close similarity to the PSI prion-like factor. Since apg15-1 is a mutation highly sensitive to termination suppression, it can be used as a tool to detect weak termination suppressors.