A noncanonical autophagy function of ATG9A for Golgi integrity and dynamics

ABSTRACT

In mammalian cells, the Golgi apparatus serves as the central hub for membrane trafficking. Notably, the membrane trafficking and Golgi integrity are tightly regulated by reversible post-translational modifications, such as glycosylation, phosphorylation and ubiquitination. Nonetheless, how the Golgi apparatus responses to stress to ensure appropriate membrane assembly and distribution of cargo is poorly understood. The Golgi resident protein ATG9A is the only multi-spanning membrane protein in the ATG family and has been demonstrated to traffic through the plasma membrane, endosomes and Golgi to deliver materials for the initiation of macroautophagy/autophagy. Our recent work reveals a noncanonical function of ATG9A for Golgi dynamics and identifies a pathway for sensing Golgi stress via the MARCHF9–ATG9A axis.

ATG9A is essential for autophagy initiation, which is mainly located in the trans-Golgi network (TGN) and trafficked between the plasma membrane, endosomes, and Golgi to deliver phosphatidylinositol-4-phosphate (PtdIns4P) to promote autophagosome maturation at all stages, including omegasomes, phagophores, and autolysosomes. The ATG9A vesicle trafficking is regulated by SRC and ULK1 kinases. Recently, ATG9A is reported to repair the plasma membrane in cooperation with IQGAP1 and the ESCRT, as a scramblase that flips phospholipids between the two monolayers of the membrane bilayer, and to mobilize lipid from lipid droplets to autophagosomes and mitochondria. Furthermore, ATG9A also delivers ITGB1 which mediates cell migration, highlighting the importance of the noncanonical functions of ATG9A. Here, we discover a new molecular mechanism of heat stress-induced Golgi dysfunction that is dependent on ATG9A ubiquitination [1]. Our findings broaden the noncanonical function of ATG9A beyond autophagy.

Our results demonstrated that the Golgi apparatus of HeLa cells becomes dispersed when treated at 43°C for 1 h, and then reassembles into a cisternae-like structure after cells are re-cultured at 37°C, indicating a dynamically revisable change in Golgi morphology. Interestingly, we found that ATG9A, similar to TGOLN2/TGN46 and Golgi matrix protein GOLGB1/Giantin, is also trafficked to the peripheral area under heat stress, although ATG9A does not always colocalize with the TGN. Next, co-immunoprecipitation (co-IP) showed heat stress promotes ATG9A interaction with AP-1 or AP-4 complexes, a process dependent on its membrane trafficking function. However, starvation-induced ATG9A dispersal is not involved in Golgi morphology and combined with heat stress does not affect the LC3-II level, indicating that heat stress-induced Golgi fragmentation and ATG9A trafficking is independent of the autophagy function of ATG9A.

To determine the definite roles of ATG9A in Golgi morphology, we reintroduced ATG9A into ATG9A knockout cells and found that reconstituted expression of only wild-type ATG9A promotes the Golgi fragmentation induced by heat stress but not the traffic mutants SS1 and SS22 (without the sorting signals). Furthermore, blocking PtdIns4P production by PI4KB/PI4KIIIβ inhibitor strongly prevents heat stress-induced ATG9A trafficking and Golgi fragmentation. These results suggest that ATG9A plays a key role in regulating Golgi integrity by its membrane trafficking upon heat stress.

Then, we sought to address how ATG9A regulates Golgi fragmentation. Under heat stress, we observed that the Golgi reassembly stacking protein GORASP2/GRASP55 becomes dispersed from the Golgi to the cytosol together with ATG9A. Additionally, GORASP2 has a stronger interaction with ATG9A under heat stress. At the same time, GORASP2 oligomerization is significantly reduced in ATG9A wild-type cells under heat stress, which is not beneficial for maintaining Golgi integrity. These results show that heat stress activates the ATG9A-GORASP2 axis to initiate Golgi fragmentation.

We next sought to explore how ATG9A responds to heat stress to regulate Golgi dynamics. First, liquid chromatography-tandem mass spectrometry (LC-MS/MS) analyses showed that ATG9A is ubiquitinated, and the level of ATG9A ubiquitination significantly increases under heat stress. Further studies revealed that ATG9A undergoes K63 linkage ubiquitination at five lysine residues. Consistently, only wild-type ATG9A, but not ATG9A 5K-R mutants, rescues heat stress-induced ATG9A ubiquitination, GORASP2 oligomerization and Golgi fragmentation in ATG9A knockout cells, indicating a requisite role of ATG9A ubiquitination in the Golgi. Next, we found that MARCHF9 is the E3 ligase that ubiquitinates ATG9A. In MARCHF9 knockout cells, dispersal of ATG9A to the cytosol and Golgi fragmentation are both inhibited under heat stress, and the interaction between ATG9A and GORASP2 is largely blocked, although GORASP2 oligomerization is unaffected. Collectively, these results show that MARCHF9-mediated nondegradable ATG9A ubiquitination is necessary for heat-stress-induced Golgi disassembly and dysfunction (Figure 1).

Figure 1.

The model for ATG9A regulating Golgi morphology via the MARCHF9-ATG9A axis under heat stress. Under heat stress, ATG9A is ubiquitinated by MARCH9 and interacts with GORASP2 to attenuate its Golgi stacking ability, resulting in Golgi fragmentation. ATG9A-dependent Golgi fragmentation also occurs in response to other Golgi stress and is required to maintain Golgi homeostasis.

Finally, we found that Golgi function on protein glycosylation and sorting are closely related to ATG9A ubiquitination (such as CTSD and LAMP2). Additionally, we observed that ubiquitinated ATG9A also plays an important role in golgicide A/GCA-induced Golgi fragmentation, indicating its general function in Golgi stress responses is not limited to heat stress.

In summary, our study reveals the important roles of ubiquitinated ATG9A in modulating Golgi morphology and homeostasis expanding the noncanonical functions of ATG9A under different situations. Further work is needed to verify whether MARCHF9 is a direct sensor of heat stress or is regulated by upstream signals and to test ATG9A-mediated Golgi dysfunction in vivo.

Funding Statement

This work was supported by Natural Science Foundation of China [9225430009, 32030026] to Yushan Zhu.

Disclosure statement

No potential conflict of interest was reported by the author(s).