The Golgi complex: a common platform for canonical and non-canonical autophagy?

Naydenov et al. found that siRNA-mediated downregulation of αSNAP, a component of the SNARE system, triggers autophagy as measured by the lipidation of LC3 and flux analysis. The SNARE system participates in vesicular fusion. Its key components include N-ethylmaleimide sensitive factor (NSF) and its adaptor soluble NSF-attachment protein α (αSNAP). Notably, downregulation of NSF had no effects on LC3 lipidation. This suggests that in mammalian cells, the autophagy regulatory function of αSNAP is separated from its other functions that involve its partner protein NSF.

The SNARE system is involved in the ER-Golgi trafficking. It is thus possible that disruption of the normal ER-Golgi interaction, ER function or the Golgi complex could be an autophagy signal. Indeed, knockdown of αSNAP caused Golgi fragmentation. Pharmacological agents that cause Golgi fragmentation and inhibit ER to Golgi trafficking, such as Brefeldin A (BFA) and Golgicide A (GA), are known to cause autophagy. BFA actually inhibit three Golgi-resident guanine nucleotide exchange factors (GEFs) for the Arf small GTPases, GBF1, BIG1 and BIG2, whereas GA only inhibits GBF1. Consistently, knockdown of GBF1, but not BIG1 and BIG2, induced LC3 lipidation. These observations indicate that an important autophagy trigger could be the disruption of Golgi function and/or structure.

Interestingly, the Golgi complex has been closely linked to autophagy in several ways. The Golgi complex has been considered to be one of the potential membrane sources of autophagosomes. Several well-defined autophagy molecules, such as Beclin-1, Atg9, Rab32 and Rab33B, can be located at the Golgi complex.- Moreover, as strengthened by the present work, disruption of Golgi function and/or structure can trigger an autophagic process. Fission of the Golgi membranes had been observed during starvation-induced autophagy, which is thought to be related to the re-distribution of Atg9 from the Golgi complex to vesicles that may be related to the biogenesis of autophagosomes. Could Golgi fragmentation, caused by the knockdown of αSNAP and GBF1, or by BFA and GA, be related to autophagosome generation in these cases? While this possibility exists, there is at least one important distinction between the two cases. In the starvation-induced autophagy, both Bif-1 and Beclin-1 are required for the Atg9 redistribution and fission of Golgi complex. In the study by Naydenov et al., Bif-1 but not Beclin-1 was required for the LC3 lipidation. It is not clear whether these molecules are required for Golgi fission and for Atg9 redistribution, if it ever occurs, in the scenario involving αSNAP or GBF1 knockdown. Likewise, if BFA and GA trigger autophagy in the same way as the downregulation of αSNAP, the role of Bif-1, Atg9 and Beclin-1 would have to be examined as well.

Nevertheless the Beclin-1-independent nature suggests a different type of autophagy, known as non-canonical autophagy (NCA), which has been reported in many other cases. It seems that NCA could be triggered by many different factors, and all could be independent on one or more components of the Beclin-1-PI3KC3 complex or other autophagy molecules. However, it is not clear at all, how NCA could occur in the absence of the initiation/nucleation complexes. One caveat of NCA is that it is measured in many studies by LC3 lipidation only without further functional assessment and/or strong electron microscopic morphological evidence. Thus, whether NCA is an autophagy process in the normal sense, i.e., as functional and productive as the canonical autophagy, has yet to be fully determined. The present study by Naydenov et al., however, did provide supportive evidence based on flux analysis that NCA triggered by αSNAP knockdown could be productive. Future studies would need to be directed further on its functional significance.

If fission of Golgi membrane is a key step toward autophagosome biogenesis, then molecules like αSNAP and GBF1 would be important gate keepers to restrict the fission and, thus, the membrane supply for autophagosome biogenesis. It also seems that just removing the gate keepers without additional signals could be sufficient to trigger an autophagy process that can bypass some of the key mechanisms. It could be speculated that the bypassed mechanisms, such as the one controlled by the Beclin-1-PI3KC3 complex, may function in fragmenting the Golgi complex, in addition to other possible capacities, during canonical autophagy.