Role of autophagy in unconventional protein secretion

Abstract

In the secretory pathway, the secretion of proteins to the plasma membrane or to the extracellular milieu occurs via vesicular transport from the endoplasmic reticulum, via the Golgi apparatus, to the plasma membrane. This process and the players involved are understood in considerable detail. However, the mode of secretion of proteins that lack a signal sequence and do not transit through the secretory pathway has not been described, despite the fact that the literature is replete with examples of such proteins. One such protein is an evolutionarily conserved, secreted Acyl-CoA binding protein (known as AcbA in Dictyostelium discoideum, Acb1 in yeast and diazepam-binding inhibitor in mammals). Two recent papers highlighted in this punctum have elucidated the pathways required for the unconventional secretion of Acb1 in Pichia pastoris and Saccharomyces cerevisiae. Both implicate autophagy proteins and autophagosome formation in the process, while also uncovering roles for other interesting proteins in the unconventional secretion of Acb1.

AcbA protein is important for signaling during Dictyostelium spore formation. It is produced by pre-spore cells, is proteolytically processed by the TagC protease present on pre-stalk cells and is required by the pre-spore cells for sporulation. We found that in the methylotropic yeast P. pastoris, Acb1 is also required for sporulation and it can trigger sporulation in Dictyostelium, when added in trans. Since Acb1 is secreted via an unconventional pathway, this bioassay allowed the two groups to screen for yeast mutants impaired in secretion of Acb1.

In accordance with previously published work, Acb1 secretion is dependent on the yeast homologue, Grh1, of the mammalian Golgi-associated protein, GRASP. However, to our surprise, we find several peroxisome biogenesis (but not pexophagy) mutants and autophagy mutants to be blocked in Acb1 production. Further genetic dissection suggests the requirement of peroxisomally-generated, medium chain fatty-acyl CoA (MCFA-CoA) for Acb1 secretion. Both Pex11, a MCFA transporter residing in the peroxisomal membrane and Faa2, an intraperoxisomal fatty-acyl CoA synthetase are necessary for Acb1 secretion, as are other Pex proteins required for the targeting of Pex11 and Faa2 to the peroxisome membrane and matrix, respectively. It remains to be tested if peroxisomes are also required for the secretion of other unconventionally-secreted proteins.

Interestingly, yeast Acb1 is secreted in a pulse between 3–3.5 h after initiating starvation for nitrogen, a condition that triggers an autophagic response. Rapamycin, a drug that releases the inhibitory action of the TOR pathway on autophagy, induces Acb1 secretion, whereas autophagy mutants, including those required for autophagosome formation, are blocked in Acb1 secretion. Based on these results, we hypothesize that autophagosome formation is necessary for Acb1 secretion and that the protein bound to MCFA-CoA is captured in autophagosome-like vesicles.

All autophagy-related processes (general nonselective, as well as selective, autophagy of specific cargoes such as organelles) invoke the formation of a double-membrane structure called the autophagosome (during macroautophagy), which captures the cargoes and delivers them to the vacuole (the yeast lysosome) for degradation and recycling. Enroute to the lysosomes, autophagosomes fuse also with other vesicles, such as the endosomes, to form secondary vesicles (amphisomes), especially in mammalian cells. Normally, these amphisomes subsequently fuse with lysosomes to form autolysosomes in which cargo degradation occurs.

However, many extracellular pathogens that enter cells by endocytosis are entrapped in autophagosomes or amphisomes, whose further maturation and fusion with lysosomes is blocked or delayed by proteins encoded by the pathogens. In such situations, autophagosome or amphisome fusion with lysosomes is pre-empted or bypassed. There are also physiological conditions known in other systems wherein amphisomes fuse with the plasma membrane. This is exemplified by the human-immortalized myelogenous leukemia line K562, in which fusion of autophagosomes with MVBs to generate amphisomes is induced by starvation, and the MVBs fuse with the plasma membrane to secrete exosomes.

Interestingly, both studies find that in the case of Acb1 secretion, mutants affected in autophagosome/vacuole fusion do secrete Acb1, showing that Acb1-containing autophagosomes (called Acb1 vesicles to distinguish their distinct fate) do not suffer the same fate as ordinary autophagosomes that fuse with lysosomes.

How are autophagosomes destined for fusion with the vacuole discriminated from Acb1 vesicles that are directed towards secretion? Our data implicate Atg11, the autophagy scaffolding protein required for selective autophagy, such as the Cvt and pexophagy pathways, in the process of Acb1 secretion, indicating that the capture of Acb1 into autophagosome-like vesicles might be a selective process.

What then is the fate of these proposed Acb1 vesicles and how is this related to Acb1 secretion? Duran et al. found proteins required for endosome fusion as well as the fusion of the isolation membrane of autophagosomes (e.g., Ypt6 and Tlg2) and multi-vesicular body (MVB) formation (e.g., Vps4 and Vps23), respectively, are blocked in Acb1 secretion. This led the two groups to propose that the Acb1 vesicles fuse with endosomes and or MVBs, to form amphisomes in yeast. It is proposed that fusion of these Acb1-containing amphisomes with the plasma membrane would cause secretion of Acb1 into the extracellular space. In this model, it would be Acb1 surrounded by a single membrane that would, in fact, be released into the extracellular space, but the membrane might pop in the extracellular milieu, to release Acb1 into the medium. Consistent with this hypothesis both labs discovered that mutants defective in vesicle fusion events at the plasma membrane (specifically the SNARE Sso1 or a phospholipase D, Spo14, that recruits Sso1 to the plasma membrane) are unable to secrete Acb1.

The two papers also show that as in Dictyostelium, yeast Acb1 is processed by an unidentified trypsin-like protease into a peptide known as SDF2, which is the bioactive molecule responsible for sporulation in the bioassay of yeast extracts using Dictyostelium pre-spore cells. Thus the secretion and processing of Acb1, and the role of the SDF2 generated from Acb1, are conserved in evolution.

In summary, although a couple of decades of intensive research has elucidated the proteins and mechanisms used in the secretory pathway in great detail, the unconventional secretion of proteins is hardly understood. These papers define about a dozen proteins required for the unconventional secretion of Acb1 and establish the framework for more detailed investigations of the molecular steps along the route taken by Acb1. It will be interesting to see how widely this pathway is used by other unconventionally-secreted cargoes.

Manjithaya R, Anjard C, Loomis WF, Subramani S. Unconventional secretion of Pichia pastoris Acb1 is dependent on GRASP protein, peroxisomal functions, and autophagosome formation. J Cell Biol. 2010;188:537–46. doi: 10.1083/jcb.200911149.

Duran JM, Anjard C, Stefan C, Loomis WF, Malhotra V. Unconventional secretion of Acb1 is mediated by autophagosomes. J Cell Biol. 2010;188:527–36. doi: 10.1083/jcb.200911154.