The Golgi apparatus of the cell has long baffled biologists, mainly because it is unclear how proteins are conveyed through it on their way to the cell surface. Some innovative microscopy may resolve the issue.
The Golgi apparatus is the enigmatic organelle responsible for modifying newly synthesized proteins that are destined to be secreted from the cell. Discovered by Camillo Golgi in 1898, the apparatus consists of a stack of disc-shaped membranes called cisternae. Newly made secretory proteins are delivered to the cis face of the Golgi. They exit, fully processed, at the opposite end — the trans face. But how do the proteins traverse the Golgi stack? Debate on this hotly contested question has persisted, in one form or another, for about 40 years. Now, two papers in this issue1,2 claim to resolve the matter in favour of one of the two competing transport mechanisms — the so-called ‘cisternal maturation model’.
This model proposes that the cisternae progress through the Golgi, gradually moving through the stack as new layers form at the cis face and old layers disperse from the trans face, and that they carry the secretory proteins with them (Fig. 1a). By contrast, the ‘vesicle-shuttle model’ (Fig. 1b) proposes that Golgi cisternae are long-lived structures, with secretory proteins being transported from layer to layer in small bubble-like membrane vesicles, which bud off one cisterna and fuse with the next one to disgorge their protein cargo.
At the heart of the controversy are the transport vesicles called COPI vesicles — after the coat-protein complex I (COPI) proteins that encrust their surface. There is a broad consensus that COPI vesicles bud from Golgi cisternae3, but no clear evidence about what these vesicles carry, or in which direction they travel — cis to trans, or the reverse4. In yeast, some secretory proteins can be secreted in the absence of COPI function, suggesting that there must be another mechanism for their transport5. Furthermore, algae secrete large sugar–protein conjugates (called scales), which are processed in the Golgi and can be 20 times the size of a COPI vesicle6. These observations argue in favour of the Golgi cisternae carrying the material forward, as proposed in the cisternal maturation model.
But what then do the COPI vesicles do? One proposal is that they carry Golgi proteins in the retrograde direction, recycling resident Golgi proteins from cisternae that are fragmenting at the trans face and incorporating them into new layers at the cis face. In this scheme, the cisternae break up at the trans face to make COPI vesicles, as well as the secretory vesicles that carry secretory proteins for the final step of their journey. The proteins end up either at the plasma membrane, where they are expelled from the cell, or at an organelle called an endosome. This version of the cisternal maturation model received a boost when Bonfanti et al.7 showed that 300-nm procollagen bundles, which are eventually secreted, travel forward through mammalian Golgi stacks without leaving the cisternae.
How do proponents of the vesicle-shuttle model interpret these results? Orci and colleagues8 proposed that bulky secretory cargoes are transported by ‘mega-vesicles’ that can be substantially larger than conventional COPI vesicles. This is not impossible. The kinetics of membrane fission (the mechanism by which a budding vesicle separates from the membrane on which it forms) might regulate the size of a vesicle, so that bulkier cargoes delay the final fission event until a vesicle of suitable size has been generated9. Another argument against cisternal maturation is that, contrary to its predictions, several studies report that resident Golgi proteins are not concentrated in COPI vesicles4.
These unresolved issues meant that the cisternal maturation concept remained unproved. But, just as it looked as if the debate would drag on for ages yet, Losev et al. (page 1002)1 and Matsuura et al. (page 1007)2 have independently produced striking images that purport to show direct evidence for cisternal maturation.
The authors exploited the fact that in the budding yeast Saccharomyces cerevisiae, the Golgi cisternae are not stacked. This organization makes it possible to follow cisternal dynamics by video microscopy of live cells. Resident proteins characteristic of early and late Golgi cisternae were tagged with two differently coloured fluorescent tags to mark the different cisternae (for example, green for early cisternae and red for late ones). Remarkably, the tagged cisternae changed colour over time, showing a consistent progression. For instance, a green early Golgi cisterna became gradually yellow and then red, as it lost the early Golgi marker and acquired the late Golgi marker. (In such imaging studies, the presence of both green and red fluorescence gives a yellow colour.) These findings are consistent with a cisternal maturation model for Golgi trafficking in budding yeast.
The story is not yet complete: a few key predictions of the cisternal maturation model remain to be tested. For instance, secretory cargoes should remain within the cisternae as they mature, while the resident Golgi proteins come and go — the obvious experiment to confirm this would be to fluorescently tag secretory cargoes at the same time as resident Golgi proteins and watch their relative progress.
Another prediction is that secretory cargoes present in an early cisterna should move through the Golgi together. Preliminary evidence is that they do: Losev et al.1 report that two yeast secretory cargoes (α-factor and carboxypeptidase Y) traverse the Golgi at roughly the same rate. Moreover, Mironov et al.10 showed that two mammalian secretory cargoes, procollagen and the vesicular stomatitis virus glycoprotein (VSV-G), travel through the Golgi at approximately the same speed. But VSV-G at high concentrations might make large protein complexes, and therefore behave as a large cargo, so additional experiments are needed to look at smaller cargoes in mammalian cells.
The fate of the late cisternae also needs to be clarified. If the cisternal maturation model is correct, they should mature into secretory vesicles and other types of carrier, but this has yet to be confirmed by fluorescence microscopy. A related prediction is that blocking retrograde traffic should block cisternal maturation and send Golgi-resident proteins to the cell surface or the endosome. Surprisingly, Matsuura et al.2 find that in a yeast mutant with a defect in COPI vesicle assembly, cisternal maturation is slowed about three-fold — but it still occurs. How are resident Golgi proteins being recycled in the absence of COPI function? Do resident Golgi proteins escape from the organelle under these conditions?
Even if cisternal maturation does occur, the details of the mechanism have yet to be defined. For instance, how is the polarized cis-to-trans distribution of resident Golgi proteins maintained? One suggestion is that it might be by the differential recycling of the Golgi proteins4. And what of other cell types? It is quite possible that in S. cerevisiae, the Golgi is formed de novo and then consumed during each round of secretory protein transport. In mammalian cells, however, there is no evidence that Golgi membranes form de novo under physiologically relevant conditions during sequential rounds of secretion. Maybe these cells rely instead on a variety of trafficking modes commensurate with cargo size and Golgi organization. Whatever the answer may be, the images provided by Losev et al.1 and Matsuura et al.2 have certainly swung the scales heavily towards cisternal maturation, in yeast at least.
About this article
Histochemistry and Cell Biology (2008)