Edited by: Vladimir Lupashin, University of Arkansas for Medical Sciences, United States
Reviewed by: José A. Martínez-Menárguez, University of Murcia, Spain; Frederic A. Bard, Institute of Molecular and Cell Biology (A*STAR), Singapore
This article was submitted to Membrane Traffic, a section of the journal Frontiers in Cell and Developmental Biology
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Intracellular transport is one of the most confusing issues in the field of cell biology. Many different models and their combinations have been proposed to explain the experimental data on intracellular transport. Here, we analyse the data related to the mechanisms of endoplasmic reticulum-to-Golgi and intra-Golgi transport from the point of view of the main models of intracellular transport; namely: the vesicular model, the diffusion model, the compartment maturation–progression model, and the kiss-and-run model. This review initially describes our current understanding of Golgi function, while highlighting the recent progress that has been made. It then continues to discuss the outstanding questions and potential avenues for future research with regard to the models of these transport steps. To compare the power of these models, we have applied the method proposed by K. Popper; namely, the formulation of prohibitive observations according to, and the consecutive evaluation of, previous data, on the basis on the new models. The levels to which the different models can explain the experimental observations are different, and to date, the most powerful has been the kiss-and-run model, whereas the least powerful has been the diffusion model.
The structure of the ER–Golgi interface and the Golgi complex (GC) is well-known and has been described many times (Mironov et al.,
Main molecular machines involved in the ER-to-Golgi and intra-Golgi transport steps (summarized from current reviews).
SNAREs (Hong) | Syntaxin5/STX5(Qa-SNARE); GS27/Membrin (Qb-SNARE); BET1 (Qc-SNARE); Sec22(R-SNARE) | Syntaxin5/STX5(Qa-SNARE); GS27/Membrin ( |
Rabs (Lamber et al., |
Rab1a/b; Rab2a/b; | Rab6a/b/c; Rab30; Rab33b; Rab43 |
COPII (Peotter et al., |
+ | – |
COPI/ARF (Béthune and Wieland, |
+ | + |
The multisubunit tethering complexes [TRAPP, Dsl1/Zw10, COG (Smith and Lupashin, |
+ | + |
Cargo receptors (p24 family, ERGIC53, KDELR, TGN46; Stanley, |
+ | + |
Golgins and matrix proteins (Ungermann and Kümmel, |
USO1/p115 | USO1/p115; GM130; Giantin; GRASP55; GRASP65; Golgin45; Golgin67; Golgin84; Golgin97/Arl1; Golgin160; Golgin245. GCC185; Syne1; CASP; Bicaudal;… |
Glycosylation enzymes and Nucleotide sugar transporters perform glycosylation regulating the transport (Stanley, |
Too many |
Scheme showing the two variants of the KARM: symmetric and asymmetric. In the asymmetric case (bottom), the cargo domain and a cellular compartment are separated by a thin tubule. Fusion takes place in one site, and fission occurs in another site where this tubule is localized. To function more precisely, it is important to concentrate SNAREs over the cargo domain, which then fuses with the distal compartment (the upper circle).
Scheme showing the main principles of intracellular transport.
For the VM, the main falsifying (prohibitive) observation is IGT of megacargoes. For example, the VM poses that IGT is carried out by COPI vesicles. However, COPI vesicles have a diameter of 52 nm, and thus megacargoes cannot be transported by COPI vesicles. However, they are transported (Bonfanti et al.,
The prohibitive observations for the DM (see
Increased concentrations (augmentation of the numeric density) of any cargo, and especially of megacargoes, are the prohibitive observation for the CMPM. Also, the speed of the cargo delivery from
The main principle of the KARM is fusion before fission, and even that fusion results in fission. Fusion/ fission might occur at the same site (i.e., symmetrical variant) or at different sites (i.e., asymmetrical variant). Within the framework of the asymmetrical KARM, fusion would be between the edges of the proximal and distal compartments whereas fission would be somewhere within the proximal compartment where rows of pores or thin tubules should be localized and SNAREs should be concentrated over the cargo domains (Mironov et al.,
The prohibitive observations for the KARM are the following:
Membrane cargoes and megacargoes should be organized in domains. Large cargo domains are more effective for transport than vesicles. It is necessary to have the concentrating of SNAREs over cargo domains. SNARE should be concentrated over cargo domains. The concentrating of cargoes and SNAREs increases efficiency. Pores or thin tubules behind a cargo domain provide directionality. Between the cargo domain and the proximal compartment there should be a thin membrane tubule(s), which connects a cargo domain with the compartment domains Pores should be consumed during transport. There should be a negatively exponential regression line for the process of emptying of the GC For EGT and post-Golgi transport, the bolus-like mechanism should be optimal. All of the compartments should always be connected. However, the existence of many SNAREs already denies this falsification observation. When secretory compartments cannot attach to each other. Previously the KARM could not explain IGT in
The VM of EGT poses that the exit of cargo proteins occurs in COPII-coated buds. These buds undergo fission and form COPII-coated spherical vesicles (Antonny et al.,
Scheme of ER-to-Golgi transport according to the VM and CMPM. Upper section:
The main support for the VM of EGT was the study by Kaiser and Schekman (
The second corner-stone study assumed to be in favor of the VM is that of Barlowe et al. (
There is no doubt that COPII is important for the exit of several cargoes from the ER (Aridor and Balch,
In spite of the importance of COPII, EGT occurs even in the absence of COPII (reviewed by Mironov,
However, the main contradiction to the VM is the exit of megacargoes, such as pre-chylomicrons [in enterocytes (Sabesin and Frase,
Megacargoes cannot be inserted into 65–80 nm COPII-dependent vesicles. It is also highly unlikely that megacargoes disassemble into smaller subunits that can be packaged into conventional transport vesicles. For instance, the HSP47 protein helps PCI to form rigid 300-nm trimers already in the ER, and the environment in the GC is not suitable for their disassembly (Bruckner and Eikenberry,
To solve these contradictions, it has been proposed that large cargoes are transported by “megavesicles,” or “megacarriers” that are formed by unusual combinations of isoforms of COPII subunits (Fromme and Schekman,
To demonstrate that megavesicles exist, Gorur et al. (
Recently, McCaughey et al. (
Also, Patterson et al. (
Within the framework of the DM, EGT occurs by diffusion along constant connections between the ER and the GC. The precise characteristics of the DM of EGT have not yet been specified in the literature. Direct membrane continuities between the ER and the GC have been described many times (Flickinger,
Trucco et al. (
According to the CMPM, immature ER-to-Golgi carriers are formed by protrusion from the ER, whereas ER-resident proteins are eliminated from the ER-to-Golgi carriers by retrograde COPI-dependent vesicles (Mironov et al.,
If COPI-coated vesicles mediate retrograde, Golgi-to-ER, transport, the concentrating of proteins with KKXX motifs would be expected, such as ERGIC53/58 or p24, in COPI-coated buds. However, to date, there has been no convincing evidence that demonstrates the concentrating of either ERGIC53/58 or p24 in COPI-coated buds on ERES. Moreover, alpha 2 protein of the p24 family is not enriched in Golgi buds (Dominguez et al.,
The characteristics of the KARM of EGT have not yet been specified in the literature. Here, we proposed our variant of the KARM of EGT (
Scheme of ER-to-Golgi transport according to the KARM. Protrusions containing small cargoes
Thus, the VM, DM, and CMPM cannot explain all of the data (i.e., prohibitive observations) that are contradictory to their logic, whereas the KARM should explain corner-stone observations that support the VM, DM, and CMPM.
The main problem for the VM of IGT (Palade,
Scheme of the VM for intra-Golgi transport (see also movie
Scheme of megavesicle-based intra-Golgi transport. This hypothesis assumes that completely isolated megavesicle (in the middle of both images) can be found near the Golgi complex.
It is established now that the vast majority of cargoes are absent from COPI vesicles. We previously presented a list of cargo proteins that are excluded from COPI vesicles (Mironov et al.,
Furthermore, to provide additional support for the VM, the Rothman group (Pellett et al.,
Another problem with the interpretations presented by Pellett et al. (
There are also other problems with the VM. There is a significant decrease in the number of COPI vesicles during synchronous IGT (Rambourg and Clermont,
On the other hand, they did not observed vesicles on strings, which they had described earlier (Orci et al.,
There are several observations that favor the DM. To be relevant, the DM should be based on structures that are interconnected. Tubular connections between Golgi cisternae have been demonstrated by Marsh et al. (
Some lipids can be easily transported along the secretory pathway when the formation of vesicles is inhibited (Sleight and Pagano,
The cisternal maturation–progression model (Mironov et al.,
Scheme of intra-Golgi transport according to the CMPM. The main postulate of this model is that during intra-Golgi transport, the amount of cargo inside the cisterna during its progression is not changed, and that COPI vesicles (COPI, black dots; Golgi-resident proteins, colored dots) should be concentrated in COPI vesicles.
We demonstrated the concentrating of albumin (Beznoussenko et al.,
The main prohibitive observation for the CMPM is the concentrating of cargo during intra-Golgi transport. We have shown that albumin is concentrated during IGT (Beznoussenko et al.,
The problem of the concentrating of Golgi-resident proteins in COPI-dependent vesicles is very serious (Glick et al.,
Further, (Martinez-Menárguez et al.,
Recent observations that demonstrate that mannosidase I can be recycled by COPI vesicles are not convincing (Rizzo et al.,
Indeed, they stated, “All reconstructions indicated that the morphology of the carriers and the structure and size of the Golgi stack were similar under all experimental conditions. Moreover, tomography confirmed that most round, 50- to 80-nm structures were indeed vesicles, and that the relative frequency of vesicles and tubules was similar to that seen in thin sections (not depicted).” To demonstrate that round profiles represented separated vesicles, they showed very small sized serial electron microscopy tomography images of only one round profile. However, this round profile showed a visible neck that connected it with a Golgi cisterna. This neck is visible on frames 45–55 of Rizzo et al. (
During synchronous IGT, the number of COPI vesicles should be sufficient for the recycling of all of the resident proteins. However, when a large amount of cargo moves across the GC, even if we take into consideration the maximal possible speed of COPI vesicle formation (Mironov et al.,
Sialyltransferases and fucosyltransferases are present within the
Moreover, although there have been several statements that COPI-coated buds can be found within the TMC, or even the TGN, in reality this has not been completely established. There was no convincing evidence in the images that (Martínez-Menárguez et al.,
The third prohibitive observation for the CMPM is the situation when renovation and progression of Golgi cisternae can be blocked. However, under these conditions IGT was observed although it became slower (Dunlop et al.,
The CMPM has several other problems. The full list of the CMPM problems was presented in our previous review (Mironov et al.,
To be efficient, the KARM of IGT should be based on several prerequisites (
Scheme of intra-Golgi transport according to the KARM.
Pores that separate cisternal distensions from the rest of the Golgi cisternae were shown by Claude (
Thus, restoration of pores in the cisternal rims might be based on this mechanism. The KARM gives the following predictions: (1) if pores inside cisternae are consumed, there should be the need for resting of the Golgi stack; (2) recycling of Ykt6 is improbable. As such, there should be one use of this SNARE, and after consumption of the cytosolic pool of Ykt6, there should be the need for the resting of the whole Golgi complex in the entire cell. Thus, there could be several waves of cargo, and only then would the pores be consumed. Fission and then fusion of COPI vesicles might induce the formation of tubules and restoration of the number of pores along cisternal rims (Park et al.,
Careful analysis of images presented by Ladinsky et al. (
When membranes are transported through the GC, the asymmetric variant of the KARM should be used. According to this, to increase the efficiency of IGT, there should be cargo domains where a cargo is concentrated. These domains should contain a set of SNAREs complementary to those in COPI vesicles (GS27, GOS28) (Fusella et al.,
Thus, at the level of ER-Golgi and IGT the VM faces with the problem of the transport of megacargoes. Attempts to modify the VM by addition of so called megabuds and megavesicles were not convincing. We suggest that the megavesicles observed by Gorur et al. (
CMPM has problems at both steps of intracellular transport, namely, at the level of the exit from the ER, it cannot explain the problem of different concentration of different cargoes whereas at the Golgi level, it cannot explain concentration of megacagoes during IGT. Although when we faced similar discrepancy between the
The DM cannot explain the necessity of SNAREs for intracellular transport and concentration of cargo at different level of the transport. Thus, the VM, DM, and CMPM cannot overcome their prohibitive observations. Also the VM, DM, and CMPM cannot explain the mechanisms of Golgi ribbon formation and the disappearance of the GC in
Scheme of intra-Golgi transport in
The vesicle delivery is a very important problem for both VM and CMPM and especially for
The CMPM cannot explain the observation that demonstrates that in
The scheme explains why the Golgi complex disappears after block of cargo exit from the ER in
The important issue, which the KARM should explain, is the role of COPI and COPI-dependent vesicles. Within the framework of the KARM, COPI vesicles are important for: (1) elimination of excessive membrane curvature (Beznoussenko et al.,
There are two transport steps where this scheme might be not very obvious; namely, EGT and post-Golgi transport. There, according to the KARM, the tubule from the GC has to move toward the ER exit site (ERES), and this should induce a bolus-like delivery of ER-to-Golgi carriers. At the post-Golgi stage, the tubule or endosome
Thus, the KARM give the following predictions. (1) The cargo should be organized in the domain more or less clearly separated from the domains, which are formed by Golgi-resident proteins. (2) Initially, there should be lateral contact between the cargo domain and the Golgi domain. Then the cargo has to go together with the Golgi compartment. (3) After its arrival at ERES, the
All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
We thank Drs. C. Berrie, C. Wilson and M. Kreft for discussion, critical suggestions and editing of the manuscript.
The Supplementary Material for this article can be found online at:
compartment (cisterna) maturation–progression model
coatomer
diffusion model
ER-to-Golgi transport
endoplasmic reticulum
ER exit site
Golgi complex
intra-Golgi transport
kiss-and-run model
procollagen I
very low-density lipoprotein
vesicular model
G protein of the VSV virus.