Traffic 2002; 3: 521–529
Blackwell Munksgaard
Copyright C Blackwell Munksgaard 2002
ISSN 1398-9219
Review
The Golgi Apparatus: Balancing New with Old
Brian Storrie1,* and Tommy Nilsson2,*
1
Department of Biochemistry, Virginia Polytechnic Institute
and State University, Blacksburg, VA, 24061–0308, USA
2
Cell Biology and Biophysics Programme, European
Molecular Biology Laboratory, D-69012 Heidelberg,
Germany
* Corresponding author: [email protected] or
[email protected]
Most models put forward to explain cellular processes
do not stand the test of time. The ‘lucky’ few that are
able to survive extensive experimental tests and peer
critique may eventually become dogmas or paradigms.
When this happens, the amount of experimental data
required to overturn the paradigm is extensive. To
some, such inertia may seem prohibitive to scientific
progress but rather, in our opinion, this helps to maintain a degree of coherence. It is needed so that experiments and interpretations may be conducted within
relatively safe boundaries. In the field of protein transport in the secretory pathway, we have enjoyed a relatively stable and productive period for quite some time
(more than 30 years!). It is only very recently that the
field has entered into a phase where all bets seem to
be off. As in any paradigm shift, the accumulation of
experimental observations inconsistent with the old
dogma eventually reached a critical point. As we ‘reluctantly’ dispense with the long-standing paradigm of
forward vesicular transport, we face a time that is
bound to be trying as well as exciting.
Key words: COP, endoplasmic reticulum, mitosis, Golgi
apparatus, organelle inheritance, recycling, secretory
pathway, vesicles
Received 2 May 2002, revised and accepted for publication 14 May 2002
This review is intended to highlight some of the key findings
that, over the years, challenged the vesicular transport paradigm of Golgi function and to provide pointers for where we
might be heading. Obviously, we cannot cover all that has
happened in the field. We apologize to our colleagues for all
oversights.
Stable vs. Dynamic Compartments
More than four decades ago, high-quality electron microscopy studies provided the first glimpse of how the secretory
pathway operates. Based on electron microscopy, the cis-
ternal maturation model was formulated depicting Golgi cisternae as carriers of newly synthesized proteins en route to
the plasma membrane (1). In this model, cisternae continuously form at the cis side, mature progressively in the cis to
trans direction and at the trans side, disintegrate to release
the newly synthesized proteins for further transport (see Figure 1). As the cisternal maturation model was entirely based
on electron microscopy, such a dynamic model was understandably difficult to verify. Also, as intracellular transport
could be shown not to depend on protein synthesis (2), the
idea that accumulation of newly synthesized proteins could
be the driving force for continuous reformation of cisternal
membranes seemed unlikely. Eventually, the model succumbed to a more static view of the pathway where each
major membrane structure is seen as a stable and preexisting compartment.
Based on the notion that the Golgi stack appears as a regular
structure, investigators early on suggested that this structure
serves as an ideal platform over which protein and lipid modifying enzymes could be arranged in an orderly fashion (3).
Others later proposed that each cisterna contains specific subsets of enzymes arranged such that they can act on the newly
synthesized proteins and lipids in a sequential manner as these
pass through the organelle (4). Small vesicles seen in abundance between the ER and the Golgi and close to the Golgi
stack were proposed to serve as transport carriers moving the
newly synthesized proteins across the ER/Golgi boundary and
between adjacent Golgi cisternae. In agreement, biochemical
fractionation and immuno-electron microscopy confirmed
that glycosylation enzymes (as well as other marker molecules,
e.g., thiamine pyrophosphatase) indeed localize in different
parts of the stack in a manner nicely reflecting a stepwise
maturation of N-linked oligosaccharides [for review, see (5)].
Some anterograde cargo could also be observed in vesicular
structures close to the Golgi stack (6). The vesicular transport
model as we know it was born. What followed has been an unprecedented molecular elucidation of the secretory pathway
combining yeast genetics, morphology, and in vitro complementation assays to identify most of the key components that
we today take for granted (7,8).
The vesicular transport paradigm was further strengthened
by the characterization of two different types of transport vesicles termed COPI (9) and COPII (10). COPII vesicles were
shown to transport newly synthesized proteins out of the ER.
COPI vesicles were interpreted to mediate transport between
adjacent Golgi cisterna. The vesicular transport model can
thus be summarized as vectorial vesicular transport from the
ER and between adjacent cisternae where each compart521
Storrie and Nilsson
522
Traffic 2002; 3: 521–529
The Golgi Apparatus: Balancing New with Old
Figure 1: Cisternal maturation in progress. Newly synthesized
proteins concentrate and exit the ER via COPII vesicles. In A: Following uncoating, vesicles and other membranes coalesce to form vesicular tubular clusters (VTCs), which travel along microtubules towards the central and juxta nuclear Golgi apparatus. This event
takes place from 100 or so peripheral ER exit sites. In B: At the Golgi
apparatus, the VTC becomes the new cis cisterna. The previous cis
cisterna then becomes the medial cisterna, and so on. The anterograde progression of the Golgi cisterna is offset by COPI vesicles
which move glycosylation enzymes and other Golgi-resident proteins from cisterna to cisterna in the opposite direction. At the trans
side, cisternae disassemble, releasing cargo for the plasma membrane or the endocytic pathway. The remaining trans cisterna is
then possibly consumed by the ER at the ‘GERL’ via a COPI-independent but rab6-dependent Golgi-to-ER recycling pathway.
ment exists as stable and pre-existing structures having
unique compositions in terms of enzymatic activities. These
compartments must (at least at some level) remain intact
during partitioning in mitosis to give rise to new compartments and organelles. In stark contrast, the cisternal maturation model predicts that cisternae continuously form from
the ER and, therefore, that the Golgi is a self-forming organelle. In this model, the need for partitioning pre-existing structures in mitosis is therefore of less importance.
Major Headaches and Shocks to the
Vesicular Transport, Stable Compartment
Paradigm
In any vectorial transport process, there is the problem of
how to balance the flow of membranes, e.g. how to counteract the forward flow of proteins and lipids. In secretory cells,
such as exocrine pancreatic cells, during the formation of
secretory granules, cargo undergoes extensive condensation.
As a consequence, the ratio of cargo-to-membrane surface
increases dramatically, giving rise to excess membrane. A
massive membrane surplus (50-fold) was estimated at the
boundary between the Golgi and the secretory granules (11).
It was suggested that such excess membrane was removed
via a specialized region of the ER termed GERL (Golgi-ERLysosomes) (12) interdigitating the trans part of the Golgi.
This idea was discounted on the basis that the ER was unlikely to contain such a highly differentiated appendix. However, ER interdigitation with the trans-most cisterna is supported by several independent studies [see, for example
(13,14)]. Although the full details of the GERL concept may
not be experimentally supported as such, the possibility that
membrane adsorption/recycling takes place at an ER interface remains. During the same time period when the GERL
was suggested, a different membrane recycling mechanism
was put forward where bi-directional vesicles would act like
‘shuttles’ between the ER and the Golgi apparatus (15).
The idea of bi-directional vesicles was later extended by
Rothman in a review postulating that the Golgi stack serves
as a distillation tower progressively refining secretory cargo
Traffic 2002; 3: 521–529
away from the resident proteins of the ER and the Golgi cisternae (16). This model predicts extensive recycling of ER
components from the Golgi apparatus. To test this, Rothman
and colleagues examined the extent of post-ER modifications that could be found on N-linked oligosaccharides of
ER-resident proteins (17). However, only low amounts of
post-ER modifications could be observed, suggesting that
ER-resident proteins access the Golgi only to a minor extent.
This was interpreted as indicating that no or little cycling
takes place between the ER and the Golgi apparatus. With
the subsequent discovery of recycling motifs, the idea of
membrane balance/recycling was revisited and verified experimentally. These signals are present on many lumenal (18)
as well as trans membrane proteins (19,20) of the ER and
have been shown to be necessary as well as sufficient for
retrieval from the Golgi back to the ER. Recycling signals provided mechanistic means to distinguish the constituents of
the pathway from anterograde cargo. However, the machinery mediating recycling remained unclear.
The observation that coatomer subunits of the COPI coat bind
to the cytoplasmic retrieval signal, K(X)KXX, provided a crucial insight (21). In yeast, reporter molecules bearing this motif were lost from the exocytic pathway to the vacuole when
expressed in cells mutated in a-COP, a subunit of the COPI
coat (22). Together, these observations provided a firm
mechanistic link between COPI vesicles and recycling. This
did not rule out a role for COPI vesicles in forward transport,
but immediately begged the question how the same coat
could perform distinct functions. Though suggestions for bidirectional COPI vesicles have been made, in vitro experiments to directly test this have failed to reveal the presence
of anterograde COPI vesicles, at least at the level of the Golgi
stack (23). The discovery that COPI [through binding to
K(X)KXX] mediates recycling marked a turning point in the
field and helped to address part of the problem of membrane
balance, at least at the ER to Golgi boundary.
Another hallmark of the vesicular transport, stable compartment paradigm (born out in part from the idea that the Golgi
apparatus serves as a distillation tower) is that newly synthesized proteins enter anterograde transport vesicles (COPI
and COPII) by default. This idea of a ‘bulk flow’ was in part
supported by experiments using synthetic tripeptides (24).
However, later studies revealed that, in fact, newly synthesized proteins are concentrated upon ER export (25,26).
This appears to be a signal-mediated process and, indeed,
several ER export motifs have now been identified (27–29).
The findings of cargo concentration upon ER export have had
a profound effect on the field, as this eroded yet another
pillar (the bulk flow hypothesis) of the vesicular transport,
stable compartment paradigm. What then remained?
The vesicular transport, stable compartment paradigm predicted that each compartment (e.g. cisterna) is stable, preformed and that it contains a specific subset of modifying
enzymes. We will discuss stable vs. self-forming compartments more extensively below, but one major finding de523
Storrie and Nilsson
serves to be mentioned here. This is the dramatic and catastrophic collapse of the Golgi apparatus into the ER upon
addition of the fungal metabolite, Brefeldin A (BFA) (30,31).
To most, this finding revealed an unanticipated dynamic aspect of the Golgi apparatus. Added to this was the observation that each cisterna does not appear to contain distinct
sets of modifying enzymes. Rather, enzymes distribute as
gradients over several cisternae of the stack (32,33). Thus,
each cisterna could no longer be viewed as a discrete biochemical compartment. The dynamic property of the Golgi
revealed by BFA, combined with the lack of bulk flow, the
surprising role of COPI in recycling and the findings of enzyme gradients, have all helped to make the vesicular transport model less viable as a working model or paradigm. What
were then the alternatives?
One of the main arguments put forward in favor of the cisternal maturation model is that it explains how large macromolecular complexes, too large to fit into vesicular intermediates, could make it through the Golgi complex. Studies on
cell wall/scale formation in algae (34) had revealed a gradual
maturation of the cargo in the cis to trans direction. Here,
scales the width of the entire stack, seemingly form inside
the lumen of the Golgi cisternae (35). In mammalian cells,
studies on collagen revealed large parallel fiber-like structures
forming towards the trans cisternae (36). In retrospect, it is
difficult to understand how so many in the transport field
managed to ignore such striking data for such a long period
of time. It is interesting to note that in the plant Golgi field,
there has been more support for the cisternal maturation process for quite some time (37).
Cisternal Maturation as the New Paradigm?
We now turn more fully to the cisternal maturation hypothesis. In this model, membrane recycling is the driving force
for membrane balance as cisternae continuously form out of
the ER and move in a cis to trans direction. Modifying enzymes and other constituents would quickly be lost if they
were not continuously recycled. Recycling driven maturation
implies the presence of Golgi-resident enzymes in recycling
vesicles (and/or tubules). This has proven to be a proposition
for which supporting evidence has come only recently. At the
end of the 1990s, small, pleiomorphic, non-cisternal structures high in Golgi enzymatic activity were isolated from
tissue culture cells (38). These were interpreted to be recycling carriers. Also, it could be shown that the in vitro Golgi
complementation assay (8), assumed to measure the transport of VSV-G proteins from N-acetylglucosaminyltransferase
I (GlcNAc-T1)-deficient 15B Golgi membranes, actually
measures COPI-dependent transport of GlcNAc-T1 from the
wild-type membrane to the VSV-G containing 15B membrane (39). In other words, this complementation assay
measures COPI-dependent, intra-Golgi recycling.
Consistent with glycosylation enzymes in recycling carriers,
COPI vesicles formed in vitro under physiological conditions
524
where GTP hydrolysis is permitted were shown to contain
large amounts of mannosidase II and other Golgi-resident
proteins, but only low amounts of anterograde cargo (40).
Incorporation of glycosylation enzymes was found also to be
dependent on GTP hydrolysis. Previous studies examining the
cargo content of COPI vesicles had mainly been carried out
using vesicles formed in the presence of the non-hydrolyzable analog, GTPgS. This negative GTPgS effect on sorting
has also been observed in vivo (41), and provides a plausible
explanation for why glycosylation enzymes in COPI vesicles
were missed previously. In an in vivo assay, careful morphometric analysis of immunoelectron microscopy labeling also
revealed mannosidase II but not cargo proteins in COPI vesicles (42), as do studies showing exclusion of serum albumin
from budding COPI vesicles (43). Consistent with recycling
at multiple levels of the Golgi apparatus, more than one
population of COPI vesicles has been observed, each containing different resident proteins (23,42). Also, recent
studies on the anterograde transport of collagen and of VSVG show that both appear to be excluded from COPI vesicles
and to travel through the Golgi apparatus at the same rate as
the cisternae mature (44). Together, these observations point
to a major role for COPI vesicles in recycling, both from the
Golgi to the ER and between adjacent cisternae of the stack,
and for cisternae serving as carriers of the newly synthesized
cargo.
So, are we then returning to the old cisternal maturation
model? For the time being, the predictions of the cisternal
maturation model [for listing, see (45)], such as glycosylation
enzymes accumulation in COPI vesicles, appear consistent
with most data but, as in the vesicular transport model, not
with all data. There are still unresolved issues. For example,
electron microscopy studies by Orci and colleagues found no
evidence for the presence of mannosidase II in COPI vesicles.
Rather, they concluded that mannosidase II is excluded from
vesicles, whereas VSV-G protein is included (46,47). More
than one population of COPI vesicles could also be detected,
one representing anterograde vesicles, and the other recycling intermediates (46). There is also the issue of formally
proving the existence of COPI and COPII vesicles. Vesicles
have been seen in abundance tethered to Golgi cisternae
(48). However, even when extreme care has been taken to
preserve ultrastructure (14), their very existence (and role in
transport) is being questioned. Furthermore, the incidence of
vesicles varies greatly between cell types. In HeLa and COS
cells, studies have failed to reveal large numbers of vesicles.
In contrast, dedicated secreting cells, such as HIT cells,
seemingly contain thousands of vesicles. This ‘discrepancy’
may reflect cell-specific differences in steady-state population of vesicles that can be observed, rather than differences
in transport modes per se (e.g. transport through tubular connections rather than vesicles). Further work will be needed to
resolve this and related issues.
Variations of the cisternal maturation model have also been
put forward. One suggestion has been to merge the vesicular
transport model with the cisternal maturation model (this
Traffic 2002; 3: 521–529
The Golgi Apparatus: Balancing New with Old
mainly to depolarize the field). Here, anterograde and retrograde cargo enter COPI vesicles seemingly at random, while
the cisternae mature (49). Vectorial transport could take
place in this ‘percolation’ model, but its main problem is a
lack of experimental support (see above). It was suggested
also that yeast relies more heavily on cisternal maturation
than do mammalian cells. However, yeast cytosol can successfully drive COPI-dependent transport in mammalian
cells, in vitro (50) making this an unlikely scenario. Another
suggestion is that all compartments of the secretory pathway
are connected via continuous and discontinuous tubules. This
model was put forward already in the early 1970s and offers
an alternative to vesicles in transfer of material across membrane boundaries (51–53). A variation on this was recently
outlined envisaging the secretory pathway as a larger tubular
continuum with two discontinuous gates in place, one between the ER and the Golgi, and the other between the Golgi
and the TGN (54). Indeed, recent work shows that transport
of material can take place along tubules (55,56). However,
vesicles as well as tubules appear to coexist in the pathway
(57). How they relate to each other, mechanistically, and to
what extent they contribute to different transport events is
not yet known.
To summarize, we suggest that the cisternal maturation
model is a better working paradigm for the field, at least for
the immediate time. COPI vesicles have been clearly implicated in recycling, and cisternae in the transport of newly
synthesized proteins (and lipids). The issue of a continuous
self-formation process, a logical extension of the cisternal
maturation process, however, has not yet been discussed.
For that, we turn to a recently discovered pathway for protein
and lipid recycling.
A Pre-Existing vs. a Self-Forming Golgi
Apparatus
Upon COPII-dependent ER export, newly synthesized proteins can be found in what are often termed vesicular tubular
clusters (VTCs). This structure was identified first upon subjecting cells to a temperature block at 15 æC (58). It was suggested that this structure served as the first step in the maturation process whereby the VTC, en route to the central Golgi,
would mature to become a cis cisterna (Figure 1). Later, the
VTC was suggested to be a stable pre-existing structure that
served as an ER-to-Golgi intermediate compartment (ERGIC)
(59). Though a highly dynamic nature had already been suggested from electron microscopy studies (60), it was not until
the advent of the green fluorescent protein (GFP) that this
aspect of the VTC could be fully revealed (61,62). We now
know that VTCs travel along microtubules from the periphery
of the cell towards the central Golgi. En route, the newly synthesized proteins are concentrated through COPI-dependent
segregation and removal of resident proteins (63,64),
essentially confirming one of the predictions of the cisternal
maturation model, a gradually maturing ER-to-Golgi membrane structure. How is the VTC then formed? For now,
Traffic 2002; 3: 521–529
whether or not the VTC form through fusion of ER-derived
material or are nucleated about some stable core is an open
question. In the remainder of this section, we consider various approaches to investigating this question.
Almost four decades ago, experiments with microtubule depolymerizing drugs indicated that the juxtanuclear position
of the Golgi apparatus in many mammalian cells is highly
dependent on microtubules. The key observation here was
that microtubule depolymerization leads to Golgi fragmentation; the Golgi apparatus is scattered into shortened cisternal stacks located at many peripheral sites, as shown by
electron microscopy (65). In general, such experiments were
interpreted as fragmentation of a stable, pre-existing, interconnected Golgi structures into individual stack units that
then migrated to distal sites. We now know that Golgi fragments, or indeed, individual stacks, are unlikely to diffuse
through the cytosol due to the high viscosity of the cytosol.
Rather, relocation appears to be an active process, which surprisingly, involves the ER. Live cell imaging and a clever entrapment strategy showed that Golgi-resident proteins continuously cycle from the juxtanuclear Golgi to the ER, only to
emerge at or near ER exit sites (66,67). When ER-to-Golgi
transport is inhibited by nocodazole, small but functionally
intact Golgi stacks rapidly form at distal sites, suggesting that
the Golgi stack is capable of self-formation, at least partially.
These peripheral Golgi stacks are likely to originate from the
ER exit site containing transitional ER, a term first introduced
by Palade (4), as indicated by their inhibited formation when
ER exit is blocked (66). The possibility of Golgi self-formation
can be drawn also from BFA washout experiments where
Golgi stacks rapidly form upon release of Golgi-resident proteins from the ER. Indeed, studies on Golgi inheritance in
yeast suggest that early Golgi elements emerge out of the
ER (68), and formation of peripheral Golgi stacks in mammalian cells is seemingly preceded by formation of new ER
exit sites (69). Thus, the structural state of the Golgi apparatus appears to be closely related to transitional ER.
Yet data from such experiments do not exclude the possibility
of a pre-existing template for Golgi assembly. That a template
exists is based on the observation that Golgi stacks are embedded in an organizing ‘ground substance’ (70), free of ribosomes (later termed ‘zone of exclusion’). Also, an identifiable
Golgi-related, clustered set of vesicles remains after BFA
treatment (71–73). It has been proposed that this serves as
templates onto which new Golgi can form. Matrix proteins
have indeed been identified and collectively been proposed
to serve as stacking proteins and/or molecular rulers. The first
such protein to be identified, GM130, was isolated from a
detergent and salt-resistant Golgi ‘matrix’ fraction. Upon BFA
treatment, GM130 relocates into scattered peripheral structures rather than into the ER consistent with a matrix serving
function (74). Also, its juxtanuclear location appeared relatively resistant to an ER exit block (75), though more recent
studies point towards a more dynamic behavior of GM130,
including relocation to the ER (76,77). In retrospect, this may
not be so surprising. Cis proteins such as the KDEL receptor,
525
Storrie and Nilsson
Figure 2: The Golgi apparatus is dependent on protein cycling but not on continuous protein synthesis. In A: HeLa cells were
microinjected with a plasmid encoding a dominant negative Sar1p (asterisks, ER exit block). A Golgi glycosyltransferase, here a VSV-G
epitope tagged GalNAc-T2, redistributes to the ER as indicated by peri-nuclear staining. In non-injected cells (arrowhead), a compact,
juxtanuclear Golgi apparatus distribution is seen. In B: Cycloheximide (CHX at 100 mg/ml) is used to block protein synthesis and the Golgi
apparatus structure is maintained as indicated by immunofluorescence staining of the VSV-G epitope tagged GalNAc-T2.
p58/ERGIC53 and members of the p24 protein family, do
not show typical ER distribution upon BFA treatment. In fact,
15 æC and/or BFA treatment are commonly used as an indicator to reveal cis Golgi localization, as this gives rise to relocation to peripheral structures. That GM130 does not show
typical ER distribution upon BFA treatment is consistent with
its more recently reported cis localization. The same argument applies to other tethering molecules of the cis Golgi,
such as p115 and GRASP65. That this is a likely interpretation
is supported by the observation that when establishing an ER
exit block using a GDP-restricted mutant of Sar1p, ER exit
sites can no longer form (77). Under these conditions,
GM130 and other marker proteins now become evenly distributed throughout the ER.
The issue of pre-existing templates or self-formation may to
some extent be resolved by establishing to what extent peripheral structures seen upon BFA treatment or upon ER exit
block correspond to ER exit sites, or whether they are truly
discontinuous with the ER. If they are continuous with the
ER, they may be seen as the first step of the maturation
process. In contrast, if they are separated from the ER, it will
be difficult to reconcile this with the idea that Golgi membranes are outgrowths of the ER (be it COPII vesicles and/or
larger membrane structures). In support of outgrowths, recent in vitro experiments suggest that tubular transitional ER
elements, such as those originally indicated by Palade (4),
can give rise to VTCs in the absence of Golgi membranes
(78). This possibility is supported by experiments using permeabilized cells suggesting that the VTC may well comprise
part of an overall ER exit complex (79). Along similar lines,
experiments have been performed asking whether it is possible to generate a functional Golgi apparatus, de novo, using
only ER as starting material. Cell surgery experiments in
which ER containing cytoplasts free of Golgi stacks were
526
created, suggest that the cell is unable to form a Golgi, de
novo (80). However, this experiment may have been unfair
(to the cell) as the level of synthesis of most Golgi-resident
proteins is very low, raising the possibility that the cytoplast
(devoid also of the nucleus) was unable to generate enough
material to build a new Golgi, even if it tried.
Mitosis – Divide and Conquer
The issue of self-formation vs. a pre-existing template becomes more pointed when we turn to mitosis. Here, the typical juxtanuclear mammalian Golgi apparatus must be distributed evenly into the two daughter cells. On the basis of
fluorescence and electron microscopy data, it was argued
that the Golgi apparatus undergoes complete fragmentation
and vesicularization during mitosis, resulting in mitotic clusters. These are then segregated into the respective daughter
cells (81). More recent data, however, indicate that the fragmentation process is incomplete, and that perhaps a more
ordered inheritance strategy is used (82). A relationship between the Golgi and the ER in this process cannot be ruled
out, as mitotic clusters are invariably seen in close proximity
to the ER. The number of mitotic clusters is also consistent
with the number of peripheral Golgi stacks seen upon addition of nocodazole in interphase cells. Indeed, data based
mainly on live cell imaging suggest that the entire Golgi apparatus is absorbed into the ER during mitosis (83). However,
this work suffers from a limitation of GFP, in that it folds more
slowly than native molecules, giving rise to a larger than expected pool of proteins in the ER (84). Appreciable ER accumulation was observed by time-lapse over only a limited
time-window during mitosis (83), meaning that its recognition by electron microscopy or cell fractionation would be
difficult. Absorption into the ER is still an open possibility. If
Traffic 2002; 3: 521–529
The Golgi Apparatus: Balancing New with Old
true, it will be inconsistent with the vesicular traffic, stable
compartment paradigm.
That the Golgi apparatus would be intimately related with the
ER is not far fetched (85). Besides the constant COPI-dependent recycling of proteins from all levels of the pathway,
Golgi proteins can also recycle to the ER by a COPI-independent mechanism (see Figure 2). This recycling is independent
of protein synthesis, relatively slow compared to COPI-dependent recycling and has a half-time of ⬃1.5 h. COPI-independent recycling gives rise to an ER pool of several Golgi
glycosyltransferases that has been estimated to be at most
5% of the total enzyme population (66,84). Distributed over
at least a 10-fold larger surface area, this pool is too low to
cause significant glycosylation of ER-resident proteins (17)
and, therefore, to upset the ER quality control machinery
which depends crucially on the absence of complex N-linked
oligosaccharide modifications (86,87). The COPI-independent pathway corresponds to the above recycling pathway
revealed upon nocodazole treatment or temperature entrapment, and is characterized by a clear independence of treatments which inhibit COPI function but a dependency on
rab6, a small GTPase (88,89). That this ER recycling operates
in the absence of any perturbations is supported by the observation that Shiga toxin is transported from the Golgi apparatus to the ER by the same COPI-independent, rab6-dependent pathway (88,89). Also, depletion of a Golgi matrix protein
such as golgin-45 results in the ER accumulation of NAGT-I,
a medial Golgi glycosyltransferase (90), presumably due to
recycling and a matrix-specific block of Golgi apparatus reassembly.
General Conclusions and Perspectives
In summary, the evidence revealing a dynamic nature of the
Golgi apparatus structure is permissive for the cisternal maturation paradigm. There is general agreement today that integral membrane proteins such as glycosylation enzymes cycle
between the Golgi apparatus and the ER, indicating a close
relationship between these two organelles. The issue of preexisting vs. self-forming structures is still open. At the fundamental level, it is asking the question whether there is a predetermined plan of how to build the Golgi or whether its
shape and function arise simply through molecular interactions. Examples of complex biological patterns and shapes
formed through molecular interactions and energy dissipation
can be found in recent elegant work on microtubules and
spindle formation (91). Whether or not nature has extended
this to the secretory pathway remains to be seen.
Acknowledgments
Thanks to Gareth Griffiths and Joanne Young for helpful, cheerful discussions and critical comments on the manuscript. One of us (B.S.) was
supported in part by grants from the U.S. NSF and NIH.
Traffic 2002; 3: 521–529
References
1. Grassé PP. Ultrastructure, polarité et reproduction de l’appareil de
Golgi. C R Acad Sci Paris 1957;245:1278–1281.
2. Jamieson JD, Palade GE. Intracellular transport of secretory proteins
in the pancreatic exocrine cell. 3. Dissociation of intracellular transport
from protein synthesis. J Cell Biol 1968;39:580–588.
3. Gersh I. A protein component of the Golgi apparatus. Arch Path
1949;47:99–109.
4. Palade G. Intracellular aspects of the process of protein synthesis.
Science 1975;189:347–358.
5. Pfeffer SR, Rothman JE. Biosynthetic protein transport and sorting by
the endoplasmic reticulum and Golgi. Annu Rev Biochem 1987;56:
829–852.
6. Jamieson JD, Palade GE. Role of the Golgi complex in the intracellular
transport of secretory proteins. Proc Natl Acad Sci USA 1966;55:
424–431.
7. Novick P, Field C, Schekman R. Identification of 23 complementation
groups required for post- translational events in the yeast secretory
pathway. Cell 1980;21:205–215.
8. Balch WE, Dunphy WG, Braell WA, Rothman JE. Reconstitution of the
transport of protein between successive compartments of the Golgi
measured by the coupled incorporation of N-acetylglucosamine. Cell
1984;39:405–416.
9. Orci L, Glick BS, Rothman JE. A new type of coated vesicular carrier
that appears not to contain clathrin: Its possible role in protein transport within the Golgi stack. Cell 1986;46:171–184.
10. Barlowe C, Orci L, Yeung T, Hosobuchi M, Hamamoto S, Salama N,
Rexach MF, Ravazzola M, Amherdt M, Schekman R. COPII: A membrane coat formed by Sec proteins that drive vesicle budding from
the endoplasmic reticulum. Cell 1994;77:895–907.
11. Meldolesi J. Membranes and membrane surfaces. Dynamics of cytoplasmic membranes in pancreatic acinar cells. Philos Trans R Soc
Lond B Biol Sci 1974;268:39–53.
12. Novikoff PM, Novikoff AB, Quintana N, Hauw JJ. Golgi apparatus,
GERL, and lysosomes of neurons in rat dorsal root ganglia, studied
by thick section and thin section cytochemistry. J Cell Biol
1971;50:859–886.
13. Hermo L, Clermont Y, Rambourg A. Endoplasmic reticulum-Golgi apparatus relationships in the rat spermatid. Anat Rec 1979;193:243–
255.
14. Marsh BJ, Mastronarde DN, Buttle KF, Howell KE, McIntosh JR. Organellar relationships in the Golgi region of the pancreatic beta cell
line, HIT-T15, visualized by high resolution electron tomography. Proc
Natl Acad Sci USA 2001;98:2399–2406.
15. Jamieson JD, Palade GE. Intracellular transport of secretory proteins
in the pancreatic exocrine cell. I. Role of the peripheral elements of
the Golgi complex. J Cell Biol 1967;34:577–596.
16. Rothman JE. The Golgi apparatus: two organelles in tandem. Science
1981;213:1212–1219.
17. Brands R, Snider MD, Hino Y, Park SS, Gelboin HV, Rothman JE. Retention of membrane proteins by the endoplasmic reticulum. J Cell
Biol 1985;101:1724– 32.
18. Munro S, Pelham HR. A C-terminal signal prevents secretion of luminal ER proteins. Cell 1987;48:899–907.
19. Nilsson T, Jackson M, Peterson PA. Short cytoplasmic sequences
serve as retention signals for transmembrane proteins in the endoplasmic reticulum. Cell 1989;58:707–718.
20. Jackson MR, Nilsson T, Peterson PA. Identification of a consensus
motif for retention of transmembrane proteins in the endoplasmic
reticulum. EMBO J 1990;9:3153–3162.
21. Cosson P, Letourneur F. Coatomer interaction with di-lysine endoplasmic reticulum retention motifs. Science 1994;263:1629–1631.
527
Storrie and Nilsson
22. Letourneur F, Gaynor EC, Hennecke S, Demolliere C, Duden R, Emr
SD, Riezman H, Cosson P. Coatomer is essential for retrieval of dilysine-tagged proteins to the endoplasmic reticulum. Cell 1994;79:
1199–1207.
23. Lanoix J, Ouwendijk J, Stark A, Szafer E, Cassel D, Dejgaard K, Weiss
M, Nilsson T. Sorting of Golgi resident proteins into different subpopulations of COPI vesicles: a role for ArfGAP1. J Cell Biol 2001;155:
1199–1212.
24. Wieland FT, Gleason ML, Serafini TA, Rothman JE. The rate of bulk
flow from the endoplasmic reticulum to the cell surface. Cell
1987;50:289–300.
25. Mizuno M, Singer SJ. A soluble secretory protein is first concentrated
in the endoplasmic reticulum before transfer to the Golgi apparatus.
Proc Natl Acad Sci USA 1993;90:5732–5736.
26. Balch WE, McCaffery JM, Plutner H, Farquhar MG. Vesicular stomatitis virus glycoprotein is sorted and concentrated during export from
the endoplasmic reticulum. Cell 1994;76:841–852.
27. Nishimura N, Balch WE. A di-acidic signal required for selective export
from the endoplasmic reticulum. Science 1997;277:556–558.
28. Kappeler F, Klopfenstein DR, Foguet M, Paccaud JP, Hauri HP. The
recycling of ERGIC-53 in the early secretory pathway. ERGIC-53 carries a cytosolic endoplasmic reticulum-exit determinant interacting
with COPII. J Biol Chem 1997;272:31801–31808.
29. Dominguez M, Dejgaard K, Fullekrug J, Dahan S, Fazel A, Paccaud JP,
Thomas DY, Bergeron JJ, Nilsson T. gp25L/emp24/p24 protein family
members of the cis-Golgi network bind both COP I and II coatomer.
J Cell Biol 1998;140:751–765.
30. Fujiwara T, Oda K, Yokota S, Takatsuki A, Ikehara Y. Brefeldin A causes
disassembly of the Golgi complex and accumulation of secretory proteins in the endoplasmic reticulum. J Biol Chem 1988;263:18545–
18552.
31. Lippincott-Schwartz J, Yuan LC, Bonifacino JS, Klausner RD. Rapid
redistribution of Golgi proteins into the ER in cells treated with brefeldin A: Evidence for membrane cycling from Golgi to ER. Cell
1989;56:801–813.
32. Nilsson T, Pypaert M, Hoe MH, Slusarewicz P, Berger EG, Warren G.
Overlapping distribution of two glycosyltransferases in the Golgi apparatus of HeLa cells. J Cell Biol 1993;120:5–13.
33. Rabouille C, Hui N, Hunte F, Kieckbusch R, Berger EG, Warren G, Nilsson T. Mapping the distribution of Golgi enzymes involved in the construction of complex oligosaccharides. J Cell Sci 1995;108:1617–
1627.
34. Brown RM Jr. Observations on the relationship of the Golgi apparatus
to wall formation in the marine chrysophycean alga Pleurochrysis
scherffelii Pringsheim. J Cell Biol 1969;41:109–123.
35. Melkonian M, Becker B, Becker D. Scale formation in algae. J Electron
Microsc Tech 1991;17:165–178.
36. Marchi F, Leblond CP. Radioautographic characterization of successive
compartments along the rough endoplasmic reticulum–Golgi pathway
of collagen precursors in foot pad fibroblasts of [3H]proline-injected
rats. J Cell Biol 1984;98:1705–1709.
37. Beams HW, Kessel RG. The Golgi apparatus: Structure and function.
Int Rev Cytol 1968;23:209–276.
38. Love HD, Lin CC, Short CS, Ostermann J. Isolation of functional Golgiderived vesicles with a possible role in retrograde transport. J Cell Biol
1998;140:541–551.
39. Lin CC, Love HD, Gushue JN, Bergeron JJ, Ostermann J. ER/Golgi
intermediates acquire Golgi enzymes by brefeldin A-sensitive retrograde transport in vitro. J Cell Biol 1999;147:1457–1472.
40. Lanoix J, Ouwendijk J, Lin CC, Stark A, Love HD, Ostermann J, Nilsson
T. GTP hydrolysis by arf-1 mediates sorting and concentration of Golgi
resident enzymes into functional COP I vesicles. EMBO J
1999;18:4935–4948.
41. Pepperkok R, Whitney JA, Gomez M, Kreis TE. COPI vesicles accumu-
528
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
lating in the presence of a GTP restricted arf1 mutant are depleted of
anterograde and retrograde cargo. J Cell Sci 2000;113:135–144.
Martinez-Menarguez JA, Prekeris R, Oorschot VM, Scheller R, Slot
JW, Geuze HJ, Klumperman J. Peri-Golgi vesicles contain retrograde
but not anterograde proteins consistent with the cisternal progression
model of intra-Golgi transport. J Cell Biol 2001;155:1213–1224.
Dahan S, Ahluwalia JP, Wong L, Posner BI, Bergeron JJ. Concentration
of intracellular hepatic apolipoprotein E in Golgi apparatus saccular
distensions and endosomes. J Cell Biol 1994;127:1859–1869.
Mironov AA, Beznoussenko GV, Nicoziani P, Martella O, Trucco A,
Kweon HS, Di Giandomenico D, Polishchuk RS, Fusella A, Lupetti P,
Berger EG, Geerts WJ, Koster AJ, Burger KN, Luini A. Small cargo
proteins and large aggregates can traverse the Golgi by a common
mechanism without leaving the lumen of cisternae. J Cell Biol
2001;155:1225–1238.
Storrie B, Pepperkok R, Nilsson T. Breaking the COPI monopoly on
Golgi recycling. Trends Cell Biol 2000;10:385–391.
Orci L, Stamnes M, Ravazzola M, Amherdt M, Perrelet A, Sollner TH,
Rothman JE. Bidirectional transport by distinct populations of COPIcoated vesicles. Cell 1997;90:335–349.
Orci L, Amherdt M, Ravazzola M, Perrelet A, Rothman JE. Exclusion of
Golgi residents from transport vesicles budding from Golgi cisternae in
intact cells. J Cell Biol 2000;150:1263–1270.
Orci L, Perrelet A, Rothman JE. Vesicles on strings: morphological
evidence for processive transport within the Golgi stack. Proc Natl
Acad Sci USA 1998;95:2279–2283.
Pelham HR, Rothman JE. The debate about transport in the Golgi –
two sides of the same coin? Cell 2000;102:713–719.
Dunphy WG, Pfeffer SR, Clary DO, Wattenberg BW, Glick BS, Rothman JE. Yeast and mammals utilize similar cytosolic components to
drive protein transport through the Golgi complex. Proc Natl Acad Sci
USA 1986;83:1622–1626.
Claude A. Growth and differentiation of cytoplasmic membranes in
the course of lipoprotein granule synthesis in the hepatic cell. I. Elaboration of elements of the Golgi complex. J Cell Biol 1970;47:745–
766.
Morre DJ, Keenan TW, Mollenhauer HH. Golgi apparatus function in
membrane transformations and product compartmentalization:
Studies with cell fractions isolated from rat liver. Adv Cytopharmacol
1971;1:159–182.
Thiéry G, Rambourg A. A new staining technique for studying thick
sections in the electron microscope. J Microsc Biol Cell 1976;26:103–
106.
Griffiths G. Gut thoughts on the Golgi complex. Traffic 2000;1:738–
745.
Blum R, Pfeiffer F, Feick P, Nastainczyk W, Kohler B, Schafer KH,
Schulz I. Intracellular localization and in vivo trafficking of p24A and
p23. J Cell Sci 1999;112:537–548.
Marra P, Maffucci T, Daniele T, Tullio GD, Ikehara Y, Chan EK, Luini A,
Beznoussenko G, Mironov A, De Matteis MA. The GM130 and
GRASP65 Golgi proteins cycle through and define a subdomain of
the intermediate compartment. Nat Cell Biol 2001;3:1101–1113.
Ladinsky MS, Mastronarde DN, McIntosh JR, Howell KE, Staehelin
LA. Golgi structure in three dimensions: functional insights from the
normal rat kidney cell. J Cell Biol 1999;144:1135–1149.
Saraste J, Kuismanen E. Pre- and post-Golgi vacuoles operate in the
transport of Semliki Forest virus membrane glycoproteins to the cell
surface. Cell 1984;38:535–549.
Schweizer A, Fransen JA, Matter K, Kreis TE, Ginsel L, Hauri HP. Identification of an intermediate compartment involved in protein transport
from endoplasmic reticulum to Golgi apparatus. Eur J Cell Biol
1990;53:185–196.
Saraste J, Svensson K. Distribution of the intermediate elements operating in ER to Golgi transport. J Cell Sci 1991;100:415–430.
Traffic 2002; 3: 521–529
The Golgi Apparatus: Balancing New with Old
61. Scales SJ, Pepperkok R, Kreis TE. Visualization of ER-to-Golgi transport in living cells reveals a sequential mode of action for COPII and
COPI. Cell 1997;90:1137–1148.
62. Presley JF, Cole NB, Schroer TA, Hirschberg K, Zaal KJ, LippincottSchwartz J. ER-to-Golgi transport visualized in living cells. Nature
1997;389:81–85.
63. Shima DT, Scales SJ, Kreis TE, Pepperkok R. Segregation of COPI-rich
and anterograde-cargo-rich domains in endoplasmic-reticulum-toGolgi transport complexes. Curr Biol 1999;9:821–824.
64. Martinez-Menarguez JA, Geuze HJ, Slot JW, Klumperman J. Vesicular
tubular clusters between the ER and Golgi mediate concentration of
soluble secretory proteins by exclusion from COPI- coated vesicles.
Cell 1999;98:81–90.
65. Robbins E, Gonatas N. Histochemical and structural studies on HeLa
cell cultures exposed to spindle inhibitors with special reference to
the interphase cell. J Histochem Cytochem 1964;12:704–711.
66. Storrie B, White J, Rottger S, Stelzer EH, Suganuma T, Nilsson T. Recycling of Golgi-resident glycosyltransferases through the ER reveals
a novel pathway and provides an explanation for nocodazole-induced
Golgi scattering. J Cell Biol 1998;143:1505–1521.
67. Cole NB, Ellenberg J, Song J, DiEuliis D, Lippincott-Schwartz J. Retrograde transport of Golgi-localized proteins to the ER. J Cell Biol
1998;140:1–15.
68. Rossanese OW, Reinke CA, Bevis BJ, Hammond AT, Sears IB, O’Connor J, Glick BS. A role for actin, Cdc1p, and Myo2p in the inheritance
of late Golgi elements in Saccharomyces cerevisiae. J Cell Biol
2001;153:47–62.
69. Hammond AT, Glick BS. Dynamics of transitional endoplasmic reticulum sites in vertebrate cells. Mol Biol Cell 2000;11:3013–3030.
70. Sjöstrand FS, Hanzon V. Ultrastructure of Golgi apparatus of exocrine
cells of mouse pancreas. Exp Cell Res 1954;7:415–429.
71. Hendricks LC, McClanahan SL, McCaffery M, Palade GE, Farquhar
MG. Golgi proteins persist in the tubulovesicular remnants found in
brefeldin A-treated pancreatic acinar cells. Eur J Cell Biol
1992;58:202–213.
72. De Lemos-Chiarandini C, Ivessa NE, Black VH, Tsao YS, Gumper I,
Kreibich G. A Golgi-related structure remains after the brefeldin Ainduced formation of an ER-Golgi hybrid compartment. Eur J Cell Biol
1992;58:187–201.
73. Tooze J, Hollinshead M. Evidence that globular Golgi clusters in mitotic HeLa cells are clustered tubular endosomes. Eur J Cell Biol
1992;58:228–242.
74. Nakamura N, Rabouille C, Watson R, Nilsson T, Hui N, Slusarewicz P,
Kreis TE, Warren G. Characterization of a cis-Golgi matrix protein,
GM130. J Cell Biol 1995;131:1715– 26.
75. Seemann J, Jokitalo E, Pypaert M, Warren G. Matrix proteins can generate the higher order architecture of the Golgi apparatus. Nature
2000;407:1022–1026.
76. Miles S, McManus H, Forsten KE, Storrie B. Evidence that the entire
Traffic 2002; 3: 521–529
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
Golgi apparatus cycles in interphase HeLa cells: Sensitivity of Golgi
matrix proteins to an ER exit block. J Cell Biol 2001;155:543–555.
Ward TH, Polishchuk RS, Caplan S, Hirschberg K, Lippincott-Schwartz
J. Maintenance of Golgi structure and function depends on the integrity of ER export. J Cell Biol 2001;155:557–570.
Lavoie C, Paiement J, Dominguez M, Roy L, Dahan S, Gushue JN,
Bergeron JJ. Roles for a2 p24 and COPI in endoplasmic reticulum
cargo exit site formation. J Cell Biol 1999;146:285–299.
Aridor M, Fish KN, Bannykh S, Weissman J, Roberts TH, LippincottSchwartz J, Balch WE. The Sar1 GTPase coordinates biosynthetic
cargo selection with endoplasmic reticulum export site assembly. J
Cell Biol 2001;152:213–229.
Pelletier L, Jokitalo E, Warren G. The effect of Golgi depletion on exocytic transport. Nat Cell Biol 2000;2:840–846.
Lucocq JM, Berger EG, Warren G. Mitotic Golgi fragments in HeLa
cells and their role in the reassembly pathway. J Cell Biol
1989;109:463–474.
Shima DT, Cabrera-Poch N, Pepperkok R, Warren G. An ordered inheritance strategy for the Golgi apparatus: Visualization of mitotic disassembly reveals a role for the mitotic spindle. J Cell Biol
1998;141:955–966.
Zaal KJ, Smith CL, Polishchuk RS, Altan N, Cole NB, Ellenberg J,
Hirschberg K, Presley JF, Roberts TH, Siggia E, Phair RD, LippincottSchwartz J. Golgi membranes are absorbed into and reemerge from
the ER during mitosis. Cell 1999;99:589–601.
Jokitalo E, Cabrera-Poch N, Warren G, Shima DT. Golgi clusters and
vesicles mediate mitotic inheritance independently of the endoplasmic reticulum. J Cell Biol 2001;154:317–330.
Rossanese OW, Glick BS. Deconstructing Golgi inheritance. Traffic
2001;2:589–596.
Hammond C, Braakman I, Helenius A. Role of N-linked oligosaccharide recognition, glucose trimming, and calnexin in glycoprotein folding
and quality control. Proc Natl Acad Sci USA 1994;91:913–917.
Helenius A, Aebi M. Intracellular functions of N-linked glycans.
Science 2001;291:2364–2369.
Girod A, Storrie B, Simpson JC, Johannes L, Goud B, Roberts LM,
Lord JM, Nilsson T, Pepperkok R. Evidence for a COP-I-independent
transport route from the Golgi complex to the endoplasmic reticulum.
Nat Cell Biol 1999;1:423–430.
White J, Johannes L, Mallard F, Girod A, Grill S, Reinsch S, Keller P,
Tzschaschel B, Echard A, Goud B, Stelzer EH. Rab6 coordinates a
novel Golgi to ER retrograde transport pathway in live cells. J Cell Biol
1999;147:743–760.
Short B, Preisinger C, Korner R, Kopajtich R, Byron O, Barr FA. A
GRASP55-rab2 effector complex linking Golgi structure to membrane
traffic. J Cell Biol 2001;155:877–883.
Surrey T, Nedelec F, Leibler S, Karsenti E. Physical properties determining self-organization of motors and microtubules. Science
2001;292:1167–1171.
529