FEBS Letters 583 (2009) 3764–3769
journal homepage: www.FEBSLetters.org
Review
Sorting out glycosylation enzymes in the Golgi apparatus
Tommy Nilsson *, Catherine E. Au, John J.M. Bergeron *
The Research Institute of the McGill University Health Centre, Department of Medicine, McGill University, 687 Pine Avenue West, Montreal, Quebec, Canada H3A 1A1
a r t i c l e
i n f o
Article history:
Received 15 October 2009
Accepted 20 October 2009
Available online 28 October 2009
Edited by Alberto Luini
Keywords:
Golgi apparatus
Glycosylation
COPI
Protein transport
Protein sorting
a b s t r a c t
The study of glycosylation and glycosylation enzymes has been instrumental for the advancement of
Cell Biology. After Neutra and Leblond showed that the Golgi apparatus is the main site of glycosylation, elucidation of oligosaccharide structures by Baenziger and Kornfeld and subsequent mapping
of glycosylation enzymes followed. This enabled development of an in vitro transport assay by Rothman and co-workers using glycosylation to monitor intra Golgi transport which, complemented by
yeast genetics by Schekman and co-workers, provided much of the fundamental insights and key
components of the secretory pathway that we today take for granted. Glycobiology continues to play
a key role in Cell Biology and here, we look at the use of glycosylation enzymes to elucidate intra
Golgi transport.
Ó 2009 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
1. The Golgi apparatus—a mixed bag
Some 60 years ago, Gersh proposed ‘‘that it may be possible to
conceive of the Golgi apparatus as a framework whose structure
is of such nature that it may accommodate certain enzymes or
other activities in an orderly manner” [1]. This set the stage for a
‘‘mechanized” view of Golgi structure and function such that the
cisternae of the 100–150 Golgi stacks that make up a typical mammalian Golgi apparatus were in the 70s and 80s viewed as having
distinct sets of enzymes that, ‘‘in an orderly manner” act on biosynthetic cargo as this is transported from cisterna to cisterna via COPI
vesicles (see Fig. 1, vesicular transport and [2]). Subcellular fractionation and ultrastructural localization studies supported this
view. Using differences in membrane densities, Dunphy and Rothman showed that early enzymatic events of N-linked oligosaccharide processing could be separated from later ones in a cis to trans
fashion [3,4]. At the same time, antibodies were used to localize
particular glycosylation enzymes to distinct cisternae [5–8]. In
the 1990s, this compartmental model was challenged by the observations that N-acetylglucosaminyltransferase 1 (GlcNAc-T1), galactosyltransferase I (GT1), sialyltransferase I (ST1), mannosidase II
(Mann II), N-acetylgalactosaminyltransferase (GalNAc-T) 1–3 show
gradient-like distributions across the Golgi stack upon immunogold labeling of thin frozen sections followed by electron
microscopy and quantitation [9–11]. Such gradients also showed
significant overlap such that reactions that had previously been
* Corresponding authors.
E-mail addresses: [email protected] (T. Nilsson), john.bergeron@mcgill.
ca (J.J.M. Bergeron).
assumed as separate could now take place in the same cisterna.
For example, GT1 and ST1, believed to occupy the trans cisterna
and the trans Golgi network, respectively [12], were now shown
co-localized to the trans cisternae with no discernable difference
in distributions [10]. Moreover, GlcNAc-T1 and Mann II both overlapped with GT1 and ST1 enabling all four enzymes to act in the
same cisterna [9,10]. That indeed ‘‘medial” and ‘‘trans/trans Golgi
network” specific glycosylation events may occur without any
need for intra Golgi transport has been confirmed, in vitro [13].
As such, step-wise assembly of N-linked oligosaccharides is not
governed by physical compartmentalization [2,3,12,14–17] but
rather, an outcome of multiple enzymes that often occupy the
same membrane-bound compartment which may vary from celltype to cell-type [18]. A direct consequence of co-habiting the
same compartment is that different enzymes directed to the same
substrate will compete. A typical gradient-like distribution of
EGFP-tagged GalNAc-T2 over single Golgi stacks revealed by the
microscopy method, GRAB [19] is shown in Fig. 2.
2. Glycosylation enzyme specificity in vivo—moving to the ‘‘far
side’’ of biochemistry
Given that multiple enzymes compete in the same compartment (at present, some 342 enzymes have been classified in humans [20]), how is specificity then promoted such that certain
oligosaccharide structures and linkages are favored over others?
For polypeptide GalNAc-Ts, work by Hazes, Clausen and others
have shown that GalNAc-Ts have lectin-like domains that recognize previously added sugars in the context of the polypeptide aiding in subsequent glycosylation events [21,22]. This, however,
0014-5793/$36.00 Ó 2009 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.febslet.2009.10.064
T. Nilsson et al. / FEBS Letters 583 (2009) 3764–3769
3765
Fig. 1. Cisternal maturation versus vesicular transport. On the left, vesicular tubular clusters derived from the ER containing biosynthetic cargo proteins differentiate to
become new cisternae that gradually mature as cargo is processed through glycosylation enzymes and other enzymes. These resident Golgi enzymes are transported in a
counter current (retrograde) via COPI vesicles. On the right, COPI vesicles transport biosynthetic cargo from stationary cisternae allowing their cargo to be processed by Golgi
cisternal resident enzymes. The two models differ greatly and predicts predominantly biosynthetic cargo in COPI vesicles in the vesicular transport model and predominantly
Golgi-resident cargo proteins including glycosylation enzymes in the cisternal maturation model. Overwhelming in vitro data through quantitative proteomics supports the
cisternal maturation model.
Fig. 2. Gradient distribution of GalNAc-T2 fused to EGFP. On the left, parts of the Golgi ribbon showing individual Golgi stacks linked together laterally. The intra Golgi
distribution of GalNAc-T2-EGFP was revealed by the quantitative and correlative GRAB method. On the right, following the release of a drug, propranolol to inhibit the
synthesis of diacylglycerol, GalNAc-T2-EGFP is seen in both vesicle buds (arrow) and vesicles (arrowhead). From Asp et al. (2009).
seems more an exception than a rule as other enzyme families do
not appear to have lectin-domains. Instead, most glycosylation enzymes catalyze simple donor–acceptor reactions through inversion
or retention of stereochemistry of the sugar donor with specificity
relying mostly on previous linkages and structures (with the
exception of polypeptide O-glycosylation). In some cases, more
than one enzyme can use the same acceptor to make the same linkage (e.g. some fucosyl and sialyl transferases). Glycosylation is
therefore believed to be a consequence of classical Michaelis–Menten enzyme kinetics and an assumption that all components diffuse normally. Given the combinatorial possibilities of
glycosylation (1012–1015 is often quoted) in relation to the number
of glycosylation enzymes (a couple of hundred enzymes), one is
tempted to conclude that other factors must come into play that
ensures specificity when oligosaccharide structures are assembled
in cisternal lumens of Golgi cisternae. It is indeed difficult to predict which oligosaccharide structures will be produced based on
what enzymes are expressed in a given cell-type based on their
Km. This might not be so surprising since most enzymatic Kms
have been determined in vitro assuming that components diffuse
freely and that ‘‘the law of mass action” applies. The inside of the
cell, however, is very different. Intracellular macromolecules above
a certain size do not exhibit normal diffusion behavior. This is due
to the crowded environment inside the cell, not so much by membrane structures or cytoskeletal elements but rather, by the unique
‘‘cocktail” of other proteins that together, enforce an anomalous
diffusional behavior on the observed molecule in the cytoplasm
[23] or intracellular membranes [24]. For glycosylation enzymes,
the deviation from normal diffusion can be as much as a = 0.55 (expressed as ta and observed for GT1) [24]. This offsets any in vitrodetermined enzyme kinetics by far. In layman terms, there is really
no need for ‘‘high” affinities between acceptor structures and glycosylation enzymes as in a crowded environment, mM affinities
or lower will suffice [25]. Such a ‘‘low” affinity–‘‘high specificity”
argument may seem counter intuitive to most biochemists but
must not be ignored as anomalous diffusion is clearly observable
in the living cell. In our view, the ‘‘low” affinity–‘‘high specificity”
scenario brought about by crowded environments is at the heart
of self-assembling molecular machineries allowing these to form
in a dynamic fashion. The hypothesis promoted here is that glycosylation enzymes self-assemble into molecular machineries, enzyme complexes, that are held together by weak to strong
3766
T. Nilsson et al. / FEBS Letters 583 (2009) 3764–3769
protein–protein interactions and that it is this that promotes specificity and enzymatic ‘‘efficiency” in the production of oligosaccharide structures.
3. Sorting of glycosylation enzymes in the Golgi
Self-assembled enzyme complexes or ‘‘kin” complexes were
postulated already in the early 1990s [26] as a mechanism for correct positioning of glycosylation enzymes in the Golgi apparatus
and was demonstrated between the medial/trans enzymes GlcNAc-T1 and Mann II [27]. Many glycosylation enzymes also have
the intrinsic ability to self-oligomerize such that when isolated
from cells, these exist as higher oligomeric structures. When shed
from the cell into the circulating blood, however, glycosylation enzymes appear as monomers or homodimers. Much debate has surrounded the possibility that glycosylation enzymes form ‘‘kin”
complexes but as noted above, molecular machineries do not require strong interactions between its components, in vivo. To assess whether kin complexes exist or the extent whereby these
exist, therefore, one should keep the reality of molecular crowding
enabling ‘‘low affinity”–‘‘high specificity” in mind. It should in fact
be possible to mimic crowded environments in vitro so that enzyme kinetics and complex formation can be measured using, for
example, complex dextran solutions [23]. For in vivo, complex formation could be revealed and measured through fluorescent crosscorrelation spectroscopy using stably expressed fluorescently
tagged glycosylation enzymes with distinct excitation-emission
wavelengths.
A second and complementary model for enzyme positioning
across the Golgi stack, ‘‘the membrane thickness model” postulates
that the lengths of transmembrane domains of glycosyltransferases
determine their position in the pathway as cisternal membranes
would, due to a gradual increase in cholesterol and sphingolipids, increase in thickness in the cis to trans direction. Glycosyltransferases
would here be prevented from further transport (i.e. prevented from
entering COPI vesicles involved in forward anterograde transport).
Indeed, Golgi localized glycosylation enzymes may, on average, have
shorter transmembrane domains as do ER proteins compared to
plasma membrane proteins [28]. A modified version of this model
postulates that in addition, the composition of the transmembrane
domain is important [29]. For example, aromatic amino acids are
prevalent in membrane spanning domains of Golgi-resident proteins. According to the ‘‘kin recognition model” mentioned above,
such amino acids would facilitate or indeed, mediate, protein–protein interactions aiding in complex formation. A third model postulates that glycosylation enzymes compete for the entry into
retrograde COPI vesicles [30]. This alone suffices to explain asymmetric and overlapping enzyme distributions and is supported by
the presence of glycosylation enzymes in COPI vesicles, formed
in vitro [31–33]. That model is very attractive in its simplicity but
lacks robustness [34]. A trigger for sorting was therefore proposed
based on the notion that the cisternal milieu changes in a cis to trans
direction such that the pH decreases, the lipid composition changes
and the available substrates decrease as these become modified.
When assuming a ‘‘trigger” for sorting, the competition model now
becomes robust and compatible with both the kin recognition model
as well as the membrane thickness model. Given the high degree of
conservation between species, it is likely that glycosylation enzymes
have evolved such that their structure, be it linear or tertiary, have
adapted to the milieu in which these operate. This would include
their transmembrane lengths, transmembrane composition as well
as other parts of the enzyme (e.g. stalk regions). A ‘‘competition”
model which incorporates a trigger for sorting into recycling COPI
vesicles due to changes in the cisternal milieu is attractive as it
allows for active sorting through the COPI coatomer protein complex
as this cycles on and off the membrane in an ARF/ARFGAP GTPdependent manner [34,35].
A direct interaction between coatomer and cytoplasmic domains of glycosyltransferases is difficult to envisage due to their
seemingly short domains though these may vary from just a few
amino acids up to 80. These mostly contain positive amino acids
such as R and K at their cytoplasmic N-terminal but are likely there
to ensure type II topology of glycosylation enzymes [36]. Recent
evidence in yeast, however, shows that Vps74p has a direct interaction with many glycosyltransferases (mannosyltransferases)
through their N-terminal cytoplasmic domains [37,38]. In mammalian cells, the human ortholog gpp34 (also termed GMx33 or
GOLPH3) and a gpp34 similar protein are yet to be shown to mediate such interactions. Although gpp34complements Vps74p in
yeast, the N-terminal domains of mammalian glycosylation enzymes bear little if any resemblance to those of yeast glycosylation
enzymes. Also, neither gpp34 nor a gpp34 similar protein appears
particularly abundant compared to resident Golgi proteins (about a
100-fold less) based on quantitative proteomics analysis of purified
Golgi membranes [33]. The caveat of this argument is that peripheral proteins may fall of the membranes when subjected to sucrose-gradient centrifugation and that observed ratios may not,
therefore, be reflective of in vivo conditions. Nevertheless, we think
it is unlikely that Vps74p and its mammalian orthologs have a general function in glycosylation enzyme positioning and perhaps
should be viewed as auxiliary Golgi proteins [39] such as the
COG proteins that when perturbed, have consequences on the
localization of Golgi-resident proteins including glycosyltransferases [40].
A model for ‘‘triggered sorting” which incorporates ‘‘kin recognition” and ‘‘membrane exclusion” should be viewed in the context
of intra Golgi transport (Fig. 1). Held as a paradigm for intra Golgi
transport during the 80s and the 90s, the vesicular transport model
was attractive as biosynthetic cargo would be transported from
cisterna to cisterna via COPI vesicles and modified in an orderly
manner by glycosylation enzymes that were strictly compartmentalized (see above). This model replaced the earlier cisternal maturation/progression model [41,42] that was based on the
observation that the Golgi apparatus exhibits the structural characteristics of an organelle in transit, that is, assembled at the cis
side and taken apart at the trans-side. When viewed in 3D, the
structural characteristics of an organelle in transit is even more
telling [43,44] showing extensive tubular–vesicular networks on
both the cis- and the trans-side as well as between the more compact Golgi stacks. As the cisternal maturation/progression model
was proposed before the characterization of retrograde transport
carriers such as COPI vesicles, the model assumed that Golgi stacks
required a constant input of biosynthetic material. Protein synthesis was shown not to be required for Golgi stack morphology (at
least on a time-scale relevant to protein transport) [45], the model
of cisternal maturation was discounted, this despite plain evidence
of large biosynthetic macromolecular structures forming in the lumen of Golgi cisternae maturing in a cis to trans direction [46–48],
too large to fit into 60 nm diameter COPI transport vesicles. The
discovery that the main coat component of COPI vesicles, coatomer, interacts with the ER retention/retrieval K(X)KXX [49,50] found
on many resident proteins of the ER (e.g. glucoronosyltransferases
of the liver) and the ER to Golgi interface (e.g. the p24 family of
abundant small type I transmembrane proteins) placed COPI firmly
on the retrograde transport route(s). As such, COPI vesicles were
now viewed as carriers of resident proteins playing the lead role
in the distillation hypothesis proposed by Rothman where biosynthetic cargo is ‘‘distilled” away from resident proteins as it traverses the secretory pathway [51]. This evidently happens in the
early parts of the secretory pathway but if this applies also within
the Golgi stack is still debated and the two main models, the vesic-
T. Nilsson et al. / FEBS Letters 583 (2009) 3764–3769
ular transport model and the cisternal maturation model are therefore still used to describe intra Golgi transport [52]. Modifications
of the two models have been proposed where tubular structures
connecting cisternae of the same stack [53] may form upon high
transport activity as observed in insulin-secreting cells [54] or
through a lipid phase-driven membrane partitioning of resident
proteins and biosynthetic cargo allowing for faster than normal
transport through the secretory pathway [55]. Such lipid phasedriven partitioning implies that resident proteins and biosynthetic
membrane proteins prefer different lipid environments and therefore, that lipids would drive such partitioning. Given that most
phospholipids generally diffuse some two orders of magnitude faster than membrane proteins, lipid driven phase separation seems
unlikely as a driving force for protein separation. Rather, in a
crowded membrane such as the Golgi cisternae, we suggest that
the membrane thickness and lipid domains of a particular membrane system are direct consequence of the membrane proteins
that occupy it. In other words, the membrane proteins given their
density, orders the lipids rather than the other way around. Also, at
any given time, the ratio between resident proteins and biosynthetic membrane proteins is such there is a 13-fold excess of resident transmembrane proteins over biosynthetic ones [33]. A real
3767
phase separation such that entire or parts of cisternae breaks off
seems unlikely, at least not in hepatic Golgi of the rat [33].
So how crowded are the cisternae of the Golgi apparatus? It is
estimated that the densities of integral membrane proteins in the
Golgi and the plasma membrane are 38 000 and 30 000 proteins
per square micron, respectively, in baby hamster kidney cells
[56]. In red blood cells and synaptic vesicles, transmembrane domains make up some 23% of the membrane corresponding to an
average density of 130 000 transmembrane regions per square micron [57,58]. If assuming that each transmembrane domain is surrounded by a ring of phospholipids, the tightest packing ratio
between lipids and transmembrane is 4:1 according to the space
taken up by a phospholipid versus that of a helical transmembrane
domain. A slightly lower ratio, 3.5:1, was observed in synaptic vesicles [58] suggesting that some transmembrane domains are not
fully surrounded by phospholipids. This would be expected if a
portion of transmembrane domains belongs to type III membrane
proteins (i.e. membrane proteins with two or more transmembrane domains). In Golgi membranes, having a density of 38 000
membrane proteins per square micron, there is by comparison,
ample space to enable movement of biosynthetic membrane proteins and glycosylation enzymes in the plane of cisternal mem-
Fig. 3. Cisternal maturation versus vesicular transport. Quantitative proteomics of selected proteins enriched in Golgi fractions or COPI vesicles. The ordinate in all cases
indicates the proportion of tandem mass spectra (redundant peptides) assigned to the proteins. The secretory cargo haptoglobin, the resident protein p115 and the secretory
protein albumin are Golgi enriched whereas the resident membrane proteins GalNacT2and CASP are concentrated in isolated COP I vesicle fractions as is the intraluminal
resident soluble protein nucleobindin 1 (CalNuc). Means +/ S.D. for n = 3 biological repeats are shown. Data taken from the Gilchrist et al. resource.
3768
T. Nilsson et al. / FEBS Letters 583 (2009) 3764–3769
branes [56]. A previous claim by Pelham and Rothman [52] stated
that there is only room for a 10% increase in protein density in Golgi membranes and hence, little if any capacity of COPI vesicles to
sort and transport proteins such as Golgi glycosylation enzymes.
In light of the observed protein density of synaptic vesicles, there
is theoretically room for at least a 3-fold concentration of membrane proteins in Golgi cisternae making it possible for COPI vesicles to engage in protein sorting through concentration of Golgiresident proteins [31–33].
In our proteomics study of rat hepatic Golgi membranes and
in vitro generated COPI vesicles, we observed a 2-fold concentration of resident membrane proteins over cisternal proteins. To
achieve accumulation, generated vesicles had to be prevented from
back fusion using a dominant negative mutant of the SNARE binding protein a-SNAP. Formed vesicles were readily sedimented and
when analyzed by quantitative mass spectrometry, found to contain 29.5% type I and II transmembrane proteins (i.e. those that
span the membrane once including glycosylation enzymes and
SNARE proteins) and 7.5% type III transmembrane proteins (those
that span the membrane at least twice including sugar transporters, KDEL receptor, multi spanning domain proteins of unknown
function). In Golgi cisternae, corresponding numbers were 14.5%
for type I and II and 4.1% for type III transmembrane proteins. In
contrast, biosynthetic transmembrane proteins in COPI vesicles
were 0.4% and 0.05% for type I/II and III, respectively. In Golgi
membranes, this was 1.1% and 0.3% type I/II and III, respectively.
In other words, we observed a 3-fold exclusion of biosynthetic
transmembrane proteins (type I/II/III) from entering budding COPI
vesicles. In terms of soluble proteins, we observed a similar pattern, resident calcium-binding proteins of the Golgi were concentrated about 4-fold whereas biosynthetic cargo was excluded by
a factor of 3.5. Taken together, COPI vesicles formed in vitro show
a strong preference for resident Golgi proteins. This includes proteins predicted to occupy the medial and trans parts of the Golgi
stack suggesting a role for COPI vesicles in the recycling of Golgiresident glycosylation enzymes from all parts of the Golgi stack.
Whether or not this holds true also in vivo remains to be seen.
Examples of different proteins preference for COPI or Golgi cisternae in vitro are shown in Fig. 3.
4. A few concluding remarks
The Golgi field is as dynamic as the cellular structure itself and
with the advances of technologies such as live cell imaging at higher and higher resolutions, many of the remaining questions may be
resolved. Attaining images of glycosylation enzymes in COPI vesicles and COPI vesicle buds as shown in Fig. 2 will for sure help to
settle the debate surrounding the role of these carriers in the Golgi
apparatus. Likewise, complex formation of glycosylation enzymes
should be probed through in vivo imaging techniques that detect
protein-protein interaction on a single molecule detection level
without requiring high levels of expression. This is possible
through fluorescent cross-correlation spectroscopy. Also, sub populations of COPI vesicles are predicted in a cisternal maturation
scenario. Isolation and characterization of these through quantitative proteomics is therefore a priority. We apologize to those that
have not been cited or those whose opinions have not been presented here. This review is written from the perspective that attention should be paid to literature that has had long-term or
profound effect on the field of cell biology. As such, we tried to cite
original and key references.
References
[1] Gersh, I. (1949) A protein component of the Golgi apparatus. Arch. Pathol.
(Chic.) 47, 99–109.
[2] Rothman, J.E. and Orci, L. (1990) Movement of proteins through the Golgi
stack: a molecular dissection of vesicular transport. FASEB J. 4, 1460–1468.
[3] Dunphy, W.G., Fries, E., Urbani, L.J. and Rothman, J.E. (1981) Early and late
functions associated with the Golgi apparatus reside in distinct compartments.
Proc. Natl. Acad. Sci. USA 78, 7453–7457.
[4] Dunphy, W.G. and Rothman, J.E. (1983) Compartmentation of asparaginelinked oligosaccharide processing in the Golgi apparatus. J. Cell Biol. 97, 270–
275.
[5] Roth, J. and Berger, E.G. (1982) Immunocytochemical localization of
galactosyltransferase in HeLa cells: codistribution with thiamine
pyrophosphatase in trans-Golgi cisternae. J. Cell Biol. 93, 223–229.
[6] Roth, J., Taatjes, D.J., Lucocq, J.M., Weinstein, J. and Paulson, J.C. (1985)
Demonstration of an extensive trans-tubular network continuous with the
Golgi apparatus stack that may function in glycosylation. Cell 43, 287–295.
pii:S0092-8674(85)90034-0.
[7] Dunphy, W.G., Brands, R. and Rothman, J.E. (1985) Attachment of terminal Nacetylglucosamine to asparagine-linked oligosaccharides occurs in central
cisternae of the Golgi stack. Cell 40, 463–472. pii:S0092-8674(85)90161-8.
[8] Velasco, A., Hendricks, L., Moremen, K.W., Tulsiani, D.R., Touster, O. and
Farquhar, M.G. (1993) Cell type-dependent variations in the subcellular
distribution of alpha-mannosidase I and II. J. Cell Biol. 122, 39–51.
[9] Nilsson, T., Pypaert, M., Hoe, M.H., Slusarewicz, P., Berger, E.G. and Warren, G.
(1993) Overlapping distribution of two glycosyltransferases in the Golgi
apparatus of HeLa cells. J. Cell Biol. 120, 5–13.
[10] Rabouille, C., Hui, N., Hunte, F., Kieckbusch, R., Berger, E.G., Warren, G. and
Nilsson, T. (1995) Mapping the distribution of Golgi enzymes involved in
the construction of complex oligosaccharides. J. Cell Sci. 108 (Pt 4), 1617–
1627.
[11] Rottger, S., White, J., Wandall, H.H., Olivo, J.C., Stark, A., Bennett, E.P.,
Whitehouse, C., Berger, E.G., Clausen, H. and Nilsson, T. (1998) Localization
of three human polypeptide GalNAc-transferases in HeLa cells suggests
initiation of O-linked glycosylation throughout the Golgi apparatus. J. Cell
Sci. 111 (Pt 1), 45–60.
[12] Berger, E.G. and Hesford, F.J. (1985) Localization of galactosyl- and
sialyltransferase by immunofluorescence: evidence for different sites. Proc.
Natl. Acad. Sci. USA 82, 4736–4739.
[13] Hayes, B.K., Freeze, H.H. and Varki, A. (1993) Biosynthesis of oligosaccharides
in intact Golgi preparations from rat liver. Analysis of N-linked glycans labeled
by UDP-[6-3H]N-acetylglucosamine. J. Biol. Chem. 268, 16139–16154.
[14] Goldberg, D.E. and Kornfeld, S. (1983) Evidence for extensive subcellular
organization of asparagine-linked oligosaccharide processing and lysosomal
enzyme phosphorylation. J. Biol. Chem. 258, 3159–3165.
[15] Griffiths, G. and Simons, K. (1986) The trans Golgi network: sorting at the exit
site of the Golgi complex. Science 234, 438–443.
[16] Chege, N.W. and Pfeffer, S.R. (1990) Compartmentation of the Golgi complex:
brefeldin-A distinguishes trans-Golgi cisternae from the trans-Golgi network.
J. Cell Biol. 111, 893–899.
[17] Green, S.A. and Kelly, R.B. (1990) Endocytic membrane traffic to the Golgi
apparatus in a regulated secretory cell line. J. Biol. Chem. 265, 21269–21278.
[18] Roth, J., Taatjes, D.J., Weinstein, J., Paulson, J.C., Greenwell, P. and Watkins,
W.M. (1986) Differential subcompartmentation of terminal glycosylation in
the Golgi apparatus of intestinal absorptive and goblet cells. J. Biol. Chem. 261,
14307–14312.
[19] Grabenbauer, M., Geerts, W.J., Fernadez-Rodriguez, J., Hoenger, A., Koster, A.J.
and Nilsson, T. (2005) Correlative microscopy and electron tomography of GFP
through photooxidation. Nat. Methods 2, 857–862, doi:10.1038/nmeth806.
pii:nmeth806.
[20] Cantarel, B.L., Coutinho, P.M., Rancurel, C., Bernard, T., Lombard, V. and
Henrissat, B. (2009) The Carbohydrate-Active EnZymes database (CAZy): an
expert resource for glycogenomics. Nucl. Acids Res. 37, D233–D238,
doi:10.1093/nar/gkn663. pii:gkn663.
[21] Hazes, B. (1996) The (QxW)3 domain: a flexible lectin scaffold. Protein Sci. 5,
1490–1501, doi:10.1002/pro.5560050805.
[22] Wandall, H.H., Irazoqui, F., Tarp, M.A., Bennett, E.P., Mandel, U., Takeuchi, H.,
Kato, K., Irimura, T., Suryanarayanan, G., Hollingsworth, M.A., et al. (2007) The
lectin domains of polypeptide GalNAc-transferases exhibit carbohydratebinding specificity for GalNAc: lectin binding to GalNAc-glycopeptide
substrates is required for high density GalNAc-O-glycosylation. Glycobiology
17, 374–387, doi:10.1093/glycob/cwl082. pii:cwl082.
[23] Weiss, M., Elsner, M., Kartberg, F. and Nilsson, T. (2004) Anomalous subdiffusion
is a measure for cytoplasmic crowding in living cells. Biophys. J. 87, 3518–3524,
doi:10.1529/biophysj.104.044263. pii:S0006-3495(04)73816-3.
[24] Weiss, M., Hashimoto, H. and Nilsson, T. (2003) Anomalous protein diffusion
in living cells as seen by fluorescence correlation spectroscopy. Biophys. J. 84,
4043–4052, doi:10.1016/S0006-3495(03)75130-3. pii:S0006-3495(03)751303.
[25] Weiss, M. and Nilsson, T. (2004) In a mirror dimly: tracing the movements of
molecules in living cells. Trends Cell Biol. 14, 267–273, doi:10.1016/
j.tcb.2004.03.012. pii:S0962-8924(04)00086-8.
[26] Nilsson, T., Lucocq, J.M., Mackay, D. and Warren, G. (1991) The membrane
spanning domain of beta-1, 4-galactosyltransferase specifies trans Golgi
localization. Embo J. 10, 3567–3575.
[27] Nilsson, T., Hoe, M.H., Slusarewicz, P., Rabouille, C., Watson, R., Hunte, F.,
Watzele, G., Berger, E.G. and Warren, G. (1994) Kin recognition between
medial Golgi enzymes in HeLa cells. Embo J. 13, 562–574.
T. Nilsson et al. / FEBS Letters 583 (2009) 3764–3769
[28] Bretscher, M.S. and Munro, S. (1993) Cholesterol and the Golgi apparatus.
Science 261, 1280–1281.
[29] Munro, S. (1998) Localization of proteins to the Golgi apparatus. Trends Cell
Biol. 8, 11–15.
[30] Glick, B.S., Elston, T. and Oster, G. (1997) A cisternal maturation mechanism
can explain the asymmetry of the Golgi stack. FEBS Lett 414, 177–181.
pii:S0014-5793(97)00984-8.
[31] Lanoix, J., Ouwendijk, J., Lin, C.C., Stark, A., Love, H.D., Ostermann, J. and
Nilsson, T. (1999) GTP hydrolysis by arf-1 mediates sorting and concentration
of Golgi resident enzymes into functional COP I vesicles. Embo J. 18, 4935–
4948.
[32] Lanoix, J., Ouwendijk, J., Stark, A., Szafer, E., Cassel, D., Dejgaard, K., Weiss, M.
and Nilsson, T. (2001) Sorting of Golgi resident proteins into different
subpopulations of COPI vesicles: a role for ArfGAP1. J. Cell Biol. 155, 1199–
1212.
[33] Gilchrist, A., Au, C.E., Hiding, J., Bell, A.W., Fernandez-Rodriguez, J., Lesimple, S.,
Nagaya, H., Roy, L., Gosline, S.J., Hallett, M., et al. (2006) Quantitative
proteomics analysis of the secretory pathway. Cell 127, 1265–1281.
[34] Weiss, M. and Nilsson, T. (2000) Protein sorting in the Golgi apparatus: a
consequence of maturation and triggered sorting. FEBS Lett 486, 2–9.
pii:S0014-5793(00)02155-4.
[35] Weiss, M. and Nilsson, T. (2003) A kinetic proof-reading mechanism for
protein sorting. Traffic 4, 65–73. pii:tra040202.
[36] von Heijne, G. (1992) Membrane protein structure prediction. Hydrophobicity
analysis and the positive-inside rule. J. Mol. Biol. 225, 487–494. pii:S00222836(92)90934-C.
[37] Tu, L., Tai, W.C., Chen, L. and Banfield, D.K. (2008) Signal-mediated dynamic
retention of glycosyltransferases in the Golgi. Science 321, 404–407,
doi:10.1126/science.1159411. pii:321/5887/404.
[38] Schmitz, K.R., Liu, J., Li, S., Setty, T.G., Wood, C.S., Burd, C.G. and Ferguson, K.M.
(2008) Golgi localization of glycosyltransferases requires a Vps74p oligomer.
Dev. Cell 14, 523–534, doi:10.1016/j.devcel.2008.02.016. pii:S15345807(08)00110-X.
[39] Snyder, C.M., Mardones, G.A., Ladinsky, M.S. and Howell, K.E. (2006) GMx33
associates with the trans-Golgi matrix in a dynamic manner and sorts within
tubules exiting the Golgi. Mol. Biol. Cell 17, 511–524, doi:10.1091/mbc.E0507-0682. pii:E05-07-0682.
[40] Shestakova, A., Zolov, S. and Lupashin, V. (2006) COG complex-mediated
recycling of Golgi glycosyltransferases is essential for normal protein
glycosylation. Traffic 7, 191–204, doi:10.1111/j.1600-0854.2005.00376.x.
pii:TRA376.
[41] Grasse, P.P. (1957) Ultrastructure, polarity and reproduction of Golgi
apparatus. C.R. Hebd. Seances Acad. Sci. 245, 1278–1281.
[42] Morre, D.J. and Ovtracht, L. (1977) Dynamics of the Golgi apparatus:
membrane differentiation and membrane flow. Int. Rev. Cytol. Suppl. 61, 188.
[43] Rambourg, A., Clermont, Y. and Hermo, L. (1979) Three-dimensional
architecture of the Golgi apparatus in Sertoli cells of the rat. Am. J. Anat.
154, 455–476, doi:10.1002/aja.1001540402.
3769
[44] Ladinsky, M.S., Mastronarde, D.N., McIntosh, J.R., Howell, K.E. and Staehelin,
L.A. (1999) Golgi structure in three dimensions: functional insights from the
normal rat kidney cell. J. Cell Biol. 144, 1135–1149.
[45] Jamieson, J.D. and Palade, G.E. (1968) Intracellular transport of secretory
proteins in the pancreatic exocrine cell. 3. Dissociation of intracellular
transport from protein synthesis. J. Cell Biol. 39, 580–588.
[46] Weinstock, M. and Leblond, C.P. (1974) Synthesis, migration, and release of
precursor collagen by odontoblasts as visualized by radioautography after
(3H)proline administration. J. Cell Biol. 60, 92–127.
[47] Melkonian, M., Becker, B. and Becker, D. (1991) Scale formation in algae. J.
Electron. Microsc. Technol. 17, 165–178, doi:10.1002/jemt.1060170205.
[48] Bonfanti, L., Mironov, A.A., Jr., Martinez-Menarguez, J.A., Martella, O., Fusella,
A., Baldassarre, M., Buccione, R., Geuze, H.J., Mironov, A.A. and Luini, A. (1998)
Procollagen traverses the Golgi stack without leaving the lumen of cisternae:
evidence for cisternal maturation. Cell 95, 993–1003. pii:S00928674(00)81723-7.
[49] Letourneur, F., Gaynor, E.C., Hennecke, S., Demolliere, C., Duden, R., Emr, S.D.,
Riezman, H. and Cosson, P. (1994) Coatomer is essential for retrieval of
dilysine-tagged proteins to the endoplasmic reticulum. Cell 79, 1199–1207.
[50] Cosson, P. and Letourneur, F. (1994) Coatomer interaction with di-lysine
endoplasmic reticulum retention motifs. Science 263, 1629–1631.
[51] Rothman, J.E. (1981) The golgi apparatus: two organelles in tandem. Science
213, 1212–1219.
[52] Pelham, H.R. and Rothman, J.E. (2000) The debate about transport in the Golgi
– two sides of the same coin? Cell 102, 713–719.
[53] Trucco, A., Polishchuk, R.S., Martella, O., Di Pentima, A., Fusella, A., Di
Giandomenico, D., San Pietro, E., Beznoussenko, G.V., Polishchuk, E.V.,
Baldassarre, M., et al. (2004) Secretory traffic triggers the formation of
tubular continuities across Golgi sub-compartments. Nat. Cell Biol. 6, 1071–
1081, doi:10.1038/ncb1180. pii:ncb1180.
[54] Marsh, B.J., Volkmann, N., McIntosh, J.R. and Howell, K.E. (2004) Direct
continuities between cisternae at different levels of the Golgi complex in
glucose-stimulated mouse islet beta cells. Proc. Natl. Acad. Sci. USA 101,
5565–5570, doi:10.1073/pnas.0401242101. pii:0401242101.
[55] Patterson, G.H., Hirschberg, K., Polishchuk, R.S., Gerlich, D., Phair, R.D. and
Lippincott-Schwartz, J. (2008) Transport through the Golgi apparatus by rapid
partitioning within a two-phase membrane system. Cell 133, 1055–1067,
doi:10.1016/j.cell.2008.04.044. pii:S0092-8674(08)00618-1.
[56] Quinn, P., Griffiths, G. and Warren, G. (1984) Density of newly synthesized
plasma membrane proteins in intracellular membranes II. Biochemical
studies. J. Cell Biol. 98, 2142–2147.
[57] Dupuy, A.D. and Engelman, D.M. (2008) Protein area occupancy at the center
of the red blood cell membrane. Proc. Natl. Acad. Sci. USA 105, 2848–2852,
doi:10.1073/pnas.0712379105. pii:0712379105.
[58] Takamori, S., Holt, M., Stenius, K., Lemke, E.A., Gronborg, M., Riedel, D., Urlaub,
H., Schenck, S., Brugger, B., Ringler, P., et al. (2006) Molecular anatomy of a
trafficking organelle. Cell 127, 831–846.