The Golgi apparatus: insights from lipid biochemistry
RICHARD E. PAGANO
Department of Embryology, Carnegie Institution of Washington, 115 West University Parkway, Baltimore, MD 21210, U.S.A.
Discovered by Camillo Golgi in 1898 [ l , 21, the Golgi
apparatus has long fascinated cytologists because of its
distinct morphology (Fig. 1).This organelle, which is found
in all eukaryotic cells, consists of flat cisternae arranged in
parallel to form the characteristic ‘stacks’ that are interconnected by cisternal or tubular-reticular bridges. In addition, numerous small membranous vesicles are usually found
associated with the organelle. The Golgi apparatus is
currently the focus of much attention among cell and
membrane biologists because of its central role in directing
protein traffic within cells. At least three areas, the cis,
medial, and trans subsections can be seen by electron
microscopy. The cis compartments are often situated closest
to the nucleus, while the trans elements are most distant.
From biochemical studies we know that there are at least
three functionally distinct compartments that operate
sequentially to build the oligosaccharide chains on transported glycoproteins and to ‘sort’proteins destined for transport to other regions of the cell (e.g. the lysosomes, cell
surface or secretory granules). Excellent reviews on various
Sixth Morton Lecture
Delivered on 18 December 1989 at
St Bartholomew’s Hospital Medical College, London
DR RICHARD E. PAGANO
aspects of the Golgi apparatus, including its morphology and
cytochemistry 13-51, biochemical organization 16, 71, and its
roles in glycoprotein synthesis [S- 101 and protein secretion
and sorting I 1 1- 151, are available.
Until now, most biochemical and cell biological studies of
the Golgi apparatus have concentrated on its role in the
synthesis and sorting of proteins. In this paper I will highlight
studies from my own laboratory dealing with some aspects of
the metabolism and transport of lipids through the Golgi
apparatus and indicate how such studies may enhance our
understanding of the function of this organelle.
A vital stain for the Golgi apparatus
Fig. I . Three-dimensional representation
apparatus
of the
Go@
In transverse section the Golgi apparatus is seen as stacks of
interconnected flattened cisternae ( 1). Together these often
encircle the nucleus. Branching tubules ( 2 ) connect Golgi
cisternae (3)creating a complex network. Note the numerous
pores (4) on the inner surface of the Golgi. Reproduced with
permission from R. V. Krstic ( 1979), Ultrastructure of the
Mammalian Cell. A n Atlas, pp. 67, Springer-Verlag, New
York
Vol. 18
For a number of years, my colleagues and I have systematically studied the synthesis and transport of lipid
molecules in animal cells. Our goal has been to define the
pathways and molecular mechanisms by which the many
different lipid species found in cells move from one organelle
to another after their biosynthesis. This is a fundamental
problem in membrane biology which has been difficult to
study because simple techniques for examining lipid ‘traffic’
within cells are not available. We developed a new approach
for studying this problem using a series of fluorescent lipid
analogues in which one of the naturally occurring fatty acids
is replaced by N-(4-nitrobenzo-2-oxa- 1,3-diazoIe)aminocaproic acid ((2,-NBD-fatty acid) [ 16). These lipids exhibit
high rates of spontaneous transfer between membranes in
vitro, and this property permits us to readily integrate the
fluorescent lipids into cellular membranes from exogenous
sources. In our early studies, we found that when the various
analogues were incubated with cells, they were metabolized
similarly to their endogenous counterparts. Because the
36 1
362
BIOCHEMICAL SOCIETY TRANSACTIONS
molecules were fluorescent, we were able to examine their
distribution within living cells by high-resolution fluorescence microscopy and correlate changes in intracellular distribution with lipid metabolism (for a review, see [ 161). The
observed metabolism of these molecules was particularly
significant because it demonstrated that even though these
synthetic lipids were fluorescent and therefore 'non-natural',
the enzymes involved in lipid metabolism recognized them.
Our interest in the Golgi apparatus began in 1981 when
Naomi Lipsky, a former post-doctoral fellow, joined my
laboratory. I suggested that she synthesize a fluorescent
analogue of ceramide, [ N-( E-NBD-aminocaproyl )-D- erythrosphingosine; C,-NBD-Cer; Fig. 21, and study its behavior in
cells. This analogue was potentially interesting because
endogenous ceramide is a key intermediate in sphingolipid
biosynthesis - it is metabolized to sphingomyelin, a major
structural lipid in cells, and to glycolipids and gangliosides,
relatively minor lipids which play important roles in cell
function (e.g. cell-cell recognition).
In our initial studies [ 171, using Chinese hamster fibroblasts, we found that when cells are treated with C,-NBDCer at low temperature and washed, the endoplasmic
reticulum (ER),nuclear envelope, and mitochondria become
fluorescently labelled. Upon warming the cells to 37"C, the
fluorescent ceramide is metabolized to fluorescent
sphingomyelin (SM) and glucosylceramide (GlcCer) and,
concomitantly, the Golgi apparatus becomes intensely
fluorescent. After additional time at 37"C, the plasma
membrane also becomes visibly labelled as fluorescent SM
and GlcCer are transported to the cell surface. Transport of
the newly synthesized fluorescent sphingolipids to the
plasma membrane is inhibited in mitotic cells [IS]and by the
ionophore monensin 1191, as is the movement of newly
synthesized glycoproteins to the cell surface under these conditions [20-221. These results indicate that the fluorescent
SM and GIcCer analogues are synthesized intracellularly and
then transported from the Golgi apparatus by a vesicle-
HO
0 =c
NH
l4C-Ceramide
C,-NBD-Cer
Fig. 2. Molecular structures of fluorescent and radioactive
analogues of ceramide
D-etythro-sphingosine was Nacylated with either 1[ 14C]hexanoicacid ( l4C-Ceramide)or E-NBD-aminocaproic
acid (C,-NBD-Cer).
mediated process analogous to the movement o f plasma
membrane and secretory proteins [ 231 between the Golgi
apparatus and the plasma membrane.
In a study published in 1985, we showed that C,-NBDCer prominently stained the Golgi apparatus of a number of
different cell types and was thus a vital stain for this organelle
[24]. Since then, we have used this fluorescent analogue on
numerous cell types - in every case we have found prominent labelling o f this unique organelle. To our knowledge,
C,-NBD-Cer is the only known vital stain for the Golgi
apparatus.
As a result of these findings, many laboratories are now
using C,-NBD-Cer in various cell biological studies of the
Golgi apparatus. One particularly interesting study carried
out by van Meer and colleagues demonstrated that in
polarized epithelial MDCK (Madin-Darby canine kidney)
cells, C,-NBD-Cer is metabolized to C,-NBD-SM and
-GlcCer and these lipid metabolites are preferentially
delivered to the apical cell surface [25]. This polarized
delivery is consistent with the known enrichment of glycosphingolipids and SM in apical membranes 1261, and suggests
that C,-NBD-Cer and its metabolites are recognized by the
cellular sorting and transport machinery in a manner similar
to their native counterparts. Another intriguing study using
C,-NBD-Cer is being carried out by Cooper and colleagues
at Yale University [ 271. Using fluorescence video imaging
they have studied the dynamics of the Golgi apparatus in the
living cell and have provided evidence that the Golgi apparatus rapidly extends tubulovesicular processes along microtubules. It is not yet known whether these processes are
involved in transport of materials among the various Golgi
subcompartments or from the Golgi apparatus t o other organelles.
Over the years we have attempted to define the aspects of
the structure of C,-NBD-Cer which are responsible for the
unique properties described above. In one study [28], we
synthesized a series of fluorescent N-acyl-sphingosine
analogues, similar to C,-NBD-Cer, using various long-chain
bases and fluorescent fatty acids (Fig. 3) and then studied
their metabolism and intracellular distribution in cultured
fibroblasts. While most of the fluorescent N-acylsphingosines were significantly metabolized to the corresponding fluorescent SM, metabolism of these analogues to
fluorescent GlcCer was strongly dependent o n the longchain base and the stereochemistry of the fluorcscent fatty
acid moiety. When cells were incubated with the various
fluorescent N-acyl-sphingosine analogues under appropriate
conditions, prominent labelling of the Golgi apparatus was
seen in all cases except for N-( ~-NBD-aminohexanoyl)-3keto-sphingosine. These findings provide a basis for designing other N-acyl-sphingosine derivatives which might be
useful in future studies of (g1yco)sphingolipid metabolism,
sorting, and transport. In particular, an N-acyl-sphingosine
analogue which accumulates in the Golgi apparatus and
which bears a radioactive and photoactivatable fatty acid
might be useful in identifying some of the proteins in the
Golgi apparatus which are involved in these processes.
Although a large number of membrane proteins would
probably be labelled in such an experiment, the use of
different long-chain bases, as described above, could modify
the metabolism and intracellular distribution of the labelled
N-acyl-sphingosine and aid in identifying the proteins o f
interest. [Inpreliminary experiments, we have synthesized Derythro-sphingosines which are N-acylated with short-chain
radioactive and/or photoactivatable fatty acids [ 3-(4acid (Bolton and
hydroxy, 3-1 1251]di-iodophenyl)propionic
Hunter reagent), 3-[1251]iodo,4-azidosalicylic acid] or with a
biotinylated fatty acid, and found them t o be metabolized in
a manner similar to C,-NBD-Cer in cultured fibroblasts (R.
E. Pagano, unpublished work).J
1990
363
SIXTH MORTON LECTURE
Long-chain bases
Sphingosine
Fluorescent fatty acids
. .
cnl-(cn I-cn=cn-cn-cn-cn on
211
I
on
1
N+$
2
f-NBD-aminohexanoic acid
,O.
N
Y
Dihydrosphingosine
. .
C H -(CHI)-CH-CH-CH on
3
141
I
on N%
a-OH,€-NBD-aminohexanoicacid
2
N
,o.
Y
Phytosphingosine
CH
-(cn
3
. . .
)-cn-c~-c~-c~~o~
211
bH bH
h$
a-NBD-aminohexanoic acid
0
I,
n*c-c
.
,H-(C%
\-c n,
WH
3-Ketosphingosine
c H~-(cH~)
- c H=C H-C-c
9 1
H- cn20H
1 A%
Fig. 3. Vurioirs long-chuin buses and fluorescent fatty acids used in the synthesis of
flirorescent compounds striicturully related to C,-NBD-Cer
Asterisks indicate asymmetric carbon atoms which result in stereoisomers of the
indicated compounds. The following stereoisomers were used in the synthesis of the
fluorescent N-acyl-sphingosines [ 2 8 ] :D-erythro-sphingosine, L-erythro-sphingosine,
D-threo-sphingosine, L-threo-sphingosine, D-erythro-dihydrosphingosine, L-threodihydrosphingosine, D- or L-a-OH, &-NBD-aminohexanoic acids, D- or L-a-NBDaminohexanoic acids.
Surprisingly, labelling of the Golgi apparatus even
occurred in cells which had been fixed before incubation
with the fluorescent lipid. When cells are fixed with glutaraldehyde and subsequently incubated with C,-NBD-Cer,
labelling of the Golgi apparatus is detected, but is difficult to
see clearly because of a high 'background' from other
fluorescently labelled intracellular membranes within the
treated cells. However, this background fluorescence can be
removed when the labelled cells are further incubated
('back-exchanged') with defatted albumin. The result is a
preparation of fixed cells in which only the Golgi apparatus
is prominently stained by C,-NBD-Cer (Fig. 4).
We also developed a method for visualizing the fluores1'11~.( h l g i uppururiis us u 'rnolecwlur trap 'for cerumide /-?(I,
311
cent lipid at the level of the electron microscope, based on
The labelling o f the Golgi apparatus by C,-NBD-Cer is so the photoconversion of a fluorescent marker to a diaminohighly specific and striking that we further investigated the benzidine (DAB) product [32, 331. In this procedure, cells
mechanism responsible for its accumulation in that organelle. labelled with C,-NBD-Cer are irradiated in the presence of
(Control experiments using other C,-NBD-lipids showed DAB. During irradiation, the fluorophore is photobleached
that Golgi labelling is specific for ceramide and is not simply and the photo-oxidation products catalyse the polymerization of DAB. This results in an electron-opaque osmiophilic
duc t o the presence o f the NBD fluorophore.) One possibility
is that the fluorescent lipid, and perhaps its metabolites, are polymer which appears black when visualised in the electron
transported to the Golgi apparatus by an energy- and microscope and is present at sites within the specimen which
temperature-dependent process analogous to that seen for were originally labelled by the fluorescent lipid [ 31 1. As
the delivery of newly synthesized proteins from the ER to the shown in Fig. 5, this procedure localizes C,-NBD-Cer to
Golgi apparatus. Alternatively, C,-NBD-Cer labelling might only one or two stacks of the Golgi apparatus, although
be due t o binding of the fluorescent lipid to enzymes numerous small vesicles were also labelled in some areas,
possibly resulting from tangential sections through the Golgi
involved in ccramide metabolism. or the lipid might preferentially partition into this organelle as a result of some cisternae. Using other markers for known subcompartments
special 'physical property' of the Golgi membranes. In our of the Golgi apparatus in double-label experiments, we
initial attempts t o distinguish between these possibilities, we determined that C,-NBD-Cer labelling is restricted to the
were unable to inhibit C,-NBD-Cer labelling of the Golgi [runs-Golgi cisternae [ 3 1 1.
Several pieces of evidence suggest that the trapping of
apparatus at low ternpcraturcs or in the presence o f various
C,-NBD-Cer at the Golgi apparatus of fixed cells is due to its
metabolic inhibitors.
Recently, in collaboration with Drs Richard Haugland and
Hee Chol Kang o f Molecular Probes. Inc. (Eugene, OR,
U.S.A.), we have begun to explore other ceramide analogues,
similar t o C,-NBD-Cer, in which different fluorochromes are
substituted for the NBD moiety. Our hope is to obtain additional fluorescent analogues which are metabolized and
translocated in a manner similar to C,-NBD-Cer, but which
fluoresce at different wavelengths. Such probes might then be
used with C,-NBD-Cer in resonance energy transfer microscopy studies [29]t o examine the kinetics of lipid translocation through the Golgi apparatus of single living cells.
VOl. 18
364
BIOCHEMICAL SOCIETY TRANSACTIONS
Fig. 4. Accumulation of C,-NBD-Cer at the Golgi apparatus
oftixed cells
Human skin fibroblasts were fixed with glutaraldehyde,
washed, incubated with C,-NBD-Cer, and then backexchanged with defatted bovine serum albumin before
observation and photography [31].Bar represents 20 pm.
interaction with endogenous Golgi apparatus lipids. First,
accumulation of C,-NBD-Cer occurs in cells fixed in various
ways, but this accumulation is inhibited when fixation protocols which extract or modify cellular lipids are used. For
example, glutaraldehyde-fixed cells rendered permeable by
brief treatment with detergents and then washed extensively
do not exhibit post-fixation staining of the Golgi apparatus
by C,-NBD-Cer, although such treatments have no obvious
effect on the morphology of the Golgi apparatus as determined with fluorescent antibodies at the light microscope
level. Secondly, the amount of C,-NBD-Cer present after
back-exchange indicates that 1.0-4.2 x lo8 molecules of the
fluorescent lipid are present at the labelled Golgi cisternae of
each cell. Therefore, C,-NBD-Cer labelling of the Golgi
apparatus is probably not due to its binding to a resident
Golgi apparatus protein(s), since the required number of
copies of this protein( s) would be extraordinarily high.
Finally, Filipin, a fluorescent polyene antibiotic which forms
complexes with cellular cholesterol and labels the Golgi
apparatus of fixed cells [34, 351, inhibits accumulation of C,NBD-Cer at that organelle.
Recently, in collaboration with Drs Joan BlanchetteMackie and Peter Pentchev of the U.S. National Institutes of
Health, we have begun to explore the effects of alteration of
cellular cholesterol on accumulation of C,-NBD-Cer at the
Golgi apparatus. We found that when cellular cholesterol was
depleted, either by growth of cells in lipoprotein-deficient
serum or in medium containing inhibitors of cholesterol biosynthesis, C,-NBD-Cer labelling of the Golgi apparatus was
dramatically reduced (R. E. Pagano, 0. C. Martin, M. E.
Comly, J. Blanchette-Mackie & P. Pentchev, unpublished
work). Golgi labelling could be restored by stimulating endogenous cholesterol biosynthesis using mevalonic acid, a precursor of cholesterol biosynthesis, or by supplyFg
exogenous cholesterol to the growth medium. These strilung
observations suggest that the sterol content of the tramGolgi cisternae plays a critical role in C,-NBD-Cer labelling.
However, since cholesterol is also present in other cellular
membranes to which C,-NBD-Cer does not localize, we
speculate that additional Golgi apparatus constituents,
perhaps the endogenous (glyc0)sphingolipids found there,
Fig. 5 . Electron microscope localization of C,-NBD-Cer at the Golgi apparatus of
B e d cells
Chinese hamster ovary cells were fixed, incubated with C,-NBD-Cer, and backexchanged to give prominent labelling of the Golgi apparatus. The cells were then
photobleached in the presence of DAB, washed, treated with OsO,, and processed
for electron microscopy. Black deposition product at arrows represents sites of C,NBD-Cer localization in the fixed cells. N, nucleus. Bar represents 0.5 pm. Reproduced from the Journal of Cell Biology 1989, vol. 109, p. 2073, by copyright
permission of the Rockefeller University Press. See Pagano et al. [31].
1990
SIXTH MORTON LECTURE
arc also involved in the molecular trapping of C,,-NBD-Cer.
In future studies we hope to delineate further the components responsible for molecular trapping of C,?-NBD-Cer.
Sphingotnyelin synthesis occiirs predominantly at the cis and
medial cistemue of the G d g i appurutus
Our studies with C,-NBD-Cer in living cells [ 17, 191
suggested that the Golgi apparatus was the major site of SM
synthesis. However, this conclusion was somewhat controversial since previous studies from other laboratories
suggested that the ER, plasma membrane, Golgi apparatus,
or some combination of these, was the site of SM synthesis.
On the other hand, the biochemical pathway leading to SM
synthesis is now well established [36].SM is synthesized in
mammalian tissues from ceramide by the transfer of phosphorylcholine directly from phosphatidylcholine (PC):
-
PC + ceramide SM + DAG
generating a molecule of diacylglycerol (DAG) for each
molecule of SM formed.
Given the formation of DAG, an important second
messenger in cells [37], and the possible implications for
intracellular lipid movement if SM synthesis were to take
place at the Golgi apparatus 1381, we decided to determine
rigorously the intracellular site of SM synthesis [38a]. This
work was performed by Anthony Futerman, a post-doctoral
fellow in the laboratory, in collaboration with Drs Ann
Hubbard and Bruno Steiger of the Johns Hopkins School of
Mcdicinc. Well-characterized subcellular fractions were isolated from rat liver, incubated with a radiolabelled analogue
o f ceramide, N-( 1-1 lJC]hexanoyl)-D-erythro-sphingosine
(Fig.
2 ) . and the synthesis of radioactive SM was quantified. Using
a Golgi apparatus fraction which was highly enriched in the
marker enzyme, galactosyl transferase, and which showed
very little cross-contamination with other subcellular compartments, we found a 85-%-fold enrichment of SM
synthase activity. Small amounts of SM synthase activity
were also associated with enriched plasma membrane and
rough microsome fractions, but, after accounting for the contamination o f these fractions by Golgi apparatus membranes,
together these fractions contributed less than 13% of the
total SM synthase activity in liver. We conclude that at least
87% o f the SM synthesized in rat liver occurs at the Golgi
apparatus.
We have further separated the membranes of the Golgi
apparatus into fractions of heavy, intermediate, and light
density using the procedure of Ehrenreich et al. [39]. These
various fractions result from the maturation of nascent lipoproteins as they migrate through the Golgi complex and
acquire lipid, becoming lighter in density. As a result, Golgi
fractions from truns cisternac arc lighter in density than those
obtained from the cis and medial cisternae. These subfractions of the Golgi apparatus were definitively characterized by immunoblotting and biochemical assays using
cis/mediul (mannosidase 11) and trans (sialyl transferase and
galactosyl transferase) Golgi apparatus markers. Cis and
medial fractions were enriched in SM synthesis, with far less
activity associated with a trans fraction. Thus we conclude
that SM synthesis occurs predominantly in the cis and mediul
cisternac of the Golgi apparatus. This conclusion is
supported by our earlier studies using C,-NBD-Cer in cultured fibroblasts in which it was found that the ionophore
moncnsin did not interfere with the synthesis of C,-NBD-SM
from C,-NBD-Cer, but did inhibit the transport of the newly
synthesized S M to the plasma membrane 1191. Since
monensin blocks medial to truns movement o f glycoproteins
140. 4 1 I, this result suggested that SM synthesis occurred in
early compartments of the Golgi apparatus.
Vol. 18
365
Our studies demonstrating that C,-NBD-Ccr prcfcrcntially labels the trans-Golgi stacks of (fixed) cells I 3 1 1, while
metabolism to SM occurs in the cis- and mediul-Golgi
elements [38a], presents us with a paradox since the lipid
substrate is delivered distal to the compartment(s) wherc it is
metabolized. One possible explanation for this may be that
ceramide undergoes retrograde transport within the Golgi
apparatus during SM synthesis. Alternatively, there may be
multiple pools of C,-NBD-Cer in living cells which could
account for these findings.
A TP-dependent fusion of liposomes with the Golgi apparatus
1421
Lastly, I want to mention a remarkable finding concerning
the Golgi apparatus which was made by Toshihide Kobayashi when he was a post-doctoral fellow in this laboratory.
For years I had been interested in microinjecting fluorescently-tagged lipid vesicles (liposomes)into the cytoplasm of
cells to determine whether such vesicles would preferentially
associate with certain organelles. If so, it would then be
possible to systematically evaluate the roles of vesicle size,
lipid composition, charge, and other parameters on the distribution of these vesicles within cells. Unfortunately, such
experiments were not technically possible, because,
invariably, the lipid vesicles clogged the micropipettes used
for introducing them into cells. However, about the time Dr
Kobayashi joined the laboratory, several new methods were
introduced for generating semi-intact or ‘perforated’ cells
143, 441. Importantly, these preparations were shown to
support endogenous vesicle transport and fusion. With the
perforated-cell system, we were able to readily introduce
artificial lipid vesicles, labelled with various fluorescent
markers, into cells and examine their interactions with intracellular membranes [42]. To our surprise, we found that
when incubations were performed in the presence of an ATPregenerating system, both vesicle lipids and entrapped watersoluble markers were transferred to the Golgi apparatus of
treated cells, indicative of fusion of the liposomes with the
membranes of the Golgi apparatus. Fusion occurred using
unilamellar vesicles 30-80 nm in diameter and comprised of
phosphatidylcholine, but was inhibited by pretreatment of
the cells with N-ethylmaleimide. These findings demonstrated that lipid vesicles could be used to deliver labelled
lipids, macromolecules, and dyes to the Golgi apparatus, and
suggest several possibilities for future study. Experiments by
Peter Hoffmann-Bleihauer, another post-doctoral fellow
working on this system, are currently in progress to determine whether ( i ) liposomes fuse with all stacks of the Golgi
apparatus, or only a subset, and (ii) labelled lipids, once
delivered to the Golgi apparatus, are capable of ‘recycling’
back to the ER of the treated cells.
Fiiture directions
Our studies with fluorescent and radioactive analogues of
ceramide have yielded a vital stain for the Golgi apparatus
and established this organelle as the major site for sphingomyelin biosynthesis. It is my hope that future studies will
provide additional insights on how lipid and protein sorting
and translocation through the Golgi apparatus are interrelated and controlled. In particular we hope to ( i ) examine the
interrelationship of cholesterol biosynthesis and sphingolipid
metabolism in cells and determine the molecular signal(s) for
ceramide trapping at the Golgi apparatus; (ii) determine
whether retrograde transport of ceramide within the Golgi
apparatus occurs; (iii) identify and isolate the Golgi apparatus enzyme responsible for SM synthesis; and (iv) determine
whether ‘recycling’ of lipids from the Golgi apparatus to the
ER occurs.
366
Supported by grant R37 GM-22942from the US.Public Health
Service. I am especially grateful to Ms Ona Martin for her expert
assistance and many contributions during her years of service to the
Pagano laboratory. Space will not allow me to individually thank all
of the current and former students and post-doctoral fellows who
have worked on various aspects of the methodology which led to the
developments described in this article, but I am very grateful to all
of them.
1 . Golgi, C . ( 1 898)Arch. Ital. Biol. 30,278-286
2. Lipsky, N.G. (1989)J. Microsc. (Oxford) 155,9-14
3.Goldfischer, S. ( 1982)J. Histochem. Cytochem. 30.717-733
4.Thyberg, J. & Moskalewski, S. (1985)Exp. Cell Res. 159,1-16
5. Pavelka, M. (1987)
Adv. Anat. Embryol. CellBiol. 106,1-94
6. Dunphy, W. G. & Rothman, J. E. (1985)Cell (Cambridge,
MUSS.)42,13-21
7. Fleischer, B. ( 1988)in Protein Transfer and Organelle Biogenesis, Chapt. 7,pp. 289-316,Academic Press, London and New
York
8. Kornfeld, R. & Kornfeld, S. (1985)Annu. Rev. Biochem. 54,
631-664
9. Farquhar, M. G. ( 1 986)Annu. Rev. Cell Biol. 1,447-488
10. Hirschberg, C. B. & Snider, M. D. (1987)
Annu. Rev. Biochem.
56.63-87
I I. Palade, G. E. (1975)Science 189,347-358
12. Tartakoff, A. (1980)Int. Rev. Exp. Pathol. 22,227-251
13. Farquhar, M. G. & Palade, G. E. ( 1981 ) J . Cell Biol. 91 (No. 1,
Pt. 2),77s-103s
14.Griffiths. G. & Simons. K. 1986)Science 234.438-443
15. Rothman, J. E. (1 987)bell (Cambridge, Mass.) 50,521-522
16.Pagano,R.E.&Sleight,R.G.(1985)Science229,
1051-1057
17. Lipsky, N.G. & Pagano, R. E. (1983)/'roc. Natl. Acad. Sci.
U.S.A.80,2608-2612
18. Kobayashi, T. & Pagano, R. E. (1989)J. Biol. Chem. 264,
5966-5973
19. Lipsky, N. G. & Pagano, R. E. ( 1985)J . Cell Biol. 100.27-34
20. Griffiths, G., Quinn, P. & Warren, G. (1983)J. Cell Biol. 96,
835-850
21. Quinn, P., Griffiths, G. & Warren, G. (1983)J. Cell Biol. 96,
851-856
22. Warren, G. (1985)Trends Biochem. Sci. 10,439-443
23. Burgess, T. L. & Kelly, R. B. ( 1 987)Annu. Rev. C'ell Biol. 3,
243-293
BIOCHEMICAL SOCIETY TRANSACTIONS
24. Lipsky, N. G. & Pagano, R. E. ( 1985) Science 228,745-747
25. van Meer, G., Stetzer, E. H. K., Wijnaendts-van-Resandt. W. &
Simons,K.(l987)J. CellBiol. 105,1623-1635
26. van Meer, G. & Simons, K. ( 1986)EMRO J. 5 , 1455-1464
27. Cooper, M. S., Cornell-Bell, A. H., Chernjavsky, A. & Smith,
S.J.(1989)J.
CellBiol. 107,431a
28. Pagano, R. E. & Martin, 0. C . (1988) Biochemistry 27,
4439-4445
29. Uster, P. S. & Pagano, R. E. (1986)J . C'eN Biol. 103,
1221-1234
3 0 . Pagano, R. E. ( 1 989)Methods C'ell Biol. 29.75-85
3 1. Pagano, R. E.,Sepanski, M. A. & Martin, 0. C . (1989)J. ('ell
Biol. 109,2067-2079
32. Moranto, A. R. (1982)Science 217,953-955
33. Sandell, J. H. & Masland, R. (1988)J. Histochem. C'ytochem.
36,555-559
34. Orci, L.,Montesano, R., Meda, P., Malaise-Lagae. F., Brown,
D., Perrelet, A. & Vassalli, P. ( 1 981) /'roc. Nufl. Acud. Sci.
U.S.A.78,293-297
35. Blanchette-Mackie, E. J., Dwyer, N. K., Amende, L. M., Kruth,
H. S., Butler, J. D., Sokol, J., Comly, M. E., Vanier, M. T.,
August, J. T., Brady, R. 0. & kntchev. P. G. (1988)I'roc. Natl.
Acad. Sci. U.S.A.85,8022-8026
36. Voelker, D. R. & Kennedy, E. P. (1982)Riochernisrry 21,
2753-2759
37. Kikkawa, U., Kishimoto, A. & Nishizuka, Y. ( 1989)Annu. Rev.
Biochem. 58.31-44
38. Pagano, R. E. (1988) Trends Biochem. Sci. 13,202-205
38a.Futerman, A. H., Stieger, B., Hubbard. A. L. & Pagano, R. E.
( 1990)J. Biol. C'hem. in the press
39. Ehrenreich, J. H., Bergeron, J. J. M., Siekevitz, P. & Palade,
G. E. ( 1973)J. Cell Riol. 59,4572
40.Warren, G., Featherstone, C., Griffiths. G. & Burke. R. ( 1983)
J. CellRiol. 97,1623-1628
41.Featherstone, C., Griffiths, G. & Warren, G. (1985) J. C'ell Hiol.
101,2036-2046
42. Kobayashi, T. & Pagano, R. E. (1988) Cell (Cumbridge, Muss.)
55,797-805
43. Becker, C. J. M., Keller, D. S. & Balch, W. E. (1987)C'ell
(Cambridge, Mass.) 5 0 , 5 23-5 34
44. Simons, K. & Virta, H. (1987)EMBO J. 6,2241-2247
Received I 1 December 1989
1990