After reading both articles, you should be able to see how the discovery of sorting signals paved the way for utilizing these signals to provide information on the localization and functionality of proteins. Based on the Learn material for this module, propose an experiment for localizing a protein (of your choice) to an alternative cellular compartment. Then, speculate what characteristics or symptoms this would create on a cellular, tissue/organ, and organismal level. Lastly, comment on how this level of organization and exacting localization demonstrates God’s wonderful design in creation (Psalm 139).
Rubric:
1.All key components of the Discussion prompt are answered in the thread. The thread has a clear, logical flow. Major points are stated clearly. Major points are supported by good examples or thoughtful analysis. There are abundant, relevant references to one or more of the assigned article.
TRANSFER
OF PROTEINS
ACROSS
MEMBRANES
I. P r e s e n c e o f P r o t e o l y t i c a l l y P r o c e s s e d a n d U n p r o c e s s e d
Nascent Immunoglobulin Light Chains
On Membrane-Bound Ribosomes of Murine Myeloma
GONTER BLOBEL and BERNHARD DOBBERSTEIN
From The RockefellerUniversity,New York 10021
ABSTRACT
Fractionation of MOPC 41 DL-I tumors revealed that the mRNA for the light
chain of immunoglobulin is localized exclusively in membrane-bound ribosomes. It
was shown that the translation product of isolated light chain mRNA in a
heterologous protein-synthesizing system in vitro is larger than the authentic
secreted light chain; this confirms similar results from several laboratories. The
synthesis in vitro of a precursor protein of the light chain is not an artifact of
translation in a beterologous system, because it was shown that detached
polysomes, isolated from detergent-treated rough microsomes, not only contain
nascent light chains which have already been proteolytically processed in vivo but
also contain unprocessed nascent light chains. In vitro completion of these nascent
light chains thus resulted in the synthesis of some chains having the same mol wt as
the authentic secreted light chains, because of completion of in vivo proteolytically
processed chains and of other chains which, due to the completion of unprocessed
chains, have the same tool wt as the precursor of the light chain.
In contrast, completion of the nascent light chains contained in rough
microsomes resulted in the synthesis of only processed light chains. Taken
together, these results indicate that the processing activity is present in isolated
rough microsomes, that it is localized in the membrane moiety of rough
microsomes, and, therefore, that it was most likely solubilized during detergent
treatment used for the isolation of detached polysomes. Furthermore, these results
established that processing in vivo takes place before completion of the nascent
chain.
The data also indicate that in vitro processing of nascent chains by rough
microsomes is dependent on ribosome binding to the membrane. If the latter
process is interfered with by aurintricarboxylic acid, rough microsomes also
synthesize some unprocessed chains.
The data presented in this paper have been interpreted in the light of a recently
proposed hypothesis. This hypothesis, referred to as the signal hypothesis, is
described in greater detail in the Discussion section.
THEJOURNALOFCELLBmLOCV9 VOLUME67, 1975 . pages 835-851
835
Biological membranes present a diffusion barrier
for macromolecules such as proteins, but transfer
of a large number of specific proteins across
membranes is an important physiological activity
of virtually all cells. Segregation by a membrane is
required not only for secretory proteins but also
for lysosomal and peroxysomal proteins and for
certain mitochondrial or chloroplast proteins synthesized in the cytoplasm. Transfer of proteins
across membranes may even be required for some
intramembrane proteins, e.g., if the site of insertion into the membrane were separated from the
site of synthesis by the lipid bilayer. The discovery
of an abundance of ribosome membrane junctions,
particularly in secretory cells in the mid 50’s (24,
25) and the demonstration in the mid 60’s (1, 26,
27, 29) that nascent chains synthesized on membrane-bound ribosomes are vectorially discharged
across the membrane, suggested that the ribosome membrane junction may function in the transfer of proteins across the membrane: by topologically linking the site of synthesis with the site of
transfer, the protein would transverse the membrane only in status nascendi in an extended form
before assuming its native structure, thus maintaining the membrane’s role as a diffusion barrier
to proteins.
However, the function of the ribosome membrane junction in the transfer of proteins across the
membrane did not explain the cell’s ability to
determine which proteins can traverse via the
ribosome membrane junction. Data accumulated
in the late 60’s (reviewed in reference 28) indicated
that m R N A ‘ s for some secretory proteins are
translated almost exclusively on membrane-bound
ribosomes while m R N A ‘ s for some cytosol proteins are translated on free ribosomes. In an
attempt to explain this dichotomy, a hypothesis
was suggested in 1971 (5). It was postulated that
all m R N A ‘ s to be translated on bound ribosomes
contain a unique sequence of codons to the right of
the initiation codon (henceforth referred to as the
signal codons); translation of the signal codons
results in a unique sequence of amino acid residues
on the amino terminal end of the nascent chain
(henceforth referred to as the signal sequence); the
latter triggers attachment of the ribosome to the
membrane. A somewhat more detailed version of
this hypothesis, henceforth referred to as the signal
hypothesis, is presented in this paper.
Thus far, the most compelling support for the
signal hypothesis derives from data reported by
several laboratories on the in vitro translation of
$36
lgG light chain m R N A isolated from murine
myelomas. The translation product of this m R N A
is larger than the authentic light chain (15, 19, 23,
32, 34-36). Furthermore, it was established by
peptide mapping and partial NH2-terminal sequence analysis (33) that the in vitro translation
product contains an extra sequence of ~ 2 0 amino
acid residues at the NH2 terminus. It has been
suggested independently that this extra sequence,
which is subsequently removed, may function in
the binding of the ribosome to the membrane
(23). We therefore have chosen the murine
myeloma as a model system for these studies. In
this paper we present fractionation and in vitro
protein synthesis experiments designed to examine
some aspects of the signal hypothesis.
METHODS AND MATERIALS
All experiments were carried out with a MOPC 41
murine myeloma obtained from Litton Bioneties, Kensington, Md. The light chain produced by this tumor
proved to be ~2,000 daltons smaller than that produced
by the original MOPC 41 obtained from Dr. M. Potter of
the National Institutes of Health, Bethesda, Md. We assume that this change results from the selection of a deletion mutant ctone during our initial transfers, and hereafter we wilt refer to this myeloma as MOPC 41 DL-I
(see also companion paper).
Fractionation o f M O P C 41 D L – I T u m o r
All sucrose solutions used in cell fractionation contained 50 mM triethanolamine. HCI pH 7.4 at 20~ 50
mM KCI, and 5 mM MgCI2 (TeaKM). t
Excised tumors freed of necrotic portions were passed
through an ice-cold tissue press and were homogenized in
~3 vol of 0.25 M sucrose with a few strokes in a
Potter-Elvehjem homogenizer. All subsequent operations
were performed between l-4~ The homogenate was
centrifuged for 10 rain at 10,000 g~ in an angle rotor to
yield a postmitochondrial supernate. The latter was
layered over 2.0 ml of 1.3 M sucrose and centrifuged for
15 rain at 105,000 gov in a Spinco no. 40 rotor (Beckman
Instruments, Inc., Spinco Div., Palo Alto, Calif.). The
supernate was aspirated with a syringe and subsequently
used for the preparation of free ribosomes. The pellet,
resuspended by homogenization in 2.3 M sucrose, was
used to isolate rough microsomes by the following
procedure: 3.5 ml of the suspension was loaded at the
1AbIJreviations used in thispaper: AR, autoradiography;
ATA, aurintricarboxylic acid; DTT, dithiothreitol; ER,
endoplasmic reticulum; L~ derived large ribosomal
subunits; PAGE, polyacrylamide gel electrophoresis; S~,
native small ribosomal subunits; SDS, sodium dodecyl
sulfate; TeaKM, 50 mM triethanolamine. HCl pH 7.4 at
200C, 50 mM KCI, and 5 mM MgCl~.
THE JOURNALOF CELL BIOLOGY . VOLUME67, 1975
bottom of a tube fitting the SB 283 rotor of the IEC
centrifuge (Damon/IEC Div., Damon Corp., Needham
Heights, Mass.) and overlayed with 3-ml aliquots each of
1.75 M, 1.5 M, and 0.25 M sucrose. The discontinuous
gradient was centrifuged for 12 h at 190,000 ga~. The
material banding in the 1.75 M sucrose layer was
removed with a syringe, diluted with 1 vol of TeaKM,
layered over a 2-ml cushion of 1.3 M sucrose, and
centrifuged for 30 min at 105,000 g,,~ in a Spinco no. 40
rotor to yield a pellet consisting essentially of rough
microsomes.
To isolate detached ribosomes, such pellets were
resuspended in TeaKM; a 10% solution of deoxycholate
in water was added to a final concentration of 1%, and
the mixture was layered over a 2-ml cushion of 2.0 M
sucrose. Centrifugation for 24 h at 105,000 ga~ in a
Spinco no. 40 rotor yielded a pellet of detached ribosomes.
Free ribosomes were prepared from the 30,000 g
supernate (see above) which was layered on a discontinuous sucrose gradient containing 2-ml layers each of 2.0 M
and 1.75 M sucrose. Centrifugation for 24 h at 105,000
g,~ in a Spinco no. 40 rotor yielded a pellet of free
ribosomes.
Preparations of native small ribosomal subunits (S t~)
from rabbit reticulocytes, of derived large ribosomal
subunits (L~ from rat liver free ribosomes by the
puromycin-KCI procedure, and of pH 5 enzymes from a
high speed supernate of Krebs ascites cells were as
previously described (6, 13).
Isolation of Poly (A)-Containing R N A
from M O P C 41 DL-I Rough Microsomes
Pellets containing rough microsomes (3,000-5,000
Aa~, units) were resuspended in 60 ml 150 mM NaCI, 50
mM Tris-HCl pH 8.0, and 5 mM EDTA. A 10% solution
of sodium dodecyl sulfate (SDS) was added to a final
concentration of t.5% and RNA was extracted from this
mixture with phenol-chloroform-isoamyl alcohol (2) and
fraetionated on oligo (dT) cellulose as follows. RNA was
resuspended in H~O and subsequently adjusted to 400
mM NaCI, 50 mM Tris. HCI pH 7.5, and 0.2% SDS. The
final concentration of RNA was 100-120 A ~o,units/ml.
6 ml of that solution was mixed at room temperature by
gentle swirling with 4 ml of packed oligo (dT) cellulose
which had been washed several times in 400 mM NaC1,
50 mM Tris. HCI pH 7.5, and 0.2% SDS. The cellulose
was then sedimented at 1,000 g, washed twice with 100
mM NaCI, 50 mM Tris. HCI pH 7.5, 0.2% SDS, and
transferred to a column. After washing with at least 10
bed volumes of a solution of 100 mM NaCI and 50 mM
Tris. HCl pH 7.5, the poly (A)-containing RNA was
eluted with a solution of 10 mM Tris. HCI pH 7.5, and
0.1% SDS. The poly (A)-containing RNA was twice
precipitated with ethanol, then resuspended in doubledistilled water to a concentration of ~ 10 A2.ounits/ml,
and stored at -80~
Ceil-Free Protein-Synthesizing
Systems and Assays
The terms “readout” and “initiation” system have
been adopted for the sake of brevity. In the readout
system, previously started polypeptide chains are completed, whereas in the initiation system polypeptide
chains are synthesized de novo.
INITIATION SYSTEM: The reaction mixture (250 ~1)
contained: 25 t~mol of KCI, 5 t~mol of H EPES. KOH (pH
7.3 at 20~
0.75 ~tmol of MgCI2, 0.5/~mol of dithiothreitol (DTT), 0.25 umol of ATP, 0.05 ~mot of GTP, 1.5
t~mol of creatine phosphate, a few crystals of creatine
phosphokinase, 10 #Ci of a reconstituted protein
hydrolysate (algal profile) containing 15 ~*C-amino
acids, 7.5 nmol each of the five amino acids not present
in the algal hydrolysate (asparagine, cysteine, glutamine,
methionine, and tryptophan) as well as S s (0.4 A~e,
units), L o (1.2 Auo units), 100 ~.1 pH 5 enzymes and
poly (A)-eontaining RNA (0.05 Ase, units).
READOUT SYSTEM: The composition of this system was identical to that of the initiation system except
that it contained either free or detached ribosomes or
rough microsomes instead of SN, L~ and mRNA (in one
case it also contained Sr~).
Incubation in both systems was at 37~ 10-~1 aliquots
(unless indicated otherwise in figure legends) were removed at indicated time intervals and spotted on 3M
Whatman filter paper disks, which were processed according to Marts and Novelli (21). Radioactivity was
determined in toluene-Liquifluor (New England Nuclear
Corp., Boston, Mass.) in a Beckman LS 350 liquid
scintillation counter at about 75% efficiency.
PROTEOLYSIS
OF
TRANSLATION
PROD-
25-~1 aliquots removed from the two systems
described above after incubation (see figure legends),
were cooled to 0-2~ in an ice bath, and each treated for
3 h at the same temperature with 3 ~1 of a solution containing trypsin and chymotrypsin (500 ~tg of each per
ml). Proteolysis was terminated by the addition of 1 vol
of 20% TCA, and the ensuing precipitate was prepared
for SDS-polyacrylamide gel electrophoresis (PAGE) as
described below.
UCTS:
ANALYSIS
OF
TRANSLATION
PRODUCTS
BY
SDS-PAGE” 25-~tl aliquots removed from the two
systems either after completion of, or at various times
during, incubation were cooled to 0-2~ in an ice bath
and treated with an equal volume of ice-cold 20% TCA;
after I h the ensuing precipitate was collected at 0-4~
by centrifugation in a swinging bucket rotor for 10 rain at
2,000 g. The supernate was removed as completely as
possible and the precipitate was dissolved by incubation
for 20 rain at 37~ in 30 ~1 of a solution containing 15%
sucrose, bromophenol blue (serving both as a pH indicator for the sample and as a tracking dye for electrophoresis), 100 mM Tris base and 8 mM DTT (if the
solution turned yellow [pH 3], Tris base was added in l~al aliquots to restore the blue color [pH 4.5 and
G. BLOBEL AND B. DOBBERSTEIN
Transferof Proteins across Membranes. 1
837
higher]). Solubilization was completed by incubation in
a boiling water bath for 2 rain. After cooling to room temperature, 2 ~i of a 0,5 M solution of a-iodoacetamide
was added to each sample, and the mixture was incubated for ! h at 37″C before a 25-91 aliquot was layered
into a slot of a polyacrylamide slab gel.
Rabbit globin, porcine chymotrypsinogen, ovalbumin,
and bovine albumin were treated in an identical manner
and were used as standards for mol wt determinations.
The slab gel (I mm thick) consisted of a 10-15%
acrylamide gradient serving as a resolving gel and a 5%
acrylamide stacking gel, both in SDS and buffers as
described by Maizel (20). Electrophoresis was for 20 h
and at constant current.
After electrophoresis, the slab gel was stained in a
solution containing 0.2% Coomassie Brilliant Blue, 50%
methanol, and 10% glacial acetic acid for 2 h and then
destained in 50% methanol and 10% acetic acid. After
destaining, the gel was soaked in the last solution with 5%
glycerol added; this was helpful in preventing the gels
from cracking during and after drying on Whatman 3 M
paper.
AUTORADIOGRAPHY
ACRYLAMIDE
GELS
(AR)
AND
OF
DRIED
POLY-
DENSITOMETRIC
OF BANDS: Dried gels were exposed to
medical X-ray film (Cronex 2D, du Pont de Nemours
and Co., Inc., Wilmington, Del.), generally for a few
days. The films were developed by conventional procedures.
The bands in the developed X-ray films were analyzed
using a Joyce-Loebl densitometer (Joyce, Loebl and Co.,
Inc., Burlington, Mass.). The area under the resulting
peaks was integrated and used as a quantitative measure
for the radioactivity in the bands.
The validity of this quantitation procedure was verified by analyzing the autoradiographs derived from 10,
20, and 30-#1 aliquots of a sample loaded into different
slots; it was found that the area under each peak in
the densitometry tracing was proportional to the amount
loaded into the slot.
ANALYSIS
Preparation of Labeled Secretion Product
from MOPC 41 and MOPC 41 DL-I
Freshly excised tumor freed of necrotic portions was
sliced into small pieces using a razor blade. The tumor
slices were washed several times in 150 mM NaCI, 20
mM Tris. HCI pH 7.4, and 5 mM MgC12, using centrifugation in a swinging bucket rotor at 500 g and at 4QC.
Approximately 1 ml of packed tumor slices were resuspended in 3 ml e r a medium containing minimal essential
medium balanced salt solution for suspension cultures,
vitamins, bicarbonate, and glucose as specified by Eagle
(11) and supplemented with 50 91 (=50 ~Ci) of 15
~’C-amino acids which were part of a reconstituted
protein hydrolysate (see above). The resuspended slices
were transferred to a 10-ml Erlenmeycr flask, gassed
with 95% 02-5% CO~, and incubated for 5 h at 37~
THE JOURNAL OF CELL BIOLOGY – VOLUME
67,
After incubation the tumor slices were centrifuged into a
pellet at i,000 g. The supernate was centrifuged for 1 h at
105,000 g in a Spinco no. 40 rotor. The resulting
supernate contained secreted, radioactively labeled lgG
light chains which were precipitated with 1 vol of ice-cold
20% TCA. The ensuing precipitate was prepared for
SDS-PAGE as described above.
Electron microscopy
Pellets were fixed in 2% glutaraldehyde in TeaKM for
1 h at 0~ and postfixed in 2% OsO, in TeaKM for I h at
0aC. The pellets were stained with 0.5% uranyl acetate in
acetate-Veronal buffer before dehydration and Epon
embedding (12, 18). Sections were cut on a Porter-Blum
MT2-B ultramicrotome (Dupont Instruments, Sorvall
Operations, Newtown, Conn.) equipped with a diamond
knife (Dupont Instruments, Wilmington, Del.). They
were stained with uranyl acetate (38) and lead citrate (37)
and viewed with a Siemens Elmiskop 101 at 80 kV.
Source of Materials
Oligo (dT) cellulose T-2 from Collaborative Research,
Inc., Waltham, Mass. ATP, disodium salt; GTP, sodium
salt; creatinr phosphate, disodium salt; and creatine
phosphokinase, salt-free powder from Sigma Chemical
Co., St. Louis, Me. a-iodoacetamide from Calbiochem,
San Diego, Calif. DTT from R. S. A. Corp., Ardsley, N.
Y. Coomassie Brilliant Blue and reconstituted protein
hydrolysate (~(C), algal profile (1 mCi/l ml) from
Schwarz/Mann Div., Becton, Dickinson and Co.,
Orangeburg, N.Y., bovine pancreatic trypsin, 2x crystallized (185 U/rag), and bovine pancreatic a-chymotrypsin, 3• crystallized (49 U/rag), from Worthington
Biochemical Corp., Freehold, N. J.
RESULTS
Cell Fractionation of MOPC 41 DL-I
Tumors
A m o n g the characteristic ultrastructural features of murine m y e l o m a s are greatly dilated
endoplastic reticulum (ER) cisternae and the occurrence of intracisternal A particles budding from
the E R m e m b r a n e s . These features may have
contributed to the unusually high density of rough
microsomes, the bulk of which were found to band
isopynically at 1.75 M sucrose. Some rough microsomes sedimented even through a 2.0 M sucrose cushion (conventionally used in the fractionation of liver cells (3) as a cutoff concentration for
preventing sedimentation of rough microsomes)
and therefore were present as c o n t a m i n a n t s in the
free ribosome fraction (see Fig. 2). Rough microsomes were further fractionated by detergent
t r e a t m e n t to isolate a fraction referred to as
1975
detached ribosomes. Sedimentation profiles of free
as well as detached ribosomes in sucrose gradients
are shown in Fig. I. Many of the ribosomes,
particularly in the case of detached ribosomes,
were in the form of polysomes indicating that
R N a s e action during cell fractionation was minimal. Characteristic for the profile of detached
ribosomes (Fig. I, BR) is the presence of significant amounts of large ribosomal subunits (designated L), whereas the profile of free ribosomes
(Fig. 1, FR) is distinguished by the presence of a
large amount of monosomes and a small but
significant amount of SN; large ribosomal subunits
may be present but may not be resolved from the
monosome peak. Furthermore, some material contained in the free-ribosome preparation had sedimented to the bottom of the sucrose gradient tube.
This pellet was fixed, stained, sectioned, and
inspected by electron microscopy. It showed both
I
BR
ii
FIGURE I Sedimentation profile of free (FR) and detached (BR) ribosomes from MOPC 41 DL-I. Pellets of
free and detached ribosomes were resuspended in ice-cold
double-distilled water. 0.2-ml aliquots containing 3.0
A , , , units were layered on 12.5 ml of 10-40% sucrose
gradients in 100 mM KCI, 20 mM triethanolamine. HCI
pH 7.4 at 20″C, and 3 mM MgCI,. The gradients were
centrifuged at 4~ in an SB 283 rotor of an IEC
centrifuge for 100 min at 190,000 g,,v. Fractionation of
the sucrose gradients and recording of the optical density
were as described (6). Arrow indicates direction of
sedimentation. The native small ribosomal subunit peak
is designated as S r~, the large ribosomal subunit peak as
L, the mono-, di-, tri-, and tetrasome peaks, as I, 2, 3,
and 4, respectively.
free ribosomes as well as rough microsomes (Fig.
2). A fraction collected from the sucrose gradient
(Fig. 1, FR) comprising the monosome peak as
well as the polysome region was sedimented by
centrifugation and prepared for electron microscopy. It showed only free ribosomes without contamination by rough microsomes (electron micrograph not shown). Thus, for purification of free
ribosomes, sucrose gradient centrifugation can be
used to eliminate rough microsome contamination.
Electron microscopy of the rough microsomal
fraction showed the characteristic ribosome-studded vesicles (Fig. 3). Occasionally a few lysosomes
were seen. Frequently the ER showed a thickening
and there were also intracisternal particles characteristic for murine myelomas.
In Vitro Translation of m R N A for the
Light Chain of lgG
The crude m R N A prepared from rough microsomes as described under Materials and
Methods was translated in a heterologous system,
developed in this laboratory (13) (henceforth re-
FIGURE 2 Presence of rough microsomes in the free ribosome fraction. Pellet resulting from sucrose gradient
centrifugation of free ribosomes (see Fig. 1) was prepared
for electron microscopy (see Materials and Methods).
• 25,500. The bar denotes 0.5 ~m.
G. BLOBELAND B. DOBBERSTEIN Transferof Proteins across Membranes. 1
839
wt of 25,000 and is therefore larger in mol wt by
~4,000 than the secreted light chain of IgG of
M O P C 41 DL-I (slot S), which has a mol wt of
21,000. It was tentatively identified as the “precursor” of the light chain on the basis of work by
other laboratories (15, 19, 23, 32, 34 36) showing
that the primary translation product of m R N A for
the light chain of lgG is longer than the secreted
light chain.
In Vitro Translation of m R N A ‘ s Contained
in Free Ribosomes, Bound Ribosomes, and
Rough Microsomes
The time-course of polypeptide synthesis in a
readout system, containing either free or detached
ribosomes and pH 5 enzymes, is shown in Fig. 6
and is compared to the time-course of m R N A
translation in an initiation system (see above).
Because pH 5 enzymes contain only small amounts
of initiation factors, polypeptide synthesis is essentially completed in the readout system after a
40-min incubation, whereas in the initiation system
translation continues for more than 120 rain,
FIGURE 3 Rough microsome fraction was isolated and although at a slower initial rate.
prepared for electron microscopy as described in MateAnalysis of the products by S D S – P A G E and
rials and Methods. Arrows point to local thickenings of AR (Fig. 7) showed that two major products were
the ER membrane and to intracisternal A particles. • synthesized by detached ribosomes. One of them
25,500. The bar denotes 0.5 am.
(slot B – , upward pointing arrow) is a polypeptide
of the same mol wt as the authentic secreted light
ferred to as initiation system) consisting of S N
from rabbit reticulocytes (as a source of small
ribosomal subunits as well as initiation factors), L~
prepared by the puromycin KCI procedure (6)
from rat liver free ribosomes, and pH 5 enzymes
from Krebs ascites cells. The time-course of polypeptide synthesis in this system in the presence or o
absence of m R N A is shown in Fig. 4. There was a
more than twofold stimulation of polypeptide
synthesis in the presence of m R N A . There was
inhibition of polypeptide synthesis in the presence
of aurintricarboxylic acid (ATA) at a concentration (10 -4 M) which has been reported to inhibit
initiation but not elongation in polypeptide synthesis (17). The extent of inhibition was similar in the
absence (data not shown) and in the presence of
mRNA.
15 30
60
90
120
180
Analysis of the products by S D S – P A G E and
minutes
AR (Fig. 5) showed that a prominent radioactive FIGURE 4 Time-course of polypeptide synthesis in an
band (slot B) was synthesized by the initiation initiation system in the presence of mRNA for the light
system in the presence of the crude m R N A chain of lgG (+Li), in the presence of light chain mRNA
fraction (supposed to contain mostly light chain and of 10-* M ATA (+Li + ATA), and in the absence of
m R N A ) . This polypeptide has an estimated tool mRNA (-Li). For details, see Materials and Methods.
m
.. 1
THE JOURNAL OF CELL BIOLOGY . VOLUME67, 1975
synthesized on free ribosomes from rat liver are
sensitive to mild proteolysis, except for a fragment
of ~40 amino acid residues on the carboxyl
terminal which is thought to be protected because
of its localization within the large ribosomal
subunit (4). Products synthesized by rough microsomes of rat liver, on the other hand, were
shown to be largely resistant to proteolytic attack
since they are protected by the surrounding ER
membrane (29). In agreement with these findings
are the results shown in Fig. 7 slots B and F. The
two major translation products of detached ribosomes were sensitive to mild proteolysis (slot B+);
however, more than 60% (estimated from densitometric analysis) of the band in the position of the
authentic light chain of lgG, apparently synthesized by rough microsomes present in the crude
free-ribosome fraction, was resistant to proteolytic
attack (slot F + ; see also Fig. 14 slots A + and
A-).
The synthesis by detached ribosomes of two
major polypeptides, one of them with the same mol
40-
S”.L~
30-
FmtJaE 5 Analysis by SDS-PAGE and AR of an
experiment of the type described in Fig. 4. Shown are the
labeled products synthesized at the 180-raintime point in ?
the absence of mRNA (slot A) or in the presence of
mRNA for the light chain of lgG (slot B). For compari- ~ 20son the labeled secreted light chain of lgG is shown in U
slot S (upward pointing arrow). Downward pointing
arrow (slot B) indicates precursor of the light chain of
lgG. Dots indicate the globin chains.
I0chain (shown in slot S); the other one (slot B – ,
downward pointing arrow) has the same mol wt as
the precursor of the light chain synthesized in the
initiation system (slot A). The products synthesized by crude free ribosomes (slot F – ) also
contained a band of a mol wt identical to that of
the light chain of lgG. However, this band was
shown to be due to the presence of rough microsomes in the crude free-ribosome fraction, since
it was absent when purified free ribosomes collected from a sucrose gradient (see above) were
tested in the same manner (data not shown).
It was shown recently that nascent polypeptides
o
/
~
~o
ao
d,o
minutes
FIGURE 6 Time-course of polypeptide synthesis in a
readout system containing 9.8 A2,o units of crude free
(FR) or 5.4 Aja, units of detached (BR) ribosomes. For
comparison the time-course of polypeptide synthesis in
an initiation system containing mRNA for the light chain
of lgG (S~, L~ Li) was included. In the latter case, 50-~1
aliquots were counted, whereas in the former lO-gl
aliquots were counted.
G, ]]LOBELAND ]3. DOBBERSTEIN Transfer of Proteins across Membranes. I
841
FIGURE 7 Analysis by SDS-PAGE and AR of products synthesizedat the 120-mintime point as described in
Fig. 6 after incubation in the absence or presence of
proteolytic enzymes. Shown are the labeled products
synthesized in an initiation system in the presence of light
chain mRNA (slot A), the labeled secreted light chain
(slot S), and the products synthesizedin a readout system
containing detached ribosomes (slot B-) subsequently
subjected to proteolysis (slot B+) or containing crude
free ribosomes (slot F-), subsequently subjected to
proteolysis (slot F+). Symbols used were as in Fig. 5.
wt as the precursor and the other with the same
tool wt as the secreted light chain, suggested that
isolated detached ribosomes are heterogeneous
with respect to their content of processed and
unprocessed nascent light chains. On the basis of
the predictions made in the signal hypothesis, one
could reason that those ribosomes located near the
5′ end of mRNA should contain unprocessed
nascent light chains that still have their signal
sequence, while those near the 3′ end should
contain already proteolytically processed nascent
light chains.
These assumptions were borne out by data
obtained from a time-course experiment using
detached ribosomes in a readout system; translation was stopped at various time points and the
products were analyzed by SDS-PAGE and AR.
To insure that there was no initiation in the
readout system, a condition which was not met in
the previous experiment, ATA was added in a
concentration which has been reported (17) to
inhibit initiation, but not elongation or release of
nascent chains. That such conditions were achieved
is demonstrated by the data shown in Fig. 4 and by
the lack of inhibition observed in the readout
system in the presence of ATA and detached
ribosomes (data not shown). From the autoradiograph shown in Fig. 8 it is evident (slots I and 6)
that at the earlier time points of readout, only
processed chains were synthesized, apparently as a
result of completion of chains by ribosomes near
the 3’ end of mRNA. Only at later time points
when ribosomes located further to the left on the
mRNA have completed their readout were unprocessed chains synthesized (slots 9, 18, 25, and 50).
The data of a quantitative analysis (see Materials
and Methods) of this experiment are summarized
in Fig. 9. it can be seen that synthesis of already
processed chains was essentially completed after a
10-min incubation when there was only a barely
detectable synthesis of unprocessed chains; the
latter were synthesized only in the following 30
min. no significant synthesis of either chain was
observed after a 50-min incubation (data not
shown).
While in the preceding experiment initiation had
to be ruled out in order to substantiate the
conclusion that detached ribosomes contain both
processed and unprocessed chains, the following
experiment was performed under conditions in
which initiation could take place, Such conditions
were achieved by adding to the readout system,
containing detached ribosomes, increasing
amounts of S N from rabbit reticulocytes as a
source of initiation factors (13). Up to twofold
stimulation in polypeptide synthesis was observed
as a result of the addition of increasing amounts of
S N (Fig. 10). This stimulation could have been the
result of the presence of small amounts of globin
mRNA present in the S N fraction. However, this
was ruled out by product analysis using SDS-
842 THE JOURNALOF CELL BIOLOGY 9 VOLUME67, 1975
FIGURE 8 Analysis by SDS-PAGE and AR of the products synthesized by detached ribosomes (1.8 A260
units) during the course of readout in the presence of 10-4 M ATA. Readout was terminated at 1, 6, 9, 18,
25, and 50 min (slots 1, 6, 9, 18, 25, 50, respectively)by cooling 25-~1aliquots to 0~ and adding 25 ~l 20%
TCA. For comparison, the precursor of the light chain (slot A) synthesized in an initiation system (see Fig.
5) and the labeled secreted light chain (slot S) were included. Symbols used were as in Fig. 5. Slots 25 and 50
were from a separate slab gel.
PAGE and AR (Fig. 11). Only small amounts of
globin were synthesized in response to increasing
amounts of added S N Quantitative analysis (Fig.
12) of the radioactivity in the bands corresponding
to the processed and unprocessed light chain of
IgG revealed that the stimulation by increasing
amounts of S N can be accounted for by a proportional stimulation in the synthesis of unprocessed
chains while there was no stimulation in the
synthesis of processed chains.
Finally, readout experiments using isolated
rough microsomes were performed. Fig. 13 shows
the time-course of polypeptide synthesis in rough
microsomes in the absence and presence of ATA,
which produced only a slight inhibition of amino
acid incorporation. Analysis of the products by
SDS-PAGE and AR (Fig. 14) revealed the synthe-
sis of a polypeptide corresponding to the mol wt of
the processed light chain of lgG (slot A – ) . In
contrast to the results obtained with detached
ribosomes (see above), newly synthesized unprocessed chains were not detected. However, unprocessed light chains were synthesized when in vitro
readout of rough microsomes took place in the
presence of 10-‘ M ATA (Fig. 14, slot B-).
The products synthesized in the readout experiment with rough microsomes were subjected to
mild proteolysis and subsequently analyzed by
SDS-PAGE and AR. As expected, the processed
chains, presumably inside the microsomal vesicles,
were largely protected from proteolytic attack.
Densitometric analysis revealed that ~60% was
resistant to proteolysis in agreement with the
results shown in Fig. 7, slot F+. However, the
G. BLOBELAND B. DOBBERSTEIN Transfer o f Proteins across Membranes. 1
843
PLi_
A
“~2″
i
0
ib
2’o
io
3’o
minutes
5b
FIGURE 9 Quantitation by densitometry of the autoradiograph shown in Fig. 8. PLi and Li designate the
unprocessed (downward pointing arrow in Fig. 8) and the
processed light chains of lgG (upward pointing arrow in
Fig. 8), respectively.
The majority of the data presented in this paper
were obtained from in vitro translation of the light
chain m R N A contained in the isolated rough
microsome and detached ribosome fractions of
MOPC 41 DL-I. it was shown here that both
fractions contain unprocessed light chains (i.e.,
chains still containing the signal sequence) together with processed light chains, demonstrating
that removal of the signal sequence in vivo takes
place well before the nascent chain is completed.
However, the difference between these two fractions is that in vitro only rough microsomes retain
their ability for proteolytic removal of the signal
sequence. Thus, in vitro completion of their nascent chains also results in concomitant proteolytic
removal of the signal sequence from their unprocessed chains, and thus in the synthesis of only
processed chains (see Fig. 14, slot A – ) . In contrast, detached ribosomes, having lost their ability
J
unprocessed chains synthesized on rough microsomes in the presence of ATA were degraded
(Fig. 14, slot B+), supporting the interpretation
that these chains were not segregated in the
intravesicular space.
15’
”
50
/ ~ 1 0
DISCUSSION
The results of cell fractionation reported here are
in agreement with those of Cioli and Lennox (10).
Although our cell fractionation was performed on
solid MOPC 41 DL-I tumors, a conventionally
prepared free-ribosome fraction also was found to
be contaminated by rough microsomes (see Fig. 2).
This contamination was eliminated by isokinetic
sucrose gradient centrifugation. Purified free ribosomes did not contain any detectable light chain
m R N A activity, although it cannot be ruled out
that some light chain m R N A was present and
resulted in the synthesis of unprocessed light
chains, which overlapped with the presence of
other bands in the autoradiograph (see Fig. 7, slot
F-).
Upon in vitro translation in an initiation system
(see Materials and Methods) of a crude m R N A
fraction containing the m R N A for the light chain
of lgG, a product was synthesized which was larger
by ~4,000 tool wt than the authentic secreted light
chain. Similar results have been reported by several laboratories (15, 19, 23, 32, 34 36).
-0
O
5″
0
50
minutes
I00
FIGURE 10 Time-course of polypeptide synthesis in a
readout system containing detached ribosomes ( 1.8 A ~oo
units) and S N from reticulocytes as a source for initiation
factors. Numbers next to curves designate the amount (in
microliters) of S N (10.5 A , , units/ml) present in the
250-~d assay.
84,1 THE JOURNALOF CELL BIOLOGY VOLUME67, 1975
.
IO”
(more than one half in the former, see Fig. 9)
indicated that probably more than one of the
polysomal ribosomes contain unprocessed chains.
This suggested that removal of the signal sequence
occurs only after the ribosome has already translated a considerable portion of the mRNA. If
removal of the signal sequence is an endoproteolytic event and if the processing activity is
localized in the membrane trans rather than cis
with respect to the ribosome-membrane junction,
then the entire signal sequence would be required
to have traversed the 70-A distance which comprises the thickness of the ER membrane before
processing could take place. Assuming that the
signal sequence comprises ~20 amino acid residues, and adding to these ~ 19 and ~39 residues
which comprise the portions of the nascent chain
(3.6 A per residue in the extended configuration) in
the membrane and in the ribosome (4), respectively, then a total of ~78 amino acid residues have
to be polymerized before processing can take
place. Since the interribosomal distance on a
ribosome-saturated polysome amounts to 90 nucleotides, or 30 codons, it is possible for the
m R N A to accommodate two to three ribosomes
A L
I0
PLi
FIGUR[ I I Analysis by SDS-PAGE and AR of the
products synthesized at 120-rain time point as described
in Fig. 10. Slot numbers refer to microliters ofS Npresent
in the readout system with detached ribosomes (see Fig.
10). For comparison, the labeled secreted light chain of
IgG is included in slot S. Symbols used were as in Fig. 5
C
for in vitro processing, yielded both unprocessed
as well as processed light chains upon completion
of their nascent chains in vitro. This result can be
rationalized by assuming that the processing activity is part of the membrane and was lost during the
preparation of detached ribosomes by detergent
solubilization of rough microsomes. Alternatively,
the processing enzyme(s) may still be present in
isolated detached ribosomes but in an inactivated
form or may become inactivated rapidly during
incubation in the readout system.
The data presented here also indicated that
those polysomal ribosomes containing unprocessed chains are located on the mRNA to the left
of those containing already processed chains (Figs.
8 and 9). Furthermore, the distribution of radioactivity between unprocessed and processed chains
I-
V
Li
Li
“m
o
io
i5
go
~1 S ~
FIGURE 12 Quantitation by densitometry ofautoradiograph shown in Fig. l l. PLi and Li designate the
unprocessed and processed chains of lgG, respectively.
G. BLOBELAND B. DOBBERSTEIN Transfer o f Proteins across Membranes. 1
845
RM
RM .ATA
?
o
it
E
~4
0
I0
20
30
minutes
essed chains was not affected. This result also
demonstrated that initiation factors from rabbit
reticulocytes which contain predominantly free
ribosomes were able to perform in the translation of the light chain mRNA contained in bound
ribosomes, supporting the contention that identical
initiation factors are used for translation on free
and bound ribosomes.
It should be noted that the presence of unprocessed chains in rough microsomes was detected
only if completion of their nascent chains in vitro
occurred in the presence of ATA. This result is of
particular interest since Borgese et al, (8) have
recently shown that ribosomes will not bind to
stripped membranes in vitro in the presence of
ATA. This result can therefore be rationalized as
60
FIGURE 13 Time-courseof polypeptide synthesis in the
absence of ATA (RM) or presence (RM+ATA) of 10-‘
M ATA in a readout system containing rough microsomes (4.3 A.o units).
still containing the signal terminal of their nascent
chain, i.e., containing unprocessed chains.
The synthesis by detached ribosomes of unprocessed chains of the same mol wt as the translation
product of light chain mRNA in a heterologous
reconstituted system suggests that the latter is not
an in vitro artifact. It could have been argued,
otherwise, that translation of the light chain
mRNA, which has been performed so far in all
cases in heterologous systems, resulted in artifactual initiation at some point to the left of the
initiation codon and therefore caused the synthesis
of a larger precursor protein, while in vivo such a
precursor protein would not have been synthesized.
By using ATA to inhibit initiation but not
elongation and release of the nascent chain, it was
clearly ruled out that the synthesis of unprocessed
chains by detached ribosomes reflected in vitro
initiation rather than completion of unprocessed
chains. Conversely, the observed twofold stimulation of polypeptide synthesis by detached ribosomes, in a readout system which was supplemented by initiation factors, was shown to be
entirely the result of increased synthesis of unprocessed chains, while the level of synthesis of proc-
846
FIGURE 14 Analysis by SDS-PAGE and AR of the
products synthesized at the 60-min time point as described in Fig. 13 and subsequently incubated in the
absence or presence of proteolytic enzymes. Products
synthesized in the absence or presence of 10-‘ M ATA
are shown in slots A and B, respectively. Products not
subsequently incubated with proteolytic enzymes (see
Materials and Methods) are shown in slots marked (-),
those incubated in slots marked (+). For comparison the
labeled secreted light chain of lgG is shown in slot S.
THE JOURNAL OF CELL BIOLOGY . VOLUME 67, 1975
follows: Those ribosomes located farthest to the
left on the m R N A in a rough microsome are not
yet attached to the membrane. In the absence of
ATA, attachment of these ribosomes as well as
subsequent processing of their nascent chains can
occur in vitro during chain completion, resulting in
the synthesis of processed chains. In the presence
of ATA, however, these ribosomes will not be able
to attach. This in turn will deprive their nascent
chains of access to the processing activity, but will
not interfere with their completion in vitro. Alternatively, it is also possible that the synthesis of
unprocessed chains in the presence of ATA does
not result only from an interference in establishing
the ribosome-membrane junction but is due to a
direct inhibition of the processing activity by ATA.
Finally, it was demonstrated here that the
microsomal membrane in the rough microsome
fraction provides protection for the completed and
released chain against proteolysis. Although protection of newly synthesized chains in microsomes
(isolated from rat liver) has been demonstrated
previously (29), it was assayed not by product
analysis but by measuring the percentage of acidinsoluble radioactivity remaining after proteolysis.
However, since this protection was observed to be
significantly less than 100%, this type of assay
could not distinguish between protection resulting
from partial hydrolysis of all chains or resulting
from a combination of complete resistance of some
chains and complete hydrolysis of others. Using
product analysis by SDS-PAGE and AR after
proteolysis, it was established here (see also companion paper) that the protection of the majority
of the product was complete in that there was no
reduction in its tool wt. This constitutes important
information since resistance to proteolysis (29) is
at present the only rigorous assay for vectorially
discharged chains. The original centrifugation
assay (showing that newly made chains were
sedimented with the membranes or were solubilized by detergent treatment) which was used to
demonstrate vectorial discharge (26, 27) has subsequently been shown to be inadequate (8, 9, 31).
It should be emphasized that most of our
conclusions remain to be confirmed by further
characterization of the in vitro translation products, in particular of the band which we have
assumed to be the unprocessed precursor of the
light chain of lgG (on the basis of its mol wt).
However, preliminary data 2 obtained by sequence
2Deviilers-Thiery, A., G. Blobel, and T. J. Kindt.
Manuscript in preparation.
analysis of 50 amino terminal residues of this
putative precursor protein support our conclusions.
A n Hypothesis f o r the Transfer o f Proteins
across Membranes
Most of the data which led to the formulation of
an hypothesis for the transfer of proteins across
membranes, referred to henceforth as the signal
hypothesis, have been reviewed previously (5) and
therefore will not be discussed here. Alternative
hypotheses have been summarized previously (30)
and will be omitted here. Instead, a more detailed
version of the signal hypothesis than that presented
previously (5), based on theoretical considerations
as well as recent data from this and other laboratories, will be outlined.
As mentioned above, the essential feature of the
signal hypothesis (illustrated in Fig. 15) is the
occurrence of a unique sequence of codons, located
immediately to the right of the initiation codon,
which is present only in those m R N A ‘ s whose
translation products are to be transferred across a
membrane. No other m R N A ‘ s contain this unique
sequence. Translation of the signal codons results
in a unique sequence of amino acid residues on the
amino terminal of the nascent chain. Emergence of
this signal sequence of the nascent chain from
within a space in the large ribosomal subunit
triggers attachment of the ribosome to the membrane, thus providing the topological conditions
for the transfer of the nascent chain across the
membrane.
Following is an attempt to formulate this sequence of events in greater detail. It is suggested
that translation of m R N A ‘ s containing signal
codons begins on a free ribosome. Thus, initiation
of translation of all m R N A ‘ s whether or not they
contain signal codons proceeds by the same mechanism, eliminating the need for specialized ribosomes or initiation factors. Similarly, elongation
will proceed on a free ribosome for both categories
of m R N A ‘ s until anywhere from 10 to 40 amino
acid residues of the nascent chain have emerged
from the ribosome. Only at this point is the
membrane able to distinguish between amino
terminals of nascent chains as containing or not
containing the unique signal sequence. If they lack
the signal sequence, attachment of the ribosome to
the membrane will not occur. If they contain the
signal sequence, attachment may occur, but does
not necessarily follow. It is conceivable, for example, that the availability of ribosome binding sites
G. BLOBELAND B. DOBBERSTEIN Transfer of Proteins across Membranes. 1
~17
FIGURE 15 Illustration of the essential features of the signal hypothesis for the transfer of proteins across
membranes. Signal codons after the initiation codon AUG are indicated by a zig-zag region in the mRNA.
The signal sequence region of the nascent chain is indicated by a dashed line. Endoproteolytic removal of
the signal sequence before chain completion is indicated by the presence of signal peptides (indicated by
short dashed lines) within the intracisternal space, For details see text.
o
~176176
oo
r
FIGURE 16 Hypothetical model for the formation of a
transient tunnel in the membrane through which the
nascent chain would be transferred. Specific regions of
the signal sequence (indicated by line thickenings) of the
nascent chain emerged from the large ribosomal subunit
tunnel are recognized by one site each (indicated by a line
thickening) on three membrane proteins. Recognition
results in loose association of these proteins, subsequently “cross-linked” by interaction of sites on the large
ribosomal subunit (indicated by notches around the
tunnel exit on the large ribosomal subunit) with one site
each on the three membrane proteins (indicated by
cones). For details see text.
in the membrane may be limiting, allowing translation of signal sequence-containing peptides to
continue to completion on nonattached ribosomes.
In this event, chains containing the signal sequence
would be released into the “soluble” compartment.
These chains may be rapidly degraded, since they
may be unable to assume their distinct native
structure through the enzymatic modifications
which are confined to the intracisternal space (e.g.,
proteolytic removal of the signal sequence or
glycosylation). If, on the other hand, ribosome
attachment occurs, translation will continue on a
bound ribosome until the nascent chain is released
and vectorially discharged. It is suggested that
after discharge of the completed chain, the ribosome is detached from the membrane (evidence for
the existence of a detachment factor will be
presented elsewhereS). The released ribosome may
again start translation of any m R N A , independent
of whether the latter does or does not contain
signal codons.
Ribosome attachment to, as well as detachment
from, the membrane are likely to involve a complex sequence of events. A number of theoretical
considerations form the basis for proposing the
following model (illustrated in Fig. 16). The signal
sequence of the nascent chain emerging from
within a tunnel in the large ribosomal subunit may
dissociate one or several proteins which have been
found to be associated with the large ribosomal
subunit of free ribosomes (7, 14, 22). Dissociation
of these proteins may in turn uncover binding sites
on the large ribosomal subunit. At the same time
the emerging signal sequence also recruits two or
more membrane receptor proteins and causes their
loose association so as to form a tunnel in the
membrane (see Fig. 16). This association is stabi-
THE JOURNAL OF CELL BIOLOGY 9 VOLUME 67, 1975
3 Blobel, G. Manuscript in preparation.
lized by each of these membrane receptor proteins
interacting with the exposed sites on the large
ribosomal subunit, with the latter playing the role
of a cross-linking agent. Binding of the ribosome
would link the tunnel in the large ribosomal
subunit with the newly formed tunnel in the
membrane in continuity with the transmembrane
space. After release of the nascent chain into the
transmembrane space, ribosome detachment from
the membrane would eliminate the crosslinking
effect of the ribosome on the membrane receptor
proteins. The latter would be free again to diffuse
as individual proteins in the plane of the membrane. As a result of their disaggregation, the
tunnel would be eliminated. The tunnel, therefore,
would not constitute a permanent structure in the
membrane.
Recognition of the signal sequence by membrane receptor proteins may require precise synchronization with translation. Thus if, after emergence of the signal sequence from within the
ribosome, translation were to continue without
concomitant ribosome attachment, folding of the
signal sequence with contiguous sequences of the
nascent chain might effectively prevent subsequent
ribosome membrane attachment. For this reason
we postulate confluence of the large subunit and
membrane tunnel, rather than other conceivable
arrangements in order to provide topological conditions which would prevent the formation of
secondary structures of the nascent chain at the
ribosome membrane junction.
The model described above thus provides for a
binding of the ribosome to the membrane which is
functional, specific, and transient, that is limited in
time as well as in space. Functional binding of the
ribosome is coupled to tunnel formation in the
membrane. By definition, then, nonfunctional
binding does not involve tunnel formation and
therefore does not provide the topology for the
transfer of the nascent chain across the membrane.
Functional binding also is specific in that it is
limited only to those ribosomes which carry nascent chains containing the signal sequence. Finally,
it is limited in time, in that it is linked to ongoing
translation and it is limited in space, in that it
occurs only on those membranes which contain
ribosome binding sites and which possess sufficient
fluidity to permit specific aggregation of these
sites.
It should be noted that the signal hypothesis
does not call for a direct attachment of m R N A to
the membrane. In fact, it predicts that if initiation
of protein synthesis were blocked, while elongation
and release were to proceed, m R N A containing
signal codons would be found in the soluble
compartment of the cell rather than on rough
microsomes. Cell fractionation then would recover
these m R N A ‘ s in the free ribosome fraction, most
likely in the form of m R N P ‘ s or bound to a free
ribosome.
The original version of the signal hypothesis did
not deal with the fate of the signal sequence.
However, since the amino terminal sequence is
different among the authentic secretory proteins,
removal of the signal sequence before actual
secretion was implied. It has been suggested (23)
that the extra sequence in the amino terminal of
the light chain precursor is removed by a membrane-associated activity. Data relevant to the
proteolytic processing of the nascent chain are
presented in this and in the following paper.
As already mentioned, the sequence of events
suggested in the signal hypothesis may not be
restricted to secretory proteins but may apply to
the synthesis of all proteins which have to be
transferred across a membrane. Thus, the
m R N A ‘ s for lysosoma! and peroxysomal proteins
may contain identical signal codons as the
m R N A ‘ s for secretory proteins, if the ribosomeER junction is utilized for their transfer across the
ER membrane. The same reasoning may apply to
the synthesis of certain mitochondrial proteins
which are synthesized in the cytoplasm. A junction
of cytoplasmic ribosomes with the outer mitochondrial membrane has been described (16) and could
be utilized for the transfer of these proteins across
the outer mitochondrial membrane, presumably
involving specific signal sequences and membrane
recognition sites which are different from those of
secretory proteins and the ER membrane, respectively. Finally, the synthesis of some membrane
proteins may require transfer through the membrane before the protein could be inserted into the
membrane. Depending on what particular membrane proteins would be required to be transferred
for subsequent insertion, signal sequences and
corresponding membrane recognition sites utilized
for secretory and mitochondrial proteins could be
involved. Other membrane proteins, in particular
in those instances in which the site of synthesis is
not separated from the site of insertion by the lipid
bilayer, may be synthesized on free ribosomes.
These considerations make it evident that differ-
G. BLOBELAND B. DOBBERSTEIN Transfer o f Proteins across Membranes. I
849
ent signal sequences may exist. Moreover, the
signal sequence for one group of proteins, e.g. for
secretory proteins, may not be identical in all
cases. Phylogenetic as well as ontogenetic variability may exist. Only after the sequence of a
sufficient n u m b e r of ” s i g n a l s ” has been established
will it be possible to recognize those sequence
features which are essential for their postulated
function, namely recognition by m e m b r a n e binding sites. In addition, the signal sequence should
contain information for its correct removal by the
processing enzyme(s). One could therefore envision altered signal sequences which will not be
recognized by m e m b r a n e binding sites, with the
result that transfer across the m e m b r a n e could not
take place. Likewise, an altered sequence may still
be recognized by the m e m b r a n e and result in
transfer of the protein across the m e m b r a n e but it
m a y not serve as a substrate for the processing
enzyme. The latter condition may be physiological
for some proteins (for instance those proteins
which need to retain the signal sequence for their
proper function), pathological for others. The
former condition, however~ may be entirely pathological, since it would not lead to transfer across
the m e m b r a n e .
We thank Dr. G. Palade for his helpful comments and
Mrs. N. Dwyer for preparing the illustrations for this and
the following paper.
This investigation was supported by Grant Number
CA 12413, awarded by the National Cancer Institute,
DHEW.
Received for publication 30 June 1975, and in revised
form 2 September 1975.
REFERENCES
1. ADELMAN, M. R., D. D. SABATINI, and G. BLOBEL.
1973. Ribosome-membrane interaction. Nondestructive disassembly of rat liver rough microsomes into
ribosomal and membranous components. J. Cell
Biol. 56:206-229.
2. Avlv, H., and P. LEOER. 1972. Purification of
biologically active globin mRNA by chromatography on oligothymidylic acid-cellulose. Proc. Natl.
,4cad. Sci. U. S. .4. 69:1408 1412.
3. BLOaEL, G., and V. R. POTTER. 1967. Studies on free
and membrane-bound ribosomes in rat liver. I.
Distribution as related to total cellular RNA. J. Mol.
Biol. 26:279-292.
4. BLOBEL, G., and D. D. SABATINI. 1970. Controlled
proteolysis of nascent polypeptides in rat liver cell
fractions. 1. Location of the polypeptides within
ribosomes. J. Cell. Biol. 45:130 145.
8~0
5. BLOBEL, G., and D. D. SABATINI. 1971. Ribosomemembrane interaction in eukaryotic cells. In Biomembranes. L. A. Manson, editor. Plenum Publishing Corporation, New York. 2:193-195.
6. BLOBEL,G., and D. D. SABATINI. 1971. Dissociation
of mammalian polyribosomes into subunits by puromycin. Proc. Natl. Acad. Sci. U. S. A.
68:390-394.
7. BORGESE,N., G. BLOBEL,and D. D. SABATINI. 1973.
In vitro exchange of ribosomal subunits between free
and membrane-bound ribosomes. J. Mol. Biol.
74:415-438.
8. BORGESE, N., W. MOK, G. KREIalCH, and D. D.
SABATINI. 1974. Ribosomal-membrane interaction:
in vitro binding of ribosomes to microsomal membranes. J. MoL Biol. 88:559-580.
9. BURKE, G. T., and C. M. REOMAN. 1973. The
distribution of radioactive peptides synthesized by
polysomes and ribosomal subunits combined in vitro
with microsomal membranes. Biochim. Biophys.
Acta. 299:312 324.
10. CIOLI, D., and E. S. LENNOX. 1973. Purification and
characterization of nascent chains from immunoglobulin
producing
cells.
Biochemistry.
12:3203-3210.
11. EAGLE, H. 1959. Amino acid metabolism in mammalian cell cultures. Science (Wash. D.C.).
130:432-437.
12. FARQUHAR, M. G., and G. E. PALADE. 1965. Cell
junctions in amphibian skin. J. Cell Biol.
/,6:263 291.
13. FREIENSTEIN, C., and G. BLOBEL. 1974. Use of
eukaryotic native small ribosomal subunits for the
translation of globin messenger RNA. Proc. Natl.
Acad. Sci. U. S. A. 71:3435-3439.
14. FRIDLENDER, B. R., and F. O. WETTSTEIN. 1970.
Differences in the ribosomal protein of free and
membrane bound polysomes of chick embryo cells.
Biochem. Biophys. Res. Commun. 39:247 253.
15. GREEN, M., P. N. GRAVES, T.-L”tEHAvI-WII.ENER,J.
MCINNES, and S. PESTKA. 1975. Cell-free translation
of immunoglobulin messenger RNA from
MOPC-315 plasmacytoma and MOPC-315 NR, a
variant synthesizing only light chain. Proc. Natl.
Acad. Sci. U. S. A. 72:224-228.
16. KELLEMS, R. E., V. F. ALLISON, and R. A. BuTow.
1975. Cytoplasmic type 80S ribosomes associated
with yeast mitochondria. IV. Attachment of ribosomes to the outer membrane of isolated mitochondria. J. Cell Biol. 65:1 14.
17. LODISH, H. F., D. HOVSMAN, and M. JACOBSEN.
1971. Initiation of hemoglobin synthesis. Specific
inhibition by antibiotics and bacteriophage ribonucleic acid. Biochemistry. 10:2348-2356.
18. Lurr, G. H. 1961. Improvements in epoxy embedding methods. J. Biophys. Biochem. Cytol.
9:263-291.
19. MACH, B., C. FAUST, and P. VASALLI. 1973. Purification of 14S messenger RNA of immunoglobulin
THE JOURNAL OF CELL BIOLOGY – VOLUME67, 1975
20.
21.
22.
23.
light chain that codes for a possible light-chain
precursor. Proc. Natl. Acad. Sci. U. S. A.
70:451-455.
MAIZEL, J. V. 1969. Acrylamide gel electrophoresis
of proteins and nucleic acids. In Fundamental Techniques in Virology. K. Habel and N. P. Salzman,
editors. Academic Press, Inc., New York. 334-362.
MANS, R. J., and G. D. NOVELLt. 1961. Measurement of the incorporation of radioactive amino acids
into protein by a filter paper disk method. Arch.
Biochem. Biophys. 94:48-53.
McCoNKEY, E. H., and E. J. HAUBEg. 1975. Evidence for heterogeneity of ribosomes within the
HeLa cell. J. Biol. Chem. 250:1311 – 1318.
MILSTt~IN,C., G. G. BaOWNLEE, T. M. HARRISON,
and M. B. MATHEWS. 1972. A possible precursor of
immunoglobulin light chains. Nature New Biol.
239:117-120.
24. PALADE, G. E. 1955. A small particulate component
of the cytoplasm. J. Biophys. Biochem. Cytol.
1:59-68.
25. PALADE, G. E. 1958. In Microsomal Particles and
Protein Synthesis. First Symposium of Biophysical
Society. R. B. Roberts, editor. Pergamon Press, Inc.,
Elmsford, N.Y. 36.
26. REDMAN, C. M., and D. D. SABATINI. 1966. Vectorial discharge of peptides released by puromycin from
attached ribosomes. Proc. Natl. Acad. Sci. U. S. A.
56:608-615.
27. REDMAN,C. M., P. SIEKEVITZ,and G. E. PALADE.
1966. Synthesis and transfer of amylase in pigeon
pancreatic microsomes. J.
Biol.
Chem.
241:1150-1158.
28. ROLLESTON, F. S. 1974. Membrane-bound and free
ribosomes. Sub-Cell. Biochem. 3:91-117.
29. SABATINi, D. D., and G. BLOBEL. 1970. Controlled
proteolysis of nascent polypeptides in rat liver cell
fractions, il. Location of the polypeptides in rough
microsomes. J. Cell Biol. 45:146-157.
30. SABATINI, D. D., N. BORGESE, M. ADELMAN, G.
KREIBICH, and G. BLOBEL. 1972. Studies on the
membrane associated protein synthesis apparatus of
eukaryotic cells. RNA Viruses and Ribosomes.
Noord-Hollandsche Vitg. Mij., Amsterdam. 147171.
31. SAUER, L. A., and G. N. BuRgow. 1972. The
submicrosomal distribution of radioactive proteins
released by puromycin from the bound ribosomes of
rat liver microsomes labeled in vitro. Biochim.
Biophys. Acta. 277:179-187.
32. SCHECHTEg, I. 1973. Biologically and chemically
pure mRNA coding for mouse immunoglobulin
L-chain prepared with the aid of antibodies and
immobilized oligothymidine. Proc. Natl. Acad. Sci.
U. S. A. 70:2256-2260.
33. SCHECHTER, I., D. J. MCKEAN, R. GUYER, and W.
TERRY. 1974. Partial amino acid sequence of the
precursor of immunoglobulin light chain programmed by messenger RNA in vitro. Science
(Wash. D.C.). 1~:160-162.
34. SCHMECKPEPER, B. J., S. CORY, and J. M. ADAMS.
1974. Translation of immunoglobulin mRNAs in a
wheat germ cell-free system. Mol. Biol. Rep.
1:355-363.
35. SWAN, D., H. Avlv, and P. LEDER. 1972. Purification and properties of biologically active messenger
RNA for a myeloma light chain. Proc. Natl. Acad.
Sci. U. S. A. 69:1967-1971.
36. TONEGAWA, S., and I. BALDI. 1973. Electrophoretically homogeneous myeloma light chain mRNA and
its translation in vitro. Biochem. Biophys. Res.
Commun. 51:81-87.
37. VENABLE, J. H., and R. COGGESHALL. 1965. A
simplified lead citrate stain for use in electron
microscopy. J. Cell Biol. 25:407-408.
38. WATSON, M. L. 1958. Staining of tissue sections for
electron microscopy with heavy metals. J. Biophys.
Biochem. Cytol. 4:475-478.
G. BLOBEL AND B. DOBBERSTEIN Transfer o f Proteins across Membranes. I
851
apoA-IV tagged with the ER retention signal KDEL
perturbs the intracellular trafficking and secretion
of apoB
James W. Gallagher,* Richard B. Weinberg,† and Gregory S. Shelness1,*
Department of Pathology* and Department of Internal Medicine and Physiology & Pharmacology,†
Wake Forest University School of Medicine, Winston-Salem, NC 27157
Abstract To examine the role of apolipoprotein A-IV
(apoA-IV) in the intracellular trafficking and secretion of
apoB, COS cells were cotransfected with microsomal triglyceride transfer protein (MTP), apoB-41 (amino terminal
41% of apoB), and either native apoA-IV or apoA-IV modified with the carboxy-terminal endoplasmic reticulum (ER)
retention signal, KDEL (apoA-IV-KDEL). As expected, apoAIV-KDEL was inefficiently secreted relative to native apoAIV. Coexpression of apoB-41 with apoA-IV-KDEL reduced
the secretion of apoB-41 by 80%. The apoA-IV-KDEL effect was specific, as neither KDEL-modified forms of human serum albumin or apoA-I affected apoB-41 secretion.
Similar results were observed in McA-RH7777 rat hepatoma
cells, which express endogenous MTP. The full inhibitory
effect of apoA-IV-KDEL on apoB secretion was observed
only for forms of apoB containing a minimum of the
amino-terminal 25% of the protein (apoB-25). However,
apoA-IV-KDEL inhibited the secretion of both lipid-associated and lipid-poor forms of apoB-25. Dual-label immunofluorescence microscopy of cells transfected with native
apoA-IV and apoB-25 revealed that both apolipoproteins
were localized to the ER and Golgi, as expected. However,
when apoA-IV-KDEL was cotransfected with apoB-25, both
proteins localized primarily to the ER. These data suggest
that apoA-IV may physically interact with apoB in the secretory pathway, perhaps reflecting a role in modulating the
process of triglyceride-rich lipoprotein assembly and secretion.—Gallagher, J. W., R. B. Weinberg, and G. S. Shelness.
apoA-IV tagged with the ER retention signal KDEL perturbs
the intracellular trafficking and secretion of apoB. J. Lipid
Res. 2004. 45: 1826–1834.
Supplementary key words lipoproteins • chylomicrons • lipid absorption • lipid transport • triglycerides • protein trafficking • apolipoprotein A-IV • apolipoprotein B • endoplasmic reticulum
Apolipoprotein A-IV (apoA-IV) is a 46 kDa plasma glycoprotein (1) that is synthesized by the mammalian intestine (2) during lipid absorption, incorporated into naManuscript received 17 May 2004 and in revised form 1 July 2004.
Published, JLR Papers in Press, July 16, 2004.
DOI 10.1194/jlr.M400188-JLR200
This is an open access article under the CC BY license.
scent chylomicrons (3), and secreted into the circulation
on the surface of lymph chylomicrons (4). Although a
broad spectrum of physiological functions have been proposed for apoA-IV (5, 6), the preponderance of evidence
suggests that its primary biological function is related to
intestinal lipid absorption. In humans, apoA-IV expression is restricted to the intestine and is specifically stimulated by triglyceride absorption (7–10). The secretion of
apoA-IV into mesenteric lymph rapidly increases during
fat absorption in parallel with lymph triglycerides (11).
Plasma apoA-IV levels increase after fat feeding (11–13)
and decrease during fasting (14). Moreover, plasma apoAIV levels are correlated with dietary fat intake (15) and are
depressed in digestive disorders that cause fat malabsorption (16). Finally, the absence of the entire apoA-I/apoCIII/apoA-IV gene complex (17), but not isolated absence
of the apoA-I and apoC-III genes (18), is associated with
fat-soluble vitamin malabsorption.
Chylomicron assembly is the final, essential step in intestinal lipid absorption (19), and several additional lines
of evidence specifically implicate apoA-IV in this process.
The hydrophobic surfactant Pluronic L-81 simultaneously
and selectively blocks both chylomicron assembly and intestinal apoA-IV synthesis but not the absorption of luminal fatty acids and their intracellular esterification (20).
Enterocyte apoA-IV mRNA and protein levels do not increase during absorption of short-chain fatty acids, which,
unlike long-chain fatty acids, are absorbed directly into
the portal blood and do not require chylomicron assembly (21). Plasma apoA-IV levels are decreased in subjects
with abetalipoproteinemia (4, 12) and hypobetalipoproteinemia (1), genetic disorders in which chylomicron assembly and secretion are impaired. Nonetheless, lipid absorption is grossly normal in apoA-IV knockout mice (22),
Abbreviations: apoA-IV, apolipoprotein A-IV; DSP, dithiobis(succinimidyl propionate); ER, endoplasmic reticulum; HSA, human serum albumin; MTP, microsomal triglyceride transfer protein.
1 To whom correspondence should be addressed.
e-mail: gshelnes@wfubmc.edu
Copyright © 2004 by the American Society for Biochemistry and Molecular Biology, Inc.
1826
Journal of Lipid Research Volume 45, 2004
This article is available online at http://www.jlr.org
suggesting that apoA-IV may play a facilitating or regulatory role in chylomicron assembly and intestinal lipid
transport. Recent studies by Lu et al. (23) support this hypothesis by demonstrating that apoA-IV expression stimulates transcellular triglyceride transport in neonatal pig intestinal epithelial cells.
Although many lines of evidence support a role of
apoA-IV in intestinal lipid absorption and perhaps chylomicron assembly, the mechanism underlying its intracellular function is unknown. We have proposed that the interfacial properties of apoA-IV enable it to regulate particle
expansion in the second stage of triglyceride-rich particle
assembly (24), during which small, HDL-sized nascent
particles acquire large amounts of additional triglyceride
(19, 25). To test the hypothesis that apoA-IV interacts with
apoB within the secretory pathway, we explored the consequences of altering apoA-IV intracellular trafficking by
modifying it with the carboxy-terminal endoplasmic reticulum (ER) retention signal, KDEL (26). This approach
has been shown previously to be useful for assessing intracellular protein-protein interactions (27–29). The present
studies have revealed that ER retention of apoA-IV specifically inhibits the trafficking of apoB. Hence, apoA-IV may
interact either directly or indirectly with nascent apoBcontaining lipoprotein particles early in the triglyceriderich particle assembly process.
MATERIALS AND METHODS
Expression plasmids
Human apoA-IV cDNA was produced by reverse transcriptasecoupled PCR using human small intestine total RNA (Clontech)
as a template and 5 and 3 apoA-IV flanking oligonucleotides as
primers. Human serum albumin (HSA) cloned into the human
cytomegalovirus (CMV) immediate early promoter-based expression plasmid pBAT14 was obtained from Dr. Peter Arvan (University of Michigan). Human apoA-I cDNA cloned into expression plasmid pCMV5 was obtained from Dr. Mary Sorci-Thomas
(Wake Forest University School of Medicine). Addition of the
tetrapeptide Lys-Asp-Glu-Leu (KDEL) to the carboxy-terminal
ends of apoA-IV, HSA, and apoA-I was achieved by standard PCRbased cloning techniques. Briefly, antisense PCR primers were
designed that hybridized to the carboxy-terminal 7–10 amino acids of each cDNA and also contained sequences encoding the
KDEL tetrapeptide followed by a termination codon (26). These
were used in combination with specific 5 sense strand primers to
produce the KDEL-modified forms of each open reading frame.
PCR products were cloned into the expression vector pCMV5
(30). The validity of each construct was confirmed by DNA sequence analysis. All apoB truncation mutants contained a carboxy-terminal FLAG (DYKDDDDK) epitope (31, 32), with the
exception, as noted, where carboxy-terminal 6 His-tagged constructs were used (33).
(Roche Applied Science) (31). Transfected cells were metabolically radiolabeled with 100 Ci/ml [35S]Met/Cys (EasyTag Express Protein Labeling Mix; Perkin Elmer Life Sciences) in
Met- and Cys-deficient DMEM (ICN) for the times indicated.
Protein from media and cell lysates was immunoprecipitated
with goat anti-human apoB (Academy Biomedical, Houston,
TX), rabbit anti-HSA (Roche Applied Science), goat anti-human
apoA-I (Academy Biomedical), rabbit anti-human apoA-IV (1),
or mouse anti-FLAG M2 antibody (Sigma), as indicated. Immunoprecipitations and SDS-PAGE were performed as described
(31). Dried gels were exposed to BioMax MS film in combination with a BioMax TransScreen-LE intensifying screen (Kodak)
at 70C. For pulse-chase studies, COS-1 cells in 150 mm dishes
were cotransfected with equal mass quantities of the indicated
DNAs (15 g total). Twenty-four hours after transfection, cells
were trypsinized and replated at 50% confluence in 60 mm
dishes. Twenty-four hours after replating, cells were pulse radiolabeled with [35S]Met/Cys for 10 min and then chased with media
containing an excess of cold Met and Cys for the times indicated
(35). apoB from media and cell lysates was immunoprecipitated
and fractionated by SDS-PAGE, and bands were quantified using
a Molecular Dynamics 445 SI Phosphorimager (36).
In situ cross-linking of apoA-IV and apoB-25
COS-1 cells in 100 mm dishes were transfected, using Fugene-6, with equal mass quantities of either apoB-25 and apoA-IV
or apoB-25 and HSA-KDEL. Twenty-four hours after transfection,
cells were metabolically radiolabeled with [35S]Met/Cys for 2 h.
After washing cell monolayers with PBS, cells were incubated
on ice for 30 min with either 10 ml of PBS or 10 ml of PBS
containing 200 M dithiobis(succinimidyl propionate) (DSP;
Pierce) (37). After adjusting cells to 50 mM Tris-HCl, pH 8.2, to
inactivate unreacted DSP, monolayers were washed with PBS
and the cells were lysed as described above. Lysates were divided into equal aliquots and immunoprecipitated with antiapoA-IV or anti-HSA antibodies, as indicated. Before gel loading, samples were boiled in SDS-PAGE sample buffer containing
100 mm DTT.
Immunofluorescence of intracellular apoA-IV and apoB
COS-1 cells in 3 cm dishes were transfected with equal mass
quantities of apoB-25 and either apoA-IV-KDEL or HSA-KDEL
(1.5 g of total DNA) using Fugene-6. Twenty-four hours after
transfection, cells were fixed in 3.7% formaldehyde in PBS for 10
min and permeabilized with 0.1% Saponin in PBS (PBS-Saponin). Fixed cells were incubated with 1% BSA in PBS-Saponin for
30 min, followed by 30 min with primary antibodies in the same
buffer at the following dilutions: mouse anti-FLAG monoclonal
antibody M2, 12.5 g/ml; rabbit anti-HSA, 1:400; rabbit antiapoA-IV, 1:300. Cells were then incubated with rhodamine-conjugated goat anti-mouse IgG and FITC-conjugated goat anti-rabbit
IgG secondary antibodies (Jackson ImmunoResearch) at concentrations of 25 g/ml in PBS-Saponin containing 1% BSA.
Cells were postfixed, mounted in 90% glycerol, and viewed using
a Zeiss Axioplan 2 microscope with a 63 oil objective. Images
were captured with a Zeiss Axiocam using a gain setting of 3.
RESULTS
Transfection, metabolic labeling, and
immunoprecipitation
COS-1 cells in 100 mm dishes were transfected at 50–60% confluence with equal mass quantities of the indicated DNA plasmids (6 g total DNA) using the DEAE-Dextran method (34).
McA-RH7777 cells in 100 mm dishes were transfected with a total
of 10 g of plasmid DNA using Fugene-6 transfection reagent
apoA-IV-KDEL selectively reduces the secretion
of apoB-41
COS cells were cotransfected with apoB-41 and microsomal triglyceride transfer protein (MTP) and one of
the following: HSA, HSA-KDEL, apoA-IV, or apoA-IV-KDEL.
Gallagher, Weinberg, and Shelness apoA-IV-KDEL inhibits apoB secretion
1827
Fig. 1. KDEL-modified apolipoprotein A-IV (apoA-IV) inhibits
apoB-41 secretion. A and B: COS cells were cotransfected with 2 g
each of apoB-41 and microsomal triglyceride transfer protein
(MTP) and one of the following: human serum albumin (HSA),
HSA-KDEL, apoA-IV (AIV), or apoA-IV-KDEL (AIV-KDEL), as indicated. Cells were radiolabeled with [35S]Met/Cys for 3 h, and equal
aliquots of cell lysates (C) and media (M) were subjected to immunoprecipitation with anti-HSA (A, lanes 1–4), anti-apoA-IV (A,
lanes 5–8), or anti-apoB (B, lanes 1–8) antibodies. Immune complexes were analyzed by SDS-PAGE and fluorography. C: Parallel
dishes of cotransfected cells were pulse radiolabeled with
[35S]Met/Cys for 10 min and then chased with medium containing
excess cold Met and Cys for 0 or 120 min. After immunoprecipitation with anti-apoB antibodies, the mean percentage of newly synthesized apoB (cell-associated protein after the 10 min pulse) secreted into medium during the 120 min chase ( SD) was
calculated based on Phosphorimager analysis of dried gels. Statistically significant differences in secretion efficiencies are indicated
by different lower case letters (ANOVA, P 0.0001; Tukey/Kramer
posthoc analysis; n 3).
As expected, KDEL modification of both HSA and apoAIV markedly inhibited their secretion (Fig. 1A, compare
lanes 2 and 4 and lanes 6 and 8). HSA-KDEL and apoA-IV
had little impact on apoB-41 secretion relative to the HSA
control (Fig. 1B, lanes 1–6). In contrast, apoA-IV-KDEL
virtually eliminated apoB-41 secretion (Fig. 1B, lane 8).
1828
Journal of Lipid Research Volume 45, 2004
To quantitate the impact of the KDEL-modified proteins
on apoB secretion, cotransfected COS cells were pulse radiolabeled with medium containing [35S]Met/Cys for 10
min and then chased with medium containing an excess
of cold Met and Cys for 0 or 120 min. Native apoA-IV reduced apoB-41 secretion by 45% relative to both HSA
and HSA-KDEL, whereas apoA-IV-KDEL inhibited apoB41 secretion by greater than 80% (Fig. 1C).
The perturbing effect of apoA-IV-KDEL was specific for
assembly-competent forms of apoB, as apoB-6.6, a highly
truncated apoB that lacks the capacity to form nascent lipoproteins (31, 33), was unaffected by either apoA-IV or
apoA-IV-KDEL coexpression (Fig. 2A). We also tested
whether KDEL modification of another lipid binding protein, apoA-I, could similarly affect apoB secretion. As observed in Fig. 2B, KDEL modification of apoA-I markedly
reduced its secretion; however, there was no corresponding effect on apoB-41 secretion (Fig. 2C). Finally, to address the possibility that apoA-IV-KDEL artifactually reduced the expression of MTP in the cotransfected COS
cells, we tested whether the apoA-IV-KDEL-mediated inhibition of apoB secretion could be reproduced in McARH7777 rat hepatoma cells, which express endogenous
MTP. Results of cotransfection of apoB-34 (another assembly-competent form of apoB) (33) with native or KDELmodified apoA-IV are displayed in Fig. 2D. As observed
previously for apoB-41 in COS cells (Fig. 1), apoA-IVKDEL severely reduced the secretion of apoB-34 in McARH7777 cells (Fig. 2D, compare lanes 2, 4, and 6 with lane
8). Although it might be expected that the transfected
apoA-IV-KDEL would also reduce the secretion of endogenous apoB-48 and perhaps apoB-100 in McA-RH7777
cells, this was not observed (data not shown). This may be
because the relatively low transfection efficiency (20%
of cells) masks the impact of the transfected apoA-IVKDEL. Alternatively, it is possible that the observed interaction between human apoA-IV and human apoB is species specific, an issue currently under study.
apoA-IV-KDEL inhibits apoB secretion independently of
its lipidation state
Approximately 75% of apoB-25 coexpressed with MTP
in COS cells is assembled into a buoyant lipoprotein particle that floats at d 1.25 g/ml (33). To determine
whether the ability of apoA-IV-KDEL to perturb apoB secretion is dependent on apoB’s lipidation state, apoB-25
and MTP were cotransfected along with apoA-IV or apoAIV-KDEL into COS cells. Density gradient centrifugation
was used to separate cell medium into lipoprotein-containing (d 1.25 g/ml) and lipid-poor (d 1.25 g/ml)
fractions, as described (33). Relative to HSA-KDEL, apoAIV-KDEL abolished the secretion of apoB-25 into both the
lipoprotein and lipid-poor density fractions (Fig. 3, compare lanes 2 and 3 with lanes 5 and 6). This result favors
the interpretation that apoA-IV-KDEL can interact directly
with apoB. To confirm this finding, the experiment was repeated in the absence of MTP coexpression. Even without
the expression of MTP to induce lipoprotein assembly,
both apoA-IV and apoA-IV-KDEL reduced the efficiency
Fig. 2. Specificity of apoA-IV-KDEL-mediated inhibition of apoB secretion. A: COS cells were cotransfected
with apoB-6.6 and one of the following: HSA, HSA-KDEL, apoA-IV, or apoA-IV-KDEL, as indicated. Cells
were radiolabeled with [35S]Met/Cys for 3 h, and cell lysate (C) and media (M) samples were subjected to
immunoprecipitation with anti-apoB antibodies followed by SDS-PAGE and fluorography. B and C: COS cells
were cotransfected with apoB-41 and MTP and one of the following: HSA-KDEL, apoA-I (AI), or apoA-IKDEL (AI-KDEL), as indicated. Cells were radiolabeled with [ 35S]Met/Cys for 3 h, and equal aliquots of cell
lysate and media samples were subjected to immunoprecipitation with anti-apoA-I (B) or anti-apoB (C) antibodies. D: McA-RH7777 cells were cotransfected with apoB-34 and one of the following: HSA, HSA-KDEL,
apoA-IV, or apoA-IV-KDEL, as indicated. Twenty-four hours after transfection, cells were radiolabeled with
[35S]Met/Cys for 3 h, and cell media and lysates were subjected to immunoprecipitation with anti-FLAG antibody.
of apoB-25 secretion to a level similar to that observed for
apoB-41 (Fig. 4; see also Fig. 1).
Effect of apoB carboxy-terminal truncation on apoA-IVKDEL-mediated inhibition of apoB secretion
We previously established that a very narrow interval in
the
1 domain (38) of apoB, between residues 884
(apoB-19.5) and 912 (apoB-20.1), completes a sequence
fully capable of initiating the assembly of small emulsionlike triglyceride-rich lipoproteins (33). To examine the relationship between the structural requirements for particle assembly and the ability of apoA-IV-KDEL to perturb
apoB secretion, a panel of carboxy-terminal truncated
apoB constructs ranging from apoB-19 to apoB-25 was
cotransfected into COS cells along with either apoA-IVKDEL or HSA-KDEL. The secretion efficiency of each
construct in the presence of apoA-IV-KDEL was compared
with the efficiency in the presence of HSA-KDEL. Neither
apoB-19 nor apoB-21 secretion was affected by apoA-IVKDEL (Fig. 5); apoA-IV-KDEL caused an 75% inhibition
of apoB-23 secretion, whereas apoB-25 was inhibited to
the same extent observed previously (Fig. 4). These data
indicate that apoA-IV may interact directly with apoB at a
site that includes a domain positioned between apoB-21
and apoB-25.
apoA-IV-KDEL alters the cellular distribution of apoB
To examine whether apoA-IV-KDEL causes a redistribution of apoB within the secretory pathway, the intracellular localization of apoB-25 in apoA-IV- or apoA-IV-KDEL-
transfected COS cells was examined by dual-label immunofluorescence microscopy. When transfected together,
both apoB-25 (Fig. 6A) and apoA-IV (Fig. 6B) displayed
diffuse cytoplasmic and prominent heminuclear Golgi
staining (arrows), suggesting that both are colocalized to
the ER and Golgi compartments (Fig. 6C). In contrast,
Fig. 3. apoA-IV-KDEL inhibits the secretion of buoyant and lipidpoor forms of apoB-25. COS cells were cotransfected with apoB-25
and MTP and either HSA-KDEL or apoA-IV KDEL, as indicated.
Cells were metabolically radiolabeled with [ 35S]Met/Cys for 3 h,
and media (M) samples were subjected to density gradient centrifugation to obtain d 1.25 g/ml buoyant lipoprotein top (T) and
d 1.25 g/ml lipid-poor bottom (B) fractions (33). apoB-25 from
gradient fractions was recovered by immunoprecipitation with antiapoB antibodies and analyzed by SDS-PAGE and fluorography. C,
cell lysate.
Gallagher, Weinberg, and Shelness apoA-IV-KDEL inhibits apoB secretion
1829
Fig. 4. apoA-IV-KDEL inhibits the secretion of apoB-25 produced
in the absence of MTP. A: COS cells were transfected with apoB-25
and one of the following: HSA, HSA-KDEL, apoA-IV, or apoA-IVKDEL, as indicated. Cells were radiolabeled with [ 35S]Met/Cys for
3 h, and cell lysates (C) and media (M) were subjected to immunoprecipitation with anti-apoB antibodies followed by SDS-PAGE and
fluorography. B: Parallel dishes of cotransfected cells were subjected to pulse-chase analysis as described for Fig. 1C. The mean
percentage of newly synthesized apoB-25 secreted into media during the 120 min chase ( SD) under each condition was calculated.
Statistically significant differences in secretion efficiencies are indicated by different lower case letters (ANOVA, P 0.0001; Tukey/
Kramer posthoc analysis; n 3).
apoA-IV-KDEL displayed only diffuse cytoplasmic and perinuclear staining, consistent with predominant ER localization (Fig. 6E) (39). As predicted by the secretion results,
apoA-IV-KDEL changed the intracellular distribution of cotransfected apoB-25 from the ER-Golgi distribution observed in Fig. 6A to the predominantly ER localization observed in Fig. 6D, Hence, it appears that apoA-IV-KDEL
alters the distribution of apoB within the secretory pathway.
Control studies revealed that HSA-KDEL did not cause a redistribution of apoB-25 to the ER (Fig. 6G).
In situ cross-linking of apoA-IV and apoB-25
As another means of demonstrating an interaction between apoA-IV and apoB, in situ cross-linking was performed. COS cells were cotransfected with apoB-25 and
either apoA-IV or HSA-KDEL. After metabolic radiolabeling, cell monolayers were incubated with PBS or PBS containing 200 M of the reversible cross-linker DSP, as described (37). After inactivation of DSP, cells were lysed
and subjected to immunoprecipitation with antibodies to
1830
Journal of Lipid Research Volume 45, 2004
Fig. 5. Effect of carboxy-terminal truncation of apoB on apoA-IVKDEL-mediated inhibition of apoB secretion. COS cells were
cotransfected with the indicated apoB carboxy-terminal 6 Histagged truncation mutant and either HSA-KDEL or apoA-IV-KDEL.
A: Cells were radiolabeled for 3 h with [ 35S]Met/Cys, followed by
immunoprecipitation of media (M) and cell lysates (C) with antiapoB antibodies. B: Parallel dishes of transfected cells were pulse
radiolabeled for 10 min and chased for 120 min. The percentage of
each construct secreted during the 120 min chase in the presence
of apoA-IV-KDEL was expressed as a mean percentage ( SD; n
3) of the secretion observed in HSA-KDEL-transfected cells.
either apoA-IV or HSA. As observed in Fig. 7, lane 1, only
a small amount of apoB-25 was coimmunoprecipitated
with apoA-IV antibodies in the absence of cross-linker.
This may be attributable to the weak and/or transient nature of the intermolecular interaction between the proteins, which is destabilized by detergent lysis. However, the
addition of the cross-linking reagent DSP before cell lysis
resulted in considerable coimmunoprecipitation of apoB25 with anti-apoA-IV antibody (Fig. 7, lane 2). When apoAIV was replaced with the control protein HSA-KDEL, only
background levels of cross-linking to apoB-25 were observed (Fig. 7, lane 4). These experiments provide additional evidence that apoA-IV and apoB can interact intracellularly.
Secretion kinetics of apoA-IV
KDEL modification of apoA-IV induced an exaggerated
redistribution of both apoA-IV and cotransfected apoB.
Fig. 6. apoA-IV-KDEL alters the intracellular distribution of apoB-25. COS cells were cotransfected with
apoB-25 and one of the following: apoA-IV (A–C), apoA-IV-KDEL (D–F), or HSA-KDEL (G). Cells were fixed
and immunostained with goat anti-apoB and rabbit anti-apoA-IV (A–F) or goat anti-apoB and rabbit antiHSA (G) antibodies. Cells were then stained with rhodamine-conjugated anti-goat IgG and fluorescein-conjugated anti-rabbit IgG. A, D, and G show rhodamine fluorescence (red); B and E show fluorescein fluorescence (green); C and F show the overlap (yellow) between rhodamine and fluorescein fluorescence. Arrows
indicate Golgi staining.
However, we also observed that native apoA-IV had a partial inhibitory effect on apoB secretion (Figs. 1, 4), perhaps because of apoA-IV’s inherently slow secretion rate.
To examine the secretion kinetics of apoA-IV relative to
the more generic secretory protein HSA, transfected COS
cells were pulse radiolabeled for 10 min and chased for
0–240 min. For HSA, 50% of the newly synthesized protein was secreted within the first 30 min of chase, and the
overall secretion efficiency approached 90% (Fig. 8, closed
circles). However, native apoA-IV displayed a much slower
rate of secretion, with less than 20% secreted after the 240
min chase (Fig. 8, open circles).
DISCUSSION
Considerable correlative and some direct evidence supports a role of apoA-IV in intestinal lipid transport. Most
recently, Lu et al. (23) demonstrated that expression of
apoA-IV in IPEC-1 newborn swine intestinal epithelia cells
markedly stimulated triglyceride transport in chylomicron
particles without affecting the expression of other proteins implicated in lipid transport or metabolism. In the
current report, we explored the hypothesis that the effect
of apoA-IV on intestinal lipid transport involves a direct or
indirect intracellular interaction between apoA-IV and
apoB. To test this hypothesis, apoA-IV was modified with
the carboxy-terminal ER retention signal KDEL (26) and
the potential impact on apoB trafficking was examined.
These studies revealed that intracellular retention of apoAIV caused by KDEL modification resulted in a specific
80% reduction in the secretion of apoB-41. Native apoAIV, which displayed relatively slow secretion kinetics, even
without KDEL modification, also delayed apoB secretion.
Truncation analysis demonstrated that a minimum of
the amino-terminal 25% of apoB is required for apoA-IVKDEL to inhibit secretion. Finally, the effects of apoA-IVKDEL on apoB trafficking occurred even in the absence
of MTP, suggesting that apoA-IV may have the capacity to
interact directly with apoB, as demonstrated by cross-linking analysis.
The ability to interact with apoB and perturb its movement within the secretory pathway suggests a possible
mechanism by which apoA-IV may enhance intestinal
lipid transport. In the first stage of intestinal triglyceriderich particle assembly, apoB-48 is cotranslationally lipidated with a small amount of phospholipid and triglyceride by MTP to form small, HDL-sized nascent particles
Gallagher, Weinberg, and Shelness apoA-IV-KDEL inhibits apoB secretion
1831
Fig. 7. In situ cross-linking of apoA-IV and apoB-25. COS cells
were cotransfected with apoB-25 and either apoA-IV or HSA-KDEL,
as indicated. After metabolic radiolabeling, cell monolayers were
treated without () or with ( ) dithiobis(succinimidyl propionate) (DSP), followed by detergent lysis and immunoprecipitation
with anti-apoA-IV (lanes 1 and 2) or anti-HSA (lanes 3 and 4) antibodies. Samples were then heated in the presence of reducing
agent to break cross-links and analyzed by SDS-PAGE and fluorography.
(40). The absence or inhibition of MTP blocks this first
stage of assembly (41). In the second stage of assembly,
nascent chylomicron particles, which already have apoAIV on their surface (42), acquire large amounts of additional triglyceride and expand to diameters of 500–1,000
nm before being secreted from the enterocyte basolateral
membrane. Although MTP is believed responsible for
the trafficking of lipid into the secretory pathway for second step expansion (43–45), the mechanism by which
the resulting lipid droplets are incorporated into nascent
apoB-containing particles is unknown. Our studies raise
the possibility that apoA-IV may function as a modulatory
cofactor to reduce apoB’s rate of intracellular transport,
thereby increasing its residence time within a lipoprotein
expansion compartment. The exact intracellular compartments involved in second step expansion are under active
investigation but appear to involve the ER (25, 46) and/or
Golgi (47–49). Because the capacity to enlarge nascent
lipoproteins is the predominant means by which the intestine accommodates increased lipid flux, apoA-IV, via its
trafficking effects on apoB, may play an important role in
this process, particularly under conditions of high dietary
fat intake.
The validity of the hypothesis that apoA-IV may modulate the trafficking of apoB within the secretory pathway
requires that the intracellular apoA-IV/apoB molar ratio
be sufficiently high to ensure that each nascent apoB-containing lipoprotein interact with one or more apoA-IV
molecule. Indeed, studies in rat and swine enterocytes reveal that the intracellular apoA-IV/apoB-48 ratio ranges
from 12 to 19 (50–52). Importantly, this ratio is also ob1832
Journal of Lipid Research Volume 45, 2004
Fig. 8. Kinetics of apoA-IV and HSA secretion. COS cells transfected with apoA-IV (open circles) or HSA (closed circles) were
pulse radiolabeled with [ 35S]Met/Cys for 10 min and chased for
the times indicated. After each chase time, the percentage of protein recovered in cells and media was quantified by immunoprecipitation, SDS-PAGE, and Phosphorimager analysis. The mean percentage of newly synthesized protein that was recovered in the
media after each time point was plotted SD (n 3).
served in the cotransfection studies performed here. Inspection of the electrophoretic band intensities corresponding to apoA-IV and apoB indicate that, at steady
state, there is 2- to 3-fold more radioactivity incorporated into apoA-IV or apoA-IV-KDEL than there is into
apoB-41 (e.g., compare the band intensities in Fig. 1A, B,
lanes 5 and 7). Taking into account the different Met content of each protein, the apoA-IV/apoB ratio in the transfected cells is on the order of 16, a value within the range
observed in enterocytes. Hence, we propose that the relative levels of apoB and apoA-IV expression achieved in our
transfected cell system are comparable to those achieved
in vivo and are consistent with the ability of apoA-IV to
modulate apoB trafficking. Furthermore, it appears that
the observed effects of apoA-IV on apoB trafficking cannot be attributed to overexpression per se, as comparable
expression levels of control proteins, including HSA,
HSA-KDEL (Fig. 1), and the lipid binding proteins apoA-I
and apoA-I-KDEL (Fig. 2), had no impact on the trafficking of cotransfected apoB.
The underlying basis for the observed apoB-apoA-IV interactions observed in the present report appears to be
mediated, at least in part, by protein-protein interactions.
However, the possibility that an interaction can also arise
by a hydrophobic interaction between apoA-IV and the
lipid interface of nascent apoB-containing lipoprotein
particles cannot be ruled out. The latter theory arose
from studies of the dynamic interfacial properties of apoAIV, which noted that the ability of apoA-IV to decrease surface tension while increasing interfacial elasticity is ideally
suited to meet the thermodynamic requisites of expanding lipid emulsion particles in an aqueous substrate (24,
53). In the present study, this mechanism is favored by the
finding that the apoA-IV-KDEL inhibition effect was seen
primarily with the apoB truncations that undergo substantial lipidation, i.e., apoB-25 and higher (33). Conversely, the findings that apoA-IV-KDEL attenuated the
secretion of both lipid-associated and lipid-poor apoB25 in the presence MTP and lipid-poor apoB-25 in the
absence of MTP strongly argue for a direct protein-protein interaction with apoB at some site that includes
residues between amino acids 953 (apoB-21) and 1,134
(apoB-25).
A role of apoA-IV in modulating intestinal lipid absorption would at first appear inconsistent with two previous
studies in apoA-IV knockout (22) and human apoA-IV
transgenic (54) mice, which found no effect of apoA-IV
expression on postprandial triglyceride-rich lipoprotein
kinetics or fat-soluble vitamin absorption. However, it is
critical to note that those studies measured these parameters after a single fat bolus in animals that had been maintained on a chow diet. Thus, the maximal triglyceride absorptive capacity of these animals was not achieved, and
the ability of apoA-IV to modulate absorption of higher dietary fat loads could not be ascertained. Indeed, demonstration of the physiological impact of apoA-IV expression
on the efficiency of intestinal lipid absorption will likely
require fat balance studies, which integrate fat absorption
over a longer time period. The need for this approach is
exemplified by a study in the Mdr2 knockout mouse, in
which biliary lipid secretion is impaired: no difference in
single-bolus plasma triglyceride kinetics was found between control and Mdr / mice, whereas fat balance studies demonstrated a significant decrease in fat absorption
in Mdr / mice, but only on a high-fat diet (55).
Given that apoA-IV might modulate intestinal lipid absorption efficiency under conditions of high dietary fat intake, it is interesting that the apoA-IV T347S and Q360H
polymorphisms, which are known to have an impact upon
protein structure (56), postprandial triglyceride metabolism (57), and cholesterol absorption (58), have been
found in epidemiological studies to be associated with a
lower body mass index (Q360H) (5, 59) or increased body
mass index and adiposity (T347S) (5). This raises the possibility that these genetic polymorphisms might affect intestinal lipid absorption and thus could have important
implications for the functional genomics of obesity.
In summary, apoA-IV partially and apoA-IV-KDEL almost completely inhibits the secretion of both lipoprotein-associated and lipid-poor apoB constructs equal to or
larger than apoB-25. This effect appears to be mediated
by a delay in intracellular trafficking attributable to a protein-protein interaction between apoA-IV and a domain
near the amino terminus of apoB. These data support the
hypothesis that apoA-IV can interact with apoB to achieve
enhancement of intestinal triglyceride-rich lipoprotein
expansion and suggest that, under certain dietary conditions, apoA-IV could modulate intestinal lipid absorption
efficiency.
This work was supported by National Institutes of Health
Grants HL-49373 (G.S.S.) and HL-30897 (R.B.W.). J.W.G. was
supported by a predoctoral fellowship from the American
Heart Association, Mid-Atlantic Affiliate. The authors thank Li
Hou for technical assistance.
REFERENCES
1. Weinberg, R. B., and A. M. Scanu. 1983. Isolation and characterization of human apolipoprotein A-IV from lipoprotein-depleted
serum. J. Lipid Res. 24: 52–59.
2. Weisgraber, K. H., T. P. Bersot, and R. W. Mahley. 1978. Isolation
and characterization of an apoprotein from the d less than 1.006
lipoproteins of human and canine lymph homologous with the rat
A-IV apoprotein. Biochem. Biophys. Res. Commun. 85: 287–292.
3. Hayashi, H., D. F. Nutting, K. Fujimoto, J. A. Cardelli, D. Black,
and P. Tso. 1990. Transport of lipid and apolipoproteins A-I and
A-IV in intestinal lymph of the rat. J. Lipid Res. 31: 1613–1625.
4. Green, P. H., R. M. Glickman, C. D. Saudek, C. B. Blum, and A. R.
Tall. 1979. Human intestinal lipoproteins. Studies in chyluric subjects. J. Clin. Invest. 64: 233–242.
5. Lefevre, M., J. C. Lovejoy, S. M. DeFelice, J. W. Keener, G. A. Bray,
D. H. Ryan, D. H. Hwang, and F. L. Greenway. 2000. Common apolipoprotein A-IV variants are associated with differences in body
mass index levels and percentage body fat. Int. J. Obes. Relat. Metab.
Disord. 24: 945–953.
6. Kalogeris, T. J., M. D. Rodriguez, and P. Tso. 1997. Control of synthesis and secretion of intestinal apolipoprotein A-IV by lipid. J.
Nutr. 127 (Suppl.): 537–543.
7. Apfelbaum, T. F., N. O. Davidson, and R. M. Glickman. 1987. Apolipoprotein A-IV synthesis in the rat intestine: regulation by dietary
triglyceride. Am. J. Physiol. 252: G662–G666.
8. Go, M. F., G. Schonfeld, B. Pfleger, T. G. Cole, N. L. Sussman, and
D. H. Alpers. 1988. Regulation of intestinal and hepatic apoprotein synthesis after chronic fat and cholesterol feeding. J. Clin. Invest. 81: 1615–1620.
9. Ktistaki, E., J. M. Lacorte, N. Katrakili, V. I. Zannis, and I. Talianidis. 1994. Transcriptional regulation of the apolipoprotein A-IV
gene involves synergism between a proximal orphan receptor response element and a distant enhancer located in the upstream
promoter region of the apolipoprotein C-III gene. Nucleic Acids
Res. 22: 4689–4696.
10. Vergnes, L., T. Taniguchi, K. Omori, M. M. Zakin, and A. Ochoa.
1997. The apolipoprotein A-I/C-III/A-IV gene cluster: apoC-III
and apoA-IV expression is regulated by two common enhancers.
Biochim. Biophys. Acta. 1348: 299–310.
11. Green, P. H., R. M. Glickman, J. W. Riley, and E. Quinet. 1980. Human apolipoprotein A-IV. Intestinal origin and distribution in
plasma. J. Clin. Invest. 65: 911–919.
12. Bisgaier, C. L., O. P. Sachdev, L. Megna, and R. M. Glickman. 1985.
Distribution of ap…
Order an Essay Now & Get These Features For Free:
Turnitin Report
Formatting
Title Page
Citation
Outline
