| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Biol Chem, Vol. 275, Issue 20, 15207-15219, May 19, 2000
,,¶
From the Department of Cell and Animal Biology,
Institute of Life Sciences, Hebrew University of Jerusalem, Jerusalem
91904 and the § Department of Electronics, Jerusalem College
of Technology, Jerusalem 91160, Israel
 |
ABSTRACT |
|---|
 TOP
 ABSTRACT
 INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES
|
|---|
The compartments involved in polarized exocytosis
of membrane proteins are not well defined. In this study we
hypothesized that newly synthesized polymeric immunoglobulin receptors
are targeted from the trans-Golgi network to endosomes
prior to their appearance on the basolateral cell surface of polarized
Madin-Darby canine kidney cells. To examine this hypothesis, we have
used an assay designed to measure the meeting of newly synthesized receptors with a selective population of apical or basolateral endosomes loaded with horseradish peroxidase. We found that in the
course of basolateral exocytosis, the wild-type polymeric immunoglobulin receptor is targeted from the trans-Golgi
network to apical and basolateral endosomes. Phosphorylation of a Ser residue in the cytoplasmic tail of the receptor is implicated in this
process. The biosynthetic pathway of apically sorted polymeric immunoglobulin receptor mutants similarly traversed apical endosomes, raising the possibility that apical receptors are segregated from basolateral receptors in apical endosomes. The post-endocytic pathway
of transcytosing and recycling receptors also passed through apical
endosomes. Together, these observations are consistent with the
possibility that the biosynthetic and endocytic routes merge into
endosomes and justify a model suggesting that endosomal recycling
processes govern polarized trafficking of proteins traveling in both pathways.
Epithelial cells carry out a variety of vectorial transport and
secretory processes that profoundly depend upon the polarized distribution of proteins and lipids on their cell surface. The plasma
membrane of epithelial cells can generally be divided into two domains,
an apical surface that faces the external environment and a basolateral
surface contacting adjacent cells and underlying tissues. The two
surface domains display a completely different protein and lipid
composition. The precise intracellular compartment(s) in which proteins
destined for the apical and basolateral cell surface are sorted is not
certain. It is commonly believed that the combined ability of endosomes
and the trans-Golgi network (TGN)1 to target a membrane
protein to a particular surface domain plays a role in determining its
polarized steady-state distribution on the plasma membrane (1-4).
The TGN is thought to serve as a major sorting site for plasma membrane
proteins traveling in the biosynthetic pathway (1, 5). This notion is
based primarily on observations of a common intracellular localization
of apical and basolateral proteins throughout the Golgi stacks (6) and
the TGN (7). However, these data do not rule out the possibility that
sorting of newly synthesized proteins can take place at later
compartments, for instance, in endosomes (1, 2, 4, 8). In this case, membrane proteins would be delivered directly from the TGN to endosomes
prior to their appearance on the cell surface. Results from recent
studies in nonpolarized cells are consistent with this notion (9, 10).
For example, Futter et al. (10) showed that the transferrin
receptor (TfnR) is targeted first from the TGN to endosomes and is
subsequently delivered to the plasma membrane. More recently, Brachet
et al. (11) have suggested that the biosynthetic pathway of
MHC class II molecules traverses early endosomes and that the passage
through the endocytic compartment is required to gain access to
peptide-loading compartments. In contrast, the biosynthetic pathway of
other membrane proteins, such as glycosylphosphatidyl inositol-linked
proteins, seem to bypass endosomes (10). Hence, diverse pathways
(mechanisms) can direct membrane proteins from the TGN to the plasma
membrane in nonpolarized cells. Whether similar direct and indirect
pathways exist in polarized cells is currently unknown.
Here we analyzed the polarized exocytic transport of the polymeric
immunoglobulin receptor (pIgR) exogenously expressed in polarized
Madin-Darby canine kidney (MDCK) cells (12). The current model suggests
that the pIgR is synthesized in the endoplasmic reticulum, from where
it is targeted to the Golgi complex. From the TGN the pIgR is directed
to the basolateral surface, where it can bind polymeric
immunoglobulins, such as dimeric IgA (dIgA) or IgM, and internalize via
clathrin-mediated endocytosis into endosomes. The pIgR bound to its
ligand then transcytoses from the basolateral to the apical pole of the
cell. The process of pIgR transcytosis utilizes several endosomal
compartments, each of which is presumed to play a role in its
trafficking to the apical plasma membrane. Initialy, upon
internalization from the basolateral plasma membrane, dIgA-pIgR
complexes reach basolateral early (sorting) endosomes (BEE). BEE serve
as the first station for many other endocytic ligands internalized from
the basolateral surface (e.g. Tfn, low density lipoprotein,
and epidermal growth factor bound to their receptors (13)). These
ligands exist at high concentrations in BEE (see Ref. 14). Next,
pIgR-dIgA complexes are transported to a compartment that consists of
long, 60-nm diameter tubules lying parallel to microtubules and
oriented toward the apical cytoplasm (15). The "tubular
compartment" contains both IgA and transferrin at the same relative
concentrations as the BEE but is substantially depleted of low density
lipoprotein and epidermal growth factor receptor, which are destined
for the degradative pathway (13, 14). This compartment, which resides in the medial and apical cytoplasm of polarized MDCK and Caco-2 cells,
has been designated the "interconnected compartment" or the
"common endosomal compartment" (CE) (14-18). The final compartment in the transcytotic pathway is comprised of 100-150-nm cup-shaped vesicles distributed underneath the apical plasma membrane (13). This
compartment is likely the equivalent of the pericentriolar apical
recycling endosomal compartment (ARE), whose structure depends upon
microtubules (19, 20).
Electron microscopic and quantitative immunofluorescence localization
data suggest that ARE are enriched in the transcytotic marker, the
pIgR, but are relatively depleted in the recycling marker, the human
TfnR (hTfnR) (13, 14). The endosomal rabs, rab11a, rab17, and rab25 are
associated with this compartment (21, 22). Fluid phase ligands
internalized from the apical plasma membrane reach apical early
endosomes (AEE) but not ARE or the CE (19). ARE and CE, however, can be
reached by membrane-bound endocytic ligands such as dIgA, Tfn, or the
general endocytic tracer ricin (13, 20, 23), suggesting that CE, ARE,
and AEE are distinct compartments. Apical endosomes have become the focus of intense interest, because they appear to serve as a major intersection site for various post-endocytic (transcytotic and recycling) receptors in epithelia. On the basis of these and other observations it has been postulated that in epithelial cells, apical
and basolateral sorting as well as recycling of endocytic receptors is
mediated by CE and/or in ARE (21, 24).
A significant level of pIgR transcytosis can also occur when the pIgR
is not bound to its ligand. This process has been referred to as
constitutive transcytosis (25), whose presumptive trigger is
phosphorylation of a cytoplasmic Ser, Ser664 (26, 27). Like
ligand-bound pIgR, the unoccupied pIgR also reaches ARE en
route to the apical cell surface (19). Thus, both dIgA-mediated
and constitutive transcytosis of the pIgR are potentially regulated by
multiple endosomal compartments.
The 17-residue cytoplasmic segment positioned immediately after the
transmembrane domain of the pIgR has been characterized as an
autonomous and dominant basolateral sorting signal (28). An in-frame
deletion of fourteen amino acids in this segment results in sorting of
the pIgR from the TGN to the apical plasma membrane. Individual
residues (His656, Arg657, and
Val660) within the 17-residue peptide have been shown to be
important for mediating basolateral transport of the receptor (29, 30). However, a fraction of the pIgR mutants impaired in basolateral sorting
can reach the basolateral surface, internalize into basolateral endosomes, and transcytose to the apical surface. Intriguingly, the
mutant receptors displayed enhanced transcytotic activity (8). Hence,
mutations that inhibited basolateral targeting and increased apical
targeting from the TGN similarly decreased basolateral recycling from
endosomes and increased basolateral-to-apical transcytosis. Mutations
in the basolateral targeting signal of the low density lipoprotein
receptor that affect the polarity of biosynthetic sorting show similar
enhancement of its apical sorting after endocytosis (31). One
interpretation of these observations is that the targeting signal acts
in two different locations: in the TGN during exocytosis and in
endosomes after endocytosis. This suggests that a similar type of
sorting machinery operates in both the TGN and endosomes. An
alternative model argues that the constitutive exocytic and
post-endocytic pathways of the pIgR converge into a common sorting
compartment. This compartment is unlikely to be the TGN (8, 31) but may
be endosomes. The latter model would predict the direct targeting of
newly synthesized pIgR from the TGN to endosomes.
A casein kinase II (CKII) phosphorylation motif in the cytoplasmic tail
of the mannose-6-phosphate receptor (MPR) is thought to mediate MPR
interaction with the Golgi-associated clathrin AP-1 adaptor (32, 33). A
number of observations argue that AP-1 binding may be coupled to the
subsequent process of MPR budding into clathrin-coated vesicles and
transport from the TGN to endosomes (34-36). We have recently
suggested that phosphorylated Ser726 residing in a putative
CKII/protein kinase A phosphorylation site in the cytoplasmic tail of
pIgR mediates the association of AP-1 adaptor with the pIgR (37). These
observations further suggested to us that similar to MPR, AP-1
association with phosphorylated Ser in the cytoplasmic tail of pIgR
elicits the sequestration of pIgR into clathrin-AP-1-coated domains in
the TGN, from where it could be transported to endosomes, prior to its
appearance on the basolateral cell surface. In the present study we
took morphological and biochemical approaches to assess this hypothesis and found that the polarized exocytic pathway taken by the pIgR may
involve various endosomal compartments.
Proteins and Antibodies
Lectin-triticum vulgaris (wheat germ agglutinin (WGA)) coupled
to horseradish peroxidase (HRP) and free HRP Type II were from Sigma.
Human apo-transferrin (hTfn; Biological Industries Co, Beit hemeek,
Israel) was loaded with iron as described (38). Rabbit antiserum
directed against a peptide in the cytoplasmic domain of the rat TGN
resident protein, TGN-38 (39), was kindly provided by Dr. G. Banting
(University of Bristol, Bristol, UK) and was used at 1:500 and 1:100
dilution for Western blotting and immunofluorescence, respectively. Fab
fragments derived from affinity purified guinea pig anti-rabbit SC
polyclonal antibodies were a kind gift from Prof. K. Mostov (University
of California, San Francisco) and were prepared as described previously
(40). Radioiodination of Fabs was performed according to the iodine monochloride protocol (40).
Cell Lines
MDCK cell lines that exogenously co-express the wild-type pIgR
and hTfnR was established as described (14). We have used MDCK cells
expressing various pIgR constructs. Some of their basic characteristics
are summarized in Fig. 1. To obtain cells
with higher expression levels, the pIgR-
INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES
EXPERIMENTAL PROCEDURES
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES
R655-Y668 mutant was
re-expressed in MDCK cells using the pCB6 vector (41), as described
(29). The pIgR mutant, pIgR-S726-t, whose 30 C-terminal amino acids have been deleted was described and characterized in an earlier work
(designated 725t in Ref. 42; see also Fig. 1). This pIgR mutant also
exhibited low expression levels and was therefore re-expressed in MDCK
cells. In both cases, cell lines exhibiting pIgR expression levels
comparable with those of other pIgR-WT expressing cell lines were
isolated. Exocytic and endocytic trafficking characteristics of the
re-expressed mutant receptors were identical to those described for the
original expressors. All MDCK cells were routinely cultured on 10-cm
dishes, in minimal essential medium (MEM; Sigma) containing 5% fetal
bovine serum and 1% antibiotics (Biological industries, Beit Haemek,
Israel). Cells were maintained in culture up to 10 passages. To induce
polarity, cells were cultured on 12-mm Transwell filters (Costar) for
at least 4 days prior to each experiment. Cell polarity was estimated
by measuring gp80 secretion as described (29).
View larger version (23K):
[in a new window]
Fig. 1.
Schematic presentation of cytoplasmic sorting
signals in the pIgR and cation-dependent MPR and mutations
constructed by site-directed mutagenesis. PIgR-WT, the
autonomous 17-residue basolateral sorting signal proximal to the
transmembrane domain (TM) is indicated with
underlining. pIgR whose Arg657 and
Val660 have been doubly mutated to Ala (pIgR-R657,V660-AA),
and pIgR whose Arg655-Tyr668 segment has been
deleted in-frame (pIgR- R655-Y668) are sorted to the apical cell
surface of MDCK cells. Individual amino acids (His656,
Arg657, and Val660) that play a role in
basolateral sorting of the pIgR are marked with bold
letters. The putative RXXV basolateral sorting motif is
indicated in parentheses. The two main phosphorylation sites
in the cytoplasmic tail (marked with asterisks, postulated
to be involved in constitutive transcytosis and endocytosis, are
Ser664 and Ser726, respectively.
Ser726 resides within a putative CKII/protein kinase A
phosphorylation site located upstream to a dileucine motif with yet
undefined function (underlined and bold). In
pIgR-S726-t, the codon encoding Ser726 has been mutated to
a stop codon, resulting in the deletion of the C-terminal 30 amino
acids in the cytoplasmic tail. CD-MPR, the basolateral
sorting signal is close to the transmembrane domain and is similar to
that of the pIgR. It contains a putative RXXV sequence (74).
Phosphorylated Ser in a CKII phosphorylation site (bold and
underlined) is proposed to mediate AP-1 binding (32). The
downstream dileucine-based motif (underlined) may play a
role in sorting from the TGN to endosomes or from endosomes to the cell
surface. Amino acids are indicated with the single-letter codes.
The Diaminobenzidine (DAB) Density Shift Meeting Assay
This assay is designed to measure the meeting of newly synthesized receptors in the exocytic pathway or receptors traveling in the post-endocytic pathway with a selective population of apical or basolateral endosomes loaded with horseradish peroxidase (HRP).
Targeting of Newly Synthesized Receptors to Apical or Basolateral Endosomes-- A modified version of the original DAB density shift assay (43) used by Apodaca et al. (19) has been applied here to assay the targeting of newly synthesized pIgR to apical or basolateral endosomes. The various steps of the assay are schematically depicted in Fig. 3A. MDCK cells grown on 12-mm Transwell filters were pulse-labeled for 10 min at 37 °C with [35S]Cys,Met (0.07 mCi of Promix; Amersham Pharmacia Biotech) and subsequently chased for 15 min at 37 °C in minimal essential medium containing bovine serum albumin (MEM-BSA) (MEM containing Hanks' balanced salt solution (Sigma), 20 mM Hepes, pH 7.4, and 0.6% BSA) lacking radioactive amino acids. To label apical endosomes, WGA-HRP (12 µg/ml) or free HRP (10 mg/ml) was introduced into the medium bathing the apical surface during the chase at 37 °C. To label basolateral early endosomes, free HRP (10 mg/ml) was added to the medium in the basolateral chamber during the 15-min chase. In all cases, the medium in contact with the basolateral surface was supplemented with V8 proteinase (Roche Molecular Biochemicals) throughout the entire chase period. The proteinase does not digest HRP, but it immediately digests the lumenal domain (i.e. the SC) of radiolabeled pIgR molecules that appear on the basolateral plasma membrane. V8-digested pIgR is no longer recognized by anti-SC antibodies and cannot be immunoprecipitated from the cell (our proteinase-based delivery assay is based upon this principle (29)). The application of V8 proteinase was important for preventing the internalization of intact molecules into the cell, thereby eliminating the ability of exocytic receptors to reach endosomes via the indirect post-endocytic route. In cases where a fraction of metabolically labeled pIgR reaches the apical surface, apical endocytosis of surface arriving receptor molecules is largely diminished because of efficient cleavage of pIgR to SC by the endogenous protease. Cells were then washed three times quickly and once for 5 min with ice-cold MEM-BSA containing 20% calf serum to quench the V8 proteolytic activity. Noninternalized HRP was stripped from the cell surface as follows: WGA-HRP was removed by incubating the cells with 100 mM N-acetyl-D-glucosoamine in ice-cold MEM-BSA for three times for 10 min each time. HRP was removed by washing the cells twice quickly and once for 60 min with ice-cold glycine buffer (150 mM NaCl, 2 mM CaCl2, 10 mM glycine, pH 9.0) as described (44). Cells were then washed twice with ice-cold MEM-BSA, medium was aspirated, and DAB reaction buffer (0.5 ml) was added to both the apical and basolateral compartments of the Transwell. The DAB reaction buffer was prepared by adding 1.65 ml of 3 mg/ml DAB to 10 ml HBSS+ (1.3 mM CaCl2, 0.4 mM MgSO4, 5 mM KCl, 5.6 mM D-glucose, 25 mM Hepes titered to pH 7.4 with NaOH). The pH of the solution was then carefully adjusted to 7.0 with NaOH, filtered through a 0.45-µm pore filter unit, and 12 µl of fresh 30% H2O2 was added to the DAB reaction buffer. In control experiments, H2O2 was omitted from the DAB reaction buffer. Upon incubation with monomeric DAB and peroxide, DAB is selectively polymerized by the peroxide. Consequently, peroxide-containing organelles become detergent-insoluble. Following 45 min of incubation at 4 °C, cells were washed twice with ice-cold HBSS+, buffer was aspirated, and filters were excised from their holders, boiled for exactly 90 s in 0.5 ml of SDS-lysis buffer (0.5% SDS, 100 mM NaCl, 50 mM triethanolamine chloride, pH 8.6, 5 mM EDTA. 0.02% (w/v) sodium azide) containing 1 mM phenylmethylsulfonyl fluoride. Filters in lysis buffer were then vortexed for 15 min at 22 °C, and SDS-insoluble DAB polymers were pelleted by centrifugation (100,000 × g, 4 °C, 25 min, TL-100 Beckman ultracentrifuge). Supernatants were added to 0.5 ml of 2.5% Triton dilution buffer (TDB) (100 mM NaCl, 5 mM EDTA, 100 mM triethanolamine chloride, pH 8.6, 2.5% Triton X-100, 0.02% sodium azide), and the pIgR was immunoprecipitated with sheep anti-rabbit SC polyclonal antibodies covalently coupled to protein A-Sepharose (50 µl of 15% slurry) as described (37). The immunoprecipitated protein was analyzed by SDS-PAGE and autoradiography. The pIgR was immunoprecipitated, and the degree of pIgR meeting with endosomes is calculated from the relative loss in radiolabeled receptor signal (see Fig. 3A).
Determination of Cell-associated HRP Activity-- Internalization of large HRP concentrations may result in nonspecific precipitation of metabolically labeled pIgR. We hence carefully calibrated the concentration of HRP ligands applied to the cells in each experiments, and the concentration chosen (mentioned above) resulted in the consistent internalization of HRP activity that did not cause the pelleting of pulse-labeled receptors (namely, receptors located in the endoplasmic reticulum).
Cell-associated HRP activity in each experiment was determined using the following protocol. HRP-coupled ligands or free HRP was internalized from the apical surface of MDCK cells, and surface-bound ligand was removed. Cells were washed twice quickly with ice-cold phosphate-buffered saline, the filter was excised from its holder, and cells were scraped into 0.1 ml of ice-cold TDB containing 2.5% Triton X-100 lacking sodium azide. Cell lysates were vortexed at 4 °C for 20 min, and 1 ml of phosphate-buffered saline containing 0.75 mg/ml o-phenylenediamine dihydrochloride (Sigma) and 0.023% H2O2 was added to the lysate. The mixture was incubated further for 20 min on ice, and the reaction was stopped with 0.3 ml of 1 N HCl. Absorbance was determined at 492 nm, and the amount of cell-associated HRP was derived from a standard calibration curve using 1-50 ng of free HRP. Uptake of WGA-HRP at 37 °C or at 17 °C from the apical surface of filter-grown MDCK cells resulted in the internalization of ~30 ng of HRP. Fluid phase uptake of free HRP at 37 °C caused reproducible internalization of 30-50 ng of HRP activity. Only about 1% of pulse-labeled receptors was pelleted after ligand binding to the apical surface at 4 °C (resulting in >50 ng of cell-associated HRP activity), confirming that radiolabeled receptors are not shifted by nonspecific association with DAB polymers after lysis of cells in SDS.
Meeting of Post-endocytic Receptors with Apically Endocytosed
WGA-HRP or Free HRP--
Cells expressing the pIgR cultured on 12-mm
Transwells were placed on 30-µl drops of MEM-BSA containing 10 µg/ml 125I-labeled Fab fragments (Fabs produced from
guinea pig anti-SC antibodies were iodinated by the iodine monochloride
method (8)) yielding specific activity of 2 × 107
cpm/µg protein), and ligand was allowed to internalize from the basolateral surface for 30 min at 37 °C. During the last 15 min of
Fab internalization, WGA-HRP or fluid phase HRP were internalized from
the apical surface. Surface-bound ligand was removed, and DAB reaction
was initiated with H2O2 in one set of filters,
whereas H2O2 was omitted in control filters.
Cells were solubilized in 0.5 ml of SDS-lysis buffer, lysates were
boiled for 90 s, and SDS-insoluble DAB polymers were pelleted by
centrifugation (100,000 × g, 25 min, 4 °C, TL-100
Beckman ultracentrifuge). Radioactivity level associated with
SDS-soluble material in the supernatant (X) was determined by a
-counter. MDCK cells that do not express the pIgR were analyzed in
parallel to control for nonspecific uptake of ligand, and the values
obtained were subtracted as background (never more than 7% of the
signal from pIgR-expressing cells). Each experiment was performed in
triplicate, and the level of meeting was deduced from the percent of
reduced radioactivity in treated (+H2O2)
compared with nontreated cells (
H2O2), using the equation presented in the legend to Fig. 3A (see also
Fig. 8A). In cases where meeting with hTfn was examined, PTR
cells were first depleted of endogenous Tfn by incubating the cells for
4 h at 37 °C in MEM-BSA. Human apo-transferrin (Biological Industries, Beit Haemek, Israel) was loaded with iron (38) and iodinated by the iodine monochloride technique to yield approximately 10 µg/ml 125I-labeled hTfn at 7 × 106
cpm/µg specific activity. Ligand was continuously internalized from
the basolateral surface of PTR cells for 30 min at 37 °C to allow
the labeling of ARE or CE (19), and its association with apically
endocytosed HRP was determined as described above for pIgR labeled with
125I-Fabs.
Trafficking Assays
Basolateral Delivery--
The rate of newly synthesized pIgR
delivery to the apical or basolateral cell surface of MDCK cells, in
the absence or presence of BFA, was determined quantitatively by the
protease-based delivery assay (29). A biotinylation-based assay for
cell surface delivery was used to assess the targeting of newly
synthesized hTfnR to the basolateral and apical cell surface. PTR cells
cultured on 12-mm filter supports were pulse-labeled for 15 min at
37 °C, washed, and then chased for 7, 15, 30, or 60 min at 37 °C
in MEM-BSA. At the end of each chase time, cells were washed three
times with ice-cold HBSS+ adjusted to pH 7.4. Proteins at
the apical and basolateral cell surface were biotinylated with
NHS-LC-biotin (500 µg/ml; Pierce) on ice for 15 min. Biotinylation
was repeated, and cells were subsequently washed three times for 5 min
each time with MEM-BSA supplemented with 50 mM glycine.
Cells were then lysed in 2.5% TDB containing a mixture of protease
inhibitors, and the hTfnR was immunoprecipitated using H68.4 monoclonal
antibodies (45) coupled to protein A-Sepharose. After three washes with
mixed micelle buffer (20 mM triethanolamine chloride, pH
8.6, 150 mM NaCl, 5 mM EDTA, pH 8.0, 8% w/v
sucrose, 0.1% NaN3, 1% w/v Triton X-100, 0.2% SDS) and
one wash with final wash buffer (mixed micelle buffer lacking Triton
X-100 and SDS), the immunobead suspension in final wash buffer was
divided into two fractions, each of which was centrifuged. The bead
pellet was processed as follows. Beads in one-fifth of the original
volume were dissolved in SDS sample buffer; this fraction was used to
determine the total amount of immunoprecipitated radiolabeled
receptors. The remaining bead suspension was pelleted, boiled for 5 min
in 20 µl of 10% SDS, and biotinylated pIgR was precipitated by
streptavidin agarose (50% slurry in 2.5% TDB; Sigma) as described
(8). HTfnR precipitated from both fractions was analyzed by 10%
SDS-PAGE, and radioactivity was quantified by the FUJIX BAS 1000 bioimaging system as described. In some experiments 10 µg/ml BFA
(EpiCenter; diluted from a 10 mg/ml stock solution in Me2SO
stored at 20 °C) was included into the apical and basolateral
medium throughout the chase.
Steady-state Distribution-- To evaluate the steady-state distribution of the hTfnR on the apical and basolateral plasma membranes of PTR cells, each cell surface was selectively biotinylated in the cold. After three washes with MEM-BSA-glycine buffer, cells were lysed in TDB, and the hTfnR was immunoprecipitated. Immunoprecipitates were analyzed by SDS-PAGE electrotransferred to a nitrocellulose membrane (37), and the nitrocellulose membrane was incubated with a blocking buffer (37) lacking dry milk. Biotinylated receptors were detected with HRP-labeled streptavidin (Pierce, 1:1000 in TBST (10 mM Tris, pH 8.0, 180 mM NaCl, 0.05% Tween 20), 30 min at 22 °C) using a Luminol Reagent (Santa Cruz Biotechnology). The 90-kDa hTfnR band intensity was quantified (37).
Basolateral Recycling and Transcytosis--
Basolateral
recycling of hTfnR in PTR cells was monitored basically as described
for the recycling of the endogenous dog transferrin (19). PTR cells
were first depleted of endogenous Tfn by incubating the cells for
4 h at 37 °C in MEM-BSA. Cells were incubated from the
basolateral side with 125I-hTfn (10 µg/ml at 7 × 106 cpm/µg specific activity) for 120 min at 18 °C.
Cells were washed and incubated at 37 °C for 3, 7, 15, 30, 60, and
120 min in the absence of the radiolabeled ligand. At the various time
intervals, cells were chilled, and the apical and basolateral media
were harvested. Surface-associated hTfn was stripped by incubating the
cells for 60 min at 4 °C in acidic glycine buffer (750 mM glycine, pH 2.5, diluted 1: 5 in phosphate-buffered
saline containing 0.6% BSA), and filters were then cut out of their
holders. Radioactivity was determined for filters, the stripped
fraction, and the basolateral and apical media. Constitutive
transcytosis of empty pIgR was assessed using guinea pig anti-SC
125I-Fab fragments as described (8), except that V8
proteinase (15 µg/ml; Roche Molecular Biochemicals) was included into
the basolateral medium throughout the various incubation times. It should be noted that although most of the nascent pIgR-R655-Y668 or
pIgR-R657,V660-AA are targeted from the TGN to the apical surface, a
small fraction of these molecules is still directed to the basolateral surface, and the transcytotic activity of this fraction was monitored.
Immunofluorescent Labeling and Scanning Laser Confocal Microscopy
HRP coupled to fluorescein isothiocyanate (FITC) (5/mg/ml) or
WGA-HRP (12 µg/ml) was internalized from the apical surface for 15 min at 37 °C. Surface-bound ligand was stripped, and cells were
fixed for 10 min in methanol at 20 °C (37), blocked and immunostained as described (19, 37). Internalized WGA-HRP was labeled
with goat anti-HRP antibodies coupled to FITC (Jackson Immunochemicals;
used at 1:250 dilution). The TGN was labeled with rabbit anti-TGN38
antibodies (at 1:250 dilution; kindly provided by G. Banting,
University of Bristol) and donkey anti-rabbit secondary antibodies
conjugated to Texas Red (Jackson Immunochemicals). The R40.76 anti-ZO-1
rat monoclonal antibody (46) (used at 1:250 dilution) with
anti-rat-conjugated tetramethylrhodamine isothiocyanate (Jackson
Immunochemicals) secondary antibody was used to label the tight
junction-associated protein ZO-1. Secondary antibodies were used at the
recommended dilutions.
A Bio-Rad MRC-1024 confocal scanhead coupled to a Zeiss Axiovert 135M inverted microscope was used to acquire images of the stained cells, with a 63× oil objective (numerical aperture 1.4). The instrument was configured for detection of three labels (FITC, tetramethylrhodamine isothiocyanate, and Texas Red) as described in Ref. 37. Data presented in this paper did not demonstrate colocalization of FITC and Texas Red, when analyzed as described in Ref. 37. However, that analysis was designed to provide a strict test for colocalization. Once colocalization was not observed, the analysis was modified to provide a strict test for lack of colocalization, as described below. All images were filtered with a 3 × 3 median filter to remove point noise. Then the uniform background (estimated from areas of the field where there were no cells) was subtracted from each image. The FITC staining covered a wide dynamic range, so that some of the spots were saturated. Saturated areas of the images were discarded, because such signals are necessarily outside of the linear range of the detection system. Cross-talk between the FITC and Texas Red channels was negligible, as shown with control experiments. When testing for colocalization, an upper limit on cross-talk was used. However, to test for lack of colocalization, a lower limit on cross-talk (that is, zero) was used. Finally, the contrast was stretched to allow the stained areas to be easily seen. Areas of overlap of Texas Red and FITC appear yellow. There was no attempt to quantitate the relative strengths of the FITC and Texas Red emission. Images were saved in a tag information file format, and the contrast levels of the images were adjusted in the Photoshop 5.0 program (Adobe Co., Mountain View, CA) on a Power PC G3 Macintosh (Apple). The contrast corrected images were imported into Freehand (Macromedia, San Francisco, CA) and printed from a high quality color printer (Tektronix Phaser 440).
Electron Microscopic Analysis of Apically Endocytosed HRP Ligands
Following internalization of WGA-HRP or fluid phase HRP from the
apical surface, cells were fixed for 30 min at 22 °C with a mixture
of 2% paraformaldehyde 1% glutaraldehyde dissolved 0.1 M
phosphate buffer (0.1 M NaH2PO4 and
0.1 M Na2HPO4, pH 7.4). Cells were
rinsed three times in phosphate-buffered saline and then treated for 10 min at 22 °C with freshly prepared DAB solution (0.025% DAB
(Sigma), 0.006% H2O2 in TBS (0.05 M Tris HCl, pH 7.4, and 0.9% (w/v) NaCl). Post-fixation
was performed in a mixture of osmic acid tetroxide (1% v/v) and
potassium ferricyanide (1.5% w/v) in cacodylate buffer (0.05 M sodium cacodylate pH 7.0) for 1 h at 22 °C. After
brief washes in cacodylate, sections were dehydrated in ascending
concentrations of ethanol and infiltrated with epoxy resin (EM-BED-812,
Electron Microscopy Sciences). Flat embedding and ultrathin sectioning
was performed as described elsewhere (47). Sections were viewed in a
Philips 300 electron microscope.
| |
RESULTS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES
|
|---|
Membrane-bound WGA-HRP and Fluid Phase HRP Internalized at 37 °C or at 17 °C from the Apical Membrane Do Not Co-localize with the Golgi System-- We assayed the extent to which newly synthesized proteins meet with free HRP or HRP-coupled ligands internalized into endosomes. A prerequisite for this assay is to exclude the possibility that apically internalized ligands access the Golgi complex or the TGN compartment. It is well established that after short periods of internalization, free HRP taken up by fluid phase endocytosis enters into the vesicular portion of early sorting endosomes, and in polarized MDCK cells this marker can be selectively internalized into distinct apical and basolateral early endosomes (48, 49). On the other hand, general membrane-bound tracers like ricin and WGA have been used to label tubulovesicular recycling endosomes of nonpolarized cells and ARE of epithelial cells (19, 20, 50-52). We used morphological and biochemical approaches to confirm that in our experimental system, apically internalized WGA-HRP or fluid phase HRP indeed reach endosomes but not the Golgi/TGN compartments.
WGA-HRP or free HRP were internalized from the apical surface for 15 min at 37 °C, and surface-bound ligand was removed. Cells were fixed
and co-stained with specific anti-HRP and anti-TGN38 antibodies
followed by FITC or Texas Red-coupled secondary antibodies, respectively. The localization of HRP labeling with respect to the TGN
marker, TGN38, in the apical-basal axis of the cell was recorded by
scanning laser confocal microscopy. Confocal images clearly show that
HRP-FITC produced very intense punctate staining, most likely because
of ligand entering into the vesicular portion of early apical
endosomes. WGA-HRP-FITC, in contrast, produced more extended staining
patterns, probably because it is concentrated into tubular extensions
of sorting endosomes and/or because of its entering the tubular portion
of ARE or CE residing underneath the apical plasma membrane (Fig.
2A). At the same time, the
anti-TGN38 produced a more widely distributed reticular staining. In
all cases, HRP-FITC labeling appeared in focal planes above the tight junctions, whereas the TGN38-TR was seen in the plane of the tight junction and below (basolateral). Furthermore, even near the tight junctions, where very low HRP-FITC labeling and intense TGN38-TR labeling appeared in the same focal plane, there was practically no
overlap between them. These data suggest that internalized ligands
label morphologically distinct endosomal compartments at the apical
pole of the cells and that they do not reach the TGN. Similar results
were obtained after apical internalization of HRP ligands for 90 min at
17 °C (not shown). Here it is important to note that internalized
WGA likely labels various apical endosomal compartments, including the
CE or the ARE, but fluid phase HRP labels primarily early
endosomes.
|
Next we analyzed the localization of internalized ligands with respect to the Golgi apparatus using high resolution electron microscopy. WGA-HRP or free HRP were internalized from the apical surface as before, and cells were processed for electron microscopy (Fig. 2B). Both ligands labeled organelles residing in the apical cytoplasm, some of which are in proximity to the Golgi apparatus. However, in more than 20 sections in each case, HRP reaction product was never found to label Golgi cisternae, suggesting that internalized ligands reside predominantly in the lumen of endosomes.
Finally, we used the density shift meeting assay to assess the ability
of internalized tracers to access the TGN. WGA-HRP was internalized
from the apical surface for 15 min at 37 °C, surface-bound lectin
was removed, and HRP-mediated DAB polymerization was either induced (+)
or not () with H2O2. Cells were solubilized in SDS, insoluble-shifted material was removed by centrifugation, and
TGN38 was immunoprecipitated from the soluble fraction. We found that
comparable amounts of TGN38 were immunoprecipitated from the induced
and noninduced cells (Fig. 2C), suggesting that internalized
ligand does not reach significant portions of TGN38-containing organelles in steady state. Similar data were obtained with free HRP
(not shown). From the biochemical and morphological observations we
conclude that apically internalized WGA-HRP or free HRP reach apical
endosomes but not the Golgi apparatus or the TGN.
The Biosynthetic Pathway to the Basolateral Surface Passes through
Apical and Basolateral Endosomes--
Our first aim was to analyze the
direct delivery of newly synthesized pIgR-WT to endosomes. The exocytic
transport of the pIgR from the TGN to the basolateral cell surface was
followed by pulse labeling of pIgR-expressing MDCK cells with
35S-labeled methionine and cysteine for 10 min and chased
for 15 min in medium that lacked the radiolabeled amino acids. During the chase, V8 proteinase was added to the basolateral medium. The
proteinase digested surface arriving pIgR molecules, thereby eliminating the ability of exocytosing molecules to reach endosomes via
the post-endocytic pathway (see "Experimental Procedures"). We
found that after pulse labeling and 15 min chase, only about 20% of
radiolabeled pIgR-WT accessed the basolateral cell surface, and none
could be detected in the apical plasma membrane (37). Thus, we thought
that this chase period was appropriate for monitoring the fraction of
newly made pIgR that is targeted directly to endosomes. Concomitant
with the chase, WGA-HRP or free HRP was internalized from the apical
surface, and the extent to which nascent pIgR meets with apically
internalized ligand was assayed by the
DAB-H2O2-induced density shift of organelles
containing HRP as described under "Experimental Procedures" and in
Fig. 3A. The average of
several such experiments is shown in Fig. 3B.
|
After 10 min of pulse and 15 min of chase, ~30% of pulse-labeled pIgR reaches apical endosomes containing WGA-HRP, and less than 10% accesses free HRP taken up by fluid phase endocytosis (Fig. 3B, Pulse+Chase). Because ARE and CE can be accessed by membrane-bound endocytic ligands but are largely devoid of fluid phase endocytic tracers, these results suggest that newly synthesized pIgR is targeted from the TGN to a population of apical endosomes that probably correspond to ARE and CE, prior to reaching the basolateral plasma membrane. Interestingly, however, ~30% of newly synthesized pIgR also seems to meet with fluid phase HRP internalized from the basolateral surface. Fluid phase markers are thought to label primarily early sorting endosomes (e.g. the early endosomal rab5a has been co-localized with internal fluid phase markers (53)); thus, the exocytic pathway taken by pIgR-WT inefficiently traverses AEE but may include BEE.
Pulse-labeled pIgR is located in the endoplasmic reticulum and is sensitive to digestion by endoglycosydase H (23). In contrast, after pulse and 90 min of chase at 17 °C, the pIgR is blocked from exiting the TGN, and it acquires resistance to digestion by the enzyme (23). In both cases, the pIgR should not meet ligands in endosomes. Cells were pulse-labeled for 10 min or pulse-labeled and then chased for 90 min at 17 °C. WGA-HRP was internalized during pulse-labeling or during the 90-min chase, and the fraction of metabolically labeled pIgR pelleted with HRP-containing apical endosomes was measured. Results presented in Fig. 3B reveal that less than 5% of pulse-labeled pIgR meets with HRP tracers internalized during pulse labeling. Approximately 10% of the pulse-labeled pIgR appeared to associate with HRP-containing endosomes after 90 min chase at 17 °C. Collectively, these results are consistent with the expectation that newly synthesized pIgR in the endoplasmic reticulum or Golgi does not meet HRP-coupled ligands in endosomes and further confirm the reliability of the assay in detecting receptor meeting with endosomes. However, ~30% meeting was obtained if pulse-labeled pIgR was incubated for 90 min at 17 °C and subsequently chased for 15 min at 37 °C with WGA-HRP internalized from the apical surface during the 15 min chase (not shown). In this case, meeting is likely contributed by direct delivery of the pIgR from the TGN to apical endosomes. In summary, the results so far indicate that in the course of exocytosis, pIgR-WT is targeted from the TGN to at least two types of endosomal populations: apical endosomes (presumably ARE and/or CE) and BEE.
Kinetics of pIgR Delivery to Endosomes-- The degree of meeting between metabolically labeled pIgR and endosomes should vary as a function of chase time. Short periods of chase should result in low meeting levels because the cohort of metabolically labeled receptors has not yet entered endosomes. Meeting levels should reach maximal levels and then decrease at longer chase periods as a result of [35S]pIgR entering and subsequently leaving endosomes en route to the cell surface. These predictions have been tested by measuring the meeting levels between pulse-labeled pIgR and apically endocytosed WGA-HRP after 5, 15, and 25 min of chase in the presence of basolateral V8 proteinase. In each chase period, WGA-HRP was consistently internalized from the apical surface for 15 min. Data in Fig. 3C agree with the above expectation; after 5 min chase only 3% of [35S]pIgR was associated with endosomes. The meeting levels peaked to approximately 35% after 15 min chase and subsequently decreased to ~25% following 25 min of chase.
Previously, it was shown that after a 25-min chase, approximately 40% of pulse-labeled receptors reaches the basolateral surface (37). In the absence of basolateral V8, a fraction of these receptors would internalize and reach apical endosomes via the indirect transcytotic route. In this case, meeting levels are expected to increase because apical endosomes can now be reached by [35S]pIgR traveling in the direct (biosynthetic) and indirect (transcytotic) routes. Data in Fig. 3C indeed show that in the absence of V8, meeting levels are doubled (from 25 to 50%) after 25 min of chase, suggesting that apical endosomes can also be accessed by receptors traveling in the post-endocytic, e.g. transcytotic, route. This possibility is further investigated below.
The Biosynthetic Route of pIgR-S726-A Circumvents Endosomes,
Whereas Exocytosing pIgR-S726-t Is More Efficiently Targeted to Apical
Endosomes--
It has been previously suggested that pIgR-S726-A and
pIgR-WT utilize BFA-insensitive and BFA-sensitive routes from the TGN to the basolateral surface, respectively (37). We examined the possibility that these mechanistically distinct exocytic routes may
also be distinguished by their ability to access endosomes. In Fig.
4 we show that after pulse and 15 min of
chase, ~1% of pIgR-S726-A meets with apically internalized WGA-HRP,
and about 5% of radiolabeled receptors are associated with apically or
basolaterally internalized fluid phase HRP. Thus, after the indicated
chase period, insignificant levels of newly synthesized pIgR-S726-A appeared to reach endosomes. However, it should be noted that after 15 min of chase the pIgR-S726-A mutant barely reaches the basolateral
surface (37). Hence, it is plausible that longer chase times are
required to transport [35S]pIgR-S726-A to endosomes.
However, this does not seem to be the case because no meeting with
either tracer could be observed even after 25 min of chase (not shown),
when ~30% of the pulse-labeled pIgR-S726-A appeared to reach the
basolateral cell surface (37). These results strongly imply that
pIgR-S726-A utilizes an alternative basolateral exocytic route that
does not involve endosomes.
|
The data presented so far suggest that membrane proteins that take a slower and BFA-insensitive route bypass endosomes, whereas proteins taking a faster and BFA-sensitive pathway to the cell surface utilize endosomes. If this correlation were true, we expect that a pIgR mutant whose targeting to the basolateral cell surface occurs with faster kinetics will be more sensitive to BFA and meet endosomes more efficiently than pIgR-WT. An earlier work has shown that basolateral exocytosis of the truncated pIgR mutant, pIgR-S726-t (see also Fig. 1), occurs at significantly higher rate than pIgR-WT (42). For example, after 15 min of chase, 20% of [35S]pIgR-WT was detected on the basolateral plasma membrane compared with ~50% of [35S]pIgR-S726-t. Interestingly we find that basolateral delivery of pIgR-S726-t was also more potently inhibited by BFA; after 60 min chase ~40% of [35S]pIgR-WT reached the basolateral plasma membrane of BFA-treated MDCK cells (30), whereas 20% delivery of pIgR-S726-t is observed after the same chase time. Consistent with our expectation, results in Fig. 4 show that after 15 min of chase at 37 °C, ~60% of pulse-labeled pIgR-S726-t is targeted to WGA-HRP-loaded apical endosomes. These meeting levels are almost two times greater than those measured for pIgR-WT (Fig. 3B). The pIgR-S726-t mutant did not meet with apically internalized HRP. Together, these results suggest a correlation between the efficiency of exocytosis and meeting with apical endosomes.
The Biosynthetic Route of hTfnR Traverses Apical Endosomes--
We
next investigated whether the convergence of the biosynthetic and
endocytic pathways represents a general phenomenon and encompasses
other basolateral receptors, such as the hTfnR. To investigate this
possibility, we isolated an MDCK cell line in which we co-expressed the
pIgR and the hTfnR (called PTR). In this cell line, approximately 80%
of surface-associated hTfnR is localized on the basolateral plasma
membrane and 20% in the apical domain (Fig.
5A), and the hTfnR is
predominantly delivered from the TGN to the basolateral surface (Fig.
5B). In addition, basolateral transport of newly synthesized
hTfnR is significantly inhibited by BFA (Fig. 5B).
|
PTR cells were pulse-labeled, and the meeting of radiolabeled pIgR and hTfnR with WGA-HRP or fluid phase HRP internalized from the apical plasma membrane was examined after 15 min of chase at 37 °C, with V8 protease included into the basolateral medium during the chase. Under these conditions, 40% of the radiolabeled pIgR and hTfnR were found to associate with apical endosomes bearing WGA-HRP, and a much smaller fraction (5-10%) accessed apical endosomes loaded with fluid phase HRP (Fig. 5C). These results suggest that the biosynthetic pathways of both the pIgR and hTfnR converge in apical endosomes and are consistent with previous observations showing that the biosynthetic pathway of hTfnR operates via endosomes in nonpolarized cells (10).
Apical Exocytosis of pIgR Mutants Occurs via Apical
Endosomes--
We next analyzed the possibility that apically sorted
newly made pIgR mutants pass through endosomes. Two pIgR mutants,
pIgR-R655-Y668 and pIgR-R657,V660-AA (28, 30), were chosen for this
purpose. Examination of the kinetics of their surface delivery
revealed that after pulse and 15 min of chase, pIgR-R657,V660-AA does
not appear on the apical cell surface, and only ~10%
pIgR-R655-Y668 accessed that surface (not shown). Neither receptor
mutant could be detected on the basolateral plasma membrane after the
indicated chase time.
PIgR-R655-Y668 or pIgR-R657,V660-AA were pulse-labeled and chased
for 15 min at 37 °C. WGA-HRP or fluid phase HRP was internalized from the apical surface during the chase, and the meeting of
biosynthetically labeled receptors with ligand-containing endosomes was
assessed by the DAB density shift assay. Results show that about 60%
of pulse-labeled pIgR-R655-Y668 or pIgR-R657,V660-AA meets with apical endosomes loaded with WGA-HRP (Fig.
6). The extent of meeting is almost
2-fold greater than that measured for pIgR-WT or hTfnR (Figs.
3B and 5C).
|
WGA-HRP is likely internalized into various apical endosomal compartments, including AEE. Thus, we suggest that higher meeting efficiency could be attributed to targeting of newly made pIgR mutants to apical endosomal populations that can be accessed by membrane-bound ligands (e.g. ARE and CE) and by fluid phase ligands, namely, apical early endosomes. One can distinguish early endosomes from CE or ARE, because only early endosomes can be accessed efficiently by fluid phase markers. If so, the exocytic road of apically sorted pIgR mutants should encounter HRP selectively internalized into AEE. To test this hypothesis, the targeting of newly made pIgR mutants with AEE loaded for 15 min at 37 °C with fluid phase HRP was examined. Data summarized in Fig. 6 indeed show reduced meeting levels (30-40%) of biosynthetically labeled pIgR mutants with fluid phase HRP in AEE.
Fluid phase tracers can be chased from early endosomes into late endosomes by incubating the cells for another 15 min at 37 °C in the absence of tracers (48). To elucidate whether apically sorted pIgR encounters early versus late endosomal compartments, HRP was continuously internalized during the 15-min pulse-labeling period, and 15 min of chase at 37 °C was performed in the absence of HRP. Under these conditions, <5% of metabolically labeled receptors intersect with the fluid phase tracer (not shown), suggesting that exocytic pIgR mutants meet with HRP in early endosomes. Taken together, the results suggest that the biosynthetic route of apically sorted receptors pass through at least two apical endosomal populations: endosomes accessed by membrane-bound ligands and endosomes accessed by fluid phase tracers (i.e. AEE).
Transcytosing Receptors Interact More Efficiently with Fluid
Phase-containing Endosomes Than Do Basolateral Recycling
Receptors--
We measured recycling and trascytosis of the pIgR and
hTfnR using anti-SC 125I-Fab fragments and
125I-hTfn, respectively. The pIgR mutant pIgR-S664-A and
hTfnR display efficient basolateral recycling and minor transcytotic
activity (Fig. 7 and Ref. 27).
Conversely, internalized pIgR-R655-Y668 and pIgR-R657,V660-AA
receptors efficiently transcytose from the basolateral to the apical
surface and poorly recycle to the basolateral plasma membrane (Fig. 7).
Post-endocytic pIgR-WT exhibits an intermediate transcytotic/recycling
phenotype (Fig. 7). If pIgR trafficking in the exocytic and
post-endocytic pathways is regulated by a common mechanism (8), we
expect that like basolaterally routed exocytic receptors, basolateral
recycling receptors will encounter more efficiently apical endosomes
loaded with WGA-HRP but not with fluid phase HRP. In contrast,
analogous to apically targeted exocytic pIgR mutants, transcytosing
receptors may meet endosomes loaded with both types of endocytic
tracers. To test this hypothesis we measured the meeting of basolateral
recycling and transcytosing receptors with apically internalized
WGA-HRP or free HRP, using the DAB density shift protocol (see
"Experimental Procedures" and Fig.
8A).
|
|
Data presented in Fig. 8B clearly reveal that
125I-Fab continuously internalized from the basolateral
surface by pIgR-S664-A-expressing cells or 125I-hTfn
internalized by PTR cells meet with apically internalized WGA-HRP but
essentially not with fluid phase HRP. Conversely, 125I-Fab
internalized from the basolateral surface by cells expressing apically
targeted pIgR mutants meet more efficiently with fluid phase HRP. Low
meeting levels are also observed between transcytosing pIgR-WT and free
HRP, whereas almost 60% of 125I-Fab-labeled receptors meet
with apically internalized WGA-HRP (Fig. 8B). These results
suggest that the constitutive exocytic and transcytotic pathways taken
by apically sorted pIgR involve AEE, whereas the basolateral pathway
utilizes primarily ARE or CE but largely avoids AEE.
| |
DISCUSSION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES
|
|---|
Our findings essentially argue that in polarized MDCK cells:
(a) The BFA-sensitive exocytic road taken by basolaterally
sorted pIgR involves apical endosomes that are not early endosomes
(perhaps CE or ARE) and BEE. (b) Phosphorylated
Ser726 in the cytoplasmic domain of pIgR is essential for
receptor targeting from the TGN to endosomes; mutational inactivation
of this Ser confers pIgR incorporation into a BFA-insensitive
basolateral route that also bypasses endosomes. Thus, membrane proteins
can probably utilize mechanistically distinct but parallel exocytic routes and display variable ability in the extent to which they utilize
endosomes during exocytosis. (c) The biosynthetic pathway of
apically sorted receptors involves several apical endosomal populations, including AEE. (d) Transcytotic receptors
taking the constitutive post-endocytic route traverse AEE, whereas
basolateral recycling receptors interact with apical endosomes distinct
from AEE. Thus, the exocytic and post-endocytic pathways of membrane proteins display a parallel phenotype with respect to their ability to
interact with apical endosomes. The proposed pathways taken by the pIgR
are summarized in Fig. 9.
|
We have recently suggested that the Golgi AP-1 adaptor complex interacts with the cytoplasmic domain of the pIgR and that these interactions are abolished when phosphorylated serine 726 is mutated to alanine (37). Because BFA causes AP-1 dissociation from Golgi membranes in MDCK cells, it is reasonable to assume that pIgR interactions with AP-1 on the TGN contribute to the BFA-sensitive basolateral transport of the receptor (30, 37, 54). These interactions may also play a role in pIgR targeting from the TGN to endosomes (Fig. 9, step 1). In contrast, the pIgR-S726-A mutant does not require AP-1 adaptors for its budding from the TGN. This mutant receptor hence buds into vesicles with distinct properties in a process that is BFA-insensitive and does not involve endosomes as an intermediate compartment (Fig. 9, step 9). Thus, phosphorylation of Ser726 is proposed to function as an important physiological switch for interactions of biosynthetic pIgR with specialized AP-1-coated areas in the TGN. These AP-1 adaptors are not involved in polarized sorting of the pIgR and thereby contain the µ1a rather than the µ1b subunit recently suggested to comprise an epithelial-specific AP-1 adaptor complex involved in basolateral sorting (55).
The phenomena described herein may not be restricted to the exogenously expressed pIgR in MDCK cells. The pathways taken by the pIgR-S726-A may correspond to basolateral trafficking pathways utilized by endogenous proteins, such as the BFA-insensitive basolateral pathway of uvomorulin in MDCK cells (56). We do not know whether the rate and BFA insensitivity basolateral transport of these proteins is modulated by CKII phosphorylation motifs. In fact, nor do we know the extent to which Ser726 to Ala mutation really mimics the state of nonphosphorylated serine. It is entirely possible that the mutation artifactually uncovers a nonphysiological sorting mechanism favoring the BFA-insensitive pathway. Nevertheless, it is worth noting that CKII phosphorylation of serines has been implicated in modulating the traffic of several proteins (32, 57-59). Thus, it is quite plausible that these motifs may also play a role in basolateral trafficking of endogenous as well as exogenously expressed membrane proteins.
Our data are consistent with previous observations indicating the existence of multiple distinct pathways (mechanisms) from the TGN to the apical (60, 61) and basolateral (62) plasma membrane. Another interesting case of parallel pathways has been documented for the targeting of soluble hydrolases to the vacuole in yeast. One of these pathways is regulated by the recently characterized AP-3 adaptor coat (63). Multiple redundant routes may hence function in higher as well as lower eukaryotes. In both cases the physiological significance of multiple routes is not clear. Parallel pathways may permit flexible regulation with respect to the efficiency of cargo sorting and targeting from the TGN to the basolateral plasma membrane, allowing cells to adjust sorting and targeting in response to changing physiological and developmental requirements. In addition, redundant pathways may provide alternative modes of transportation in cases where one of the pathways is paralyzed.
Newly synthesized pIgR mutated in the basolateral sorting signal is routed to the apical surface via apical endosomes that include AEE prior to its appearance on the apical cell membrane (Figs. 6 and 9 steps 1 and 5). The Ser726 motif was kept intact in these mutants, and AP-1 adaptors may interact to some extent with this motif (37). The observation that apically and basolaterally sorted receptors meet with apically internalized membrane-bound ligand raises the intriguing possibility that the exocytic roads of these receptors merge into a common endosomal compartment, which could be the ARE or CE (Fig. 9, step 1). Consequently, certain steps in polarized sorting of these receptors may take place in endosomes. For instance, apical and basolateral proteins may be segregated from each other into discrete specialized populations of endosomes or alternatively into subdomains of individual endosomes specialized for basolateral or apical sorting (recycling). The existence of laterally segregated lipid microdomains in the membrane of tubulovesicular sorting endosomes (64) may contribute to membrane protein concentration into specialized distinct regions within this complex organelle. It is worth noting that although our data suggest that sorting in the biosynthetic pathway occurs in endosomes, we emphasize that they do not provide any direct evidence concerning the role of endosomes in polarized exocytosis. It seems equally likely, for instance, that exocytic wild-type receptors access endosomes and recycle back to the TGN, where the crucial sorting event may still take place. In this case, the observed propensity of different pIgR mutants to encounter the apical endosome could be explained by differences in the kinetics of recycling exhibited by the different mutants. Thus, additional experiments are needed to establish the precise intracellular site for polarized sorting.
Clathrin-coated buds containing -adaptin have previously been
localized mainly on the TGN (65) but have also been found to be
associated with endosomal membranes of nonpolarized cells (34). Is it
possible that like in the case of polarized exocytosis from the TGN
(30), -adaptins also control exocytosis from endosomes in polarized
cells? Ultrastructural observations revealed that in polarized MDCK
cells, dIgA and hTfn are associated with basolateral endosomes
) were internalized from the apical surface to label
apical endosomes or from the basolateral plasma membrane to label
basolateral endosomes (not shown). The density of endosomes was shifted
by HRP-catalyzed DAB polymerization (labeled with gray)
induced with peroxide (+H2O2). Consequently,
HRP-containing organelles become detergent-insoluble. The protein
content of density-shifted organelles, which in our experiment may
contain metabolically labeled receptors, cannot be extracted and simply
analyzed by immunoprecipitation. However, the fraction of receptor
reaching HRP-containing endosomes is determined indirectly.
Density-shifted or nonshifted control cells were solubilized in SDS,
and density-shifted material (endosomes) was removed by centrifugation.
Radiolabeled receptors were immunoprecipitated from the supernatants,
and immunoprecipitates were resolved by SDS-PAGE followed by
autoradiography. Receptor band intensity was quantified by
densitometric scanning. If metabolically labeled receptors reach
HRP-containing endosomes, a loss in radiolabeled receptors is observed
following cell solubilization and receptor immunoprecipitation from
density-shifted cells (+H2O2), by comparison to
control cells that were not shifted (
H2O2)],
where X is the band signal of pIgR in density-shifted
(+H2O2) and nonshifted
(
