(Received for publication, March 17, 1997, and in revised form, May 20, 1997)
From the Department of Pharmacology, Vanderbilt University Medical Center, Nashville, Tennessee 37232
G-protein-coupled receptors demonstrate differing
trafficking itineraries in polarized Madin-Darby canine kidney (MDCK
II) cells. The 2A adrenergic receptor
(
2AAR) is directly delivered to the basolateral
subdomain; the A1 adenosine receptor (A1AdoR) is apically enriched in its targeting; and the
2BAR
subtype is randomly delivered to both domains but selectively retained
basolaterally (Keefer, J. R., and Limbird, L. E. (1993) J. Biol. Chem. 268, 11340-11347; Saunders, C., Keefer, J. R.,
Kennedy, A. P., Wells, J. N., and Limbird, L. E. (1996) J. Biol. Chem. 271, 995-1002; Wozniak, M., and Limbird, L. E. (1996) J. Biol. Chem. 271, 5017-5024). The present
studies explore the role of the polarized cytoskeleton in localization
of G-protein-coupled receptors in MDCK II cells. Nocodazole or
colchicine, which disrupt microtubules, did not perturb lateral
localization of
2AR subtypes but led to a relocalization the A1AdoR to the basolateral surface, revealed by
immunocytochemical and metabolic labeling strategies. Conversely, the
apical component of the random delivery of
2BAR was not
affected by these agents, suggesting microtubule-dependent
and -independent apical targeting mechanisms for G-protein-coupled
receptors in polarized cells. Apparent rerouting of the apically
targeted A1AdoR was selective for microtubule-disrupting
agents, since cytochalasin D, which disrupts actin polymerization, did
not alter A1AdoR or
2BAR localization or
targeting. These data suggest that multiple apical targeting mechanisms
exist for G-protein-coupled receptors and that microtubule-disrupting agents serve as tools to probe their different trafficking
mechanisms.
The coordinated and vectorial functioning of polarized cells
mediated by endogenous and exogenous ligands depends on the
availability of appropriate receptors at the particular surface domains
to which the ligand has access. We are interested in elucidating the
mechanisms by which G-protein-coupled receptors
(GPCR)1 attain their localization in renal
epithelial cells, using polarized MDCK II cells as a model system. We
have demonstrated that the G-protein-coupled 2AAR is
delivered directly to and retained on the basolateral subdomain of
renal epithelial cells (1), whereas the A1AdoR
predominantly is apically targeted (65-83%) in renal epithelial cells
(2). The trafficking itineraries of the
2BAR and
2CAR subtypes differ from that of the
2AAR: the
2BAR is randomly delivered to
both the apical and basolateral domains but selectively retained
basolaterally; the
2CAR is directly delivered
basolaterally, but a substantial fraction of the receptor population
remains in a cytoplasmic compartment at steady state (3). These
different targeting itineraries are summarized in Fig. 1.
Like receptors, including the GPCR described above, the cytoskeleton in a polarized cell also is nonuniformly distributed (Fig. 1A). It is thought that this disparate distribution of cytoskeletal proteins contributes to the polarized compartmentalization of proteins, since the development of a functionally and structurally polarized phenotype is paralleled by cytoskeletal polarization (4). In MDCK II cells grown in polarized culture, Ca2+-dependent induction of cell-cell contacts via E-cadherin leads to a gradual reduction of actin microfilaments on the basal surface and their apical enrichment, in parallel with an increased stability (and polarization) of microtubules, and an alignment of ankyrin and fodrin under the lateral subdomain (4).
Since apical versus basolateral targeting of membrane proteins appears to involve discrete post-trans-Golgi network vesicular populations (5), it is not surprising that different cytoskeletal elements also have been implicated in this differential transport of apically versus basolaterally targeted vesicles (e.g. see Refs. 11, 12, and 18). To date, however, the role of the cytoskeleton in the trafficking of GPCR has not been explored.
The present studies examine the impact of cytoskeletal disrupting
agents on the localization and delivery of GPCR targeted to different
surfaces in polarized cells. We examined the effect of agents that
disrupt the cytoskeleton (colchicine, nocodazole, cytochalasin D) or
vesicular transport (monensin, brefeldin A) on the steady-state
localization and delivery of basolaterally targeted 2AAR
and
2CAR, the randomly delivered but basolaterally retained
2BAR, and the apically targeted
A1AdoR. We postulated that the use of cytoskeletal agents
in combination with the study of GPCR possessing varying trafficking
itineraries (2, 3) would allow us to ascertain criteria for basolateral
versus apical targeting as well as direct versus
random delivery. The present data provide evidence that apical delivery
of the A1AdoR and
2BAR occurs by
microtubule-dependent and -independent mechanisms,
respectively, and that cytoskeletal disrupting agents may be valuable
tools for elucidating distinct direct apical delivery pathways as well as the molecular mechanisms that govern them.
Materials
8-Cyclopentyl-1,3-di-[2,3-3H]propylxanthine
([3H]DPCPX) (109 Ci/mmol), 35S-Express
protein labeling mixture (1200 Ci/mmol),
[3H]methoxyinulin (125.6 mCi/g), and
[-35S]dATP (1389 Ci/mmol) were from NEN Life Science
Products. The A1AdoR antagonist,
1,3-dipropyl-8-(4-sulfophenyl)xanthine (6), was kindly donated by Dr.
Jack N. Wells (Vanderbilt University). Biotin hydrazide and
streptavidin-agarose were from Pierce; protein A-purified 12CA5
monoclonal antibody was from Babco; gp135 (7) and EGF receptor (8)
monoclonal antibodies were generously donated by Dr. Peter Dempsey
(Vanderbilt University); the E-cadherin antibody was from the
Developmental Studies Hybridoma Bank, NICHD, National Institutes of
Health (Iowa City, IA); Cy-3 conjugated donkey anti-mouse IgG was from
Jackson Immunochemicals; monoclonal anti-
-tubulin was from Amersham
Corp.; and rhodamine-conjugated phalloidin was from Molecular Probes.
Nocodazole, monensin, colchicine, and lumicolchicine were from Sigma;
cytochalasin D and brefeldin A were from Calbiochem.
Polarized Culture of MDCK II Cells and Functional Confirmation of Intact Monolayers
MDCK II cells were maintained as described previously (2, 3). For polarity experiments, MDCK II cells were seeded at a density of 1 × 106 cells/24.5-mm polycarbonate membrane filter (Transwell chambers, 0.4-µm pore size, Costar, Cambridge, MA), and cultured for 5-8 days with medium changes every day. The A1AdoR-expressing cell lines were grown in the presence of the receptor antagonists, 60 µM theophylline and 100 µM 1,3-dipropyl-8-(4-sulfophenyl)xanthine, since previous studies had demonstrated that, if grown in their absence, A1AdoR-expressing cells grew as multicellular layers in Transwell culture (2). Prior to each experiment, the integrity of the monolayer was assessed by monitoring [3H]methoxyinulin leak (1).
Development of Permanent Transformants of MDCK II Cells
Permanent clonal cell lines of MDCK II cells were developed as
described previously (1-3). In each case, the first 9 amino acids
after the initiating methionine encode a hemagglutinin epitope (9)
recognized by the commercially available monoclonal antibody 12CA5
(Babco). The clonal cell lines evaluated in the present study include
TAG-2AAR (25 and 7 pmol/mg of protein),
TAG-
2BAR (10 and 3 pmol/mg of protein),
TAG-
2CAR (5 and 3 pmol/mg of protein) and
TAG-A1AdoR (37 pmol/mg of protein). The
2AAR
subtype was encoded by a porcine cDNA, the
2BAR and
2CAR subtypes were encoded by a rat cDNA, and the
A1AdoR was encoded by a canine cDNA. The species origin
does not alter the trafficking patterns of these receptors. For
example, the porcine
2AAR is targeted directly to the
basolateral surface in both canine MDCK II and porcine LLC-PK1 renal
epithelial cells (1), and the canine A1AdoR is targeted
apically in both MDCK II and LLC-PKI cells (2).
Receptor Binding Assays
MDCK II particulate preparations were prepared essentially as
described (1). A1AdoR were identified using
[3H]DPCPX (2), and 2AR were identified
with [3H]rauwolscine (1, 3) as described previously.
Treatment of Polarized MDCK II Cells with Agents to Disrupt Cytoskeleton or Vesicular Traffic
ColchicineColchicine is a microtubule-disrupting drug that
binds slowly to soluble tubulin heterodimers, reducing them to large
aggregates and rendering them incapable of polymerization for
microtubule growth (10). Most incubations with colchicine were
performed for 15 h, since the half-life of all four GPCR studied
on their enriched membrane surfaces is 10-12 h. Thus, a 15-h
incubation with colchicine permitted us to evaluate the localization of
receptors already delivered to the cell surface and those that were
under synthesis and delivery. Previous reports have established that increased time of incubation with colchicine beyond 4-6 h does not
lead to nonspecific mechanisms for this agent (11, 12). We conducted a
time course of the steady-state localization of the
A1AdoR-expressing cell line in the presence of 10 µM colchicine; the appearance of the A1AdoR
on the lateral surface was evident after only 4 h of treatment and
was maximal at 18 h (data not shown). Even at 18 h, however,
some trace apical staining of A1AdoR was detectable (see
"Results"), which could be due to incomplete microtubule
disruption, incomplete turnover of the already delivered receptor
population, or a microtubule-insensitive fraction of apically delivered
receptor. Immunocytochemical analysis of treated cells using an anti
-tubulin antibody confirmed that colchicine treatment of cells had
indeed disrupted the microtubule network. Since a change in MDCK II
cell morphology was noted after 24-48 h of colchicine treatment, these
longer incubations were not used in the present studies.
Nocodazole disrupts microtubules via a molecular
mechanism different from that of colchicine (10). Nocodazole binds
rapidly to microtubule subunits and prevents heterodimers from
repolymerizing, at either 37 or 4 °C. Incubation at 4 °C
accelerates nocodazole-effected depolymerization, at least in MDCK (7)
and Caco-2 (12) cells. MDCK II cells in Transwell culture were treated
for 15 h, as described above for colchicine. Just as for
colchicine, the ability of nocodazole treatment to disrupt microtubules
was confirmed by immunocytochemical analysis with the anti--tubulin
monoclonal antibody (Amersham).
The effects of nocodazole (33 µM, equivalent to 10 µg/ml) on GPCR delivery to the cell surface were performed as follows, based on earlier reports for nocodazole treatment in MDCK cells (18): on the day of the experiment, medium was replaced with 4 °C medium and incubated for 30 min, followed by 4 °C Cys/Met-free medium with or without nocodazole for 1 h, followed by 37 °C Cys/Met-free medium with or without nocodazole for 1-1.5 h, followed by [35S]Cys/Met metabolic labeling at 37 °C for the desired "pulse" time with or without nocodazole.
Cytochalasin DThe actin polymerization inhibitor, cytochalasin D (2 µM), was incubated with polarized MDCK II cells for 15 h prior to fixation and immunocytochemical analysis. Disruption of the actin cytoskeleton was confirmed by staining control and treated cells with rhodamine phalloidin, which reveals the actin network.
Brefeldin A (BFA)BFA is a fungal metabolite that has been
shown to fuse the endoplasmic reticulum with the cis, medial, and trans
cisternae of the Golgi apparatus but not with the trans-Golgi network
(13). BFA has been demonstrated to both enhance transcytosis (14, 15)
and inhibit transcytosis (16) or targeting (17) and was tested in these
studies to explore the possible contribution of the transcytotic
pathway to the apical delivery of A1AdoR and/or the apical
component of the random delivery of 2BAR. Brefeldin A
was added to a final concentration of 3.5 µM, as
described previously for MDCK II cells (18, 19).
Monensin is a cationic ionophore that exchanges Na+ for H+, thereby dissipating transmembrane pH gradients. Although multiple intracellular consequences of monensin treatment might be expected, most widely reported are the effects of monensin to block cell surface receptor recycling of internalized molecules to the cell surface (20-23). Monensin was evaluated at a final concentration of 1.4 µM, as described previously for MDCK II cells (18, 24).
Steady-state Localization of GPCR and of Cell Surface Proteins by Immunolocalization
Immunostaining of cells grown in Transwell culture was performed as described previously (2, 3) with the following concentration of primary antibody: a 1:50 dilution of 12CA5 primary antibody, purified as described previously (3), for the localization of hemagglutinin epitope-tagged GPCR (1); 15 µg/ml mouse monoclonal EGF receptor antibody (25); a 1:10 dilution of mouse monoclonal gp135 antibody (7); or a 1:8 dilution of mouse monoclonal E-cadherin antibody (RR1, Developmental Studies Hybridoma Bank). Except when indicated otherwise, the antibody-containing and antibody wash buffers contained 0.1% Triton X-100, to permit detection of epitope either on the cell surface or in the cell interior. Treatment with the secondary Cy3 conjugated donkey anti-mouse IgG (1:200) was performed as described (2, 3). Samples were visualized by confocal microscopy on a Zeiss Axiovert 135 Micro Systems LSM (Germany). The samples were first visualized in the xy plane and then in the xz plane. In the images shown, the bottom 3/4 represent the xy plane, the conventional view of the cells as one looks down upon them. The white line that is shown in the xy plane confocal images indicates where the laser took a cross-section of the cells to generate the z scan. The top 1/4 of the images represents the xz plane (or z scan), the cortical section perpendicular to the plane of the cell layer. Images were analyzed using Showcase software on a Silicon Graphics Iris Indigo workstation.
Metabolic Labeling/Biotinylation Strategy for Determining Surface Delivery of GPCR
The amount of newly synthesized receptor delivered to the apical versus basolateral surface of polarized, metabolically labeled MDCK II cells grown in Transwell culture was quantified by surface biotinylation of either the apical or basolateral surface with NHS-biotin followed by isolation of radiolabeled receptor on the biotinylated surface by sequential Protein A- and streptavidin-agarose chromatography. These procedures were performed essentially as described (1), except that wherever drugs were used, they were added in the Cys/Met-free medium for the duration that preceded and included pulse labeling. Briefly, the cells were pretreated with Cys/Met-free medium for 2 h, followed by [35S]Cys/Met metabolic labeling for the durations given in the figure legends. If the experiment was designed to examine the retention time of the receptor on the cell surface, then chase times were included after the pulse with or without drug.
To examine the
effect of colchicine (10 µM), nocodazole (33 µM) and cytochalasin D (2 µM) on the
steady-state localization of the GPCR, 2AAR,
2BAR,
2CAR, and A1AdoR in
renal epithelial cells (Fig. 2), cell lines permanently
expressing epitope-tagged versions of these receptor populations were
incubated for 15 h with the drugs of interest as described under
"Experimental Procedures." Since the receptor half-life of the
2AR subtypes on the basolateral surface (3) and that of
the A1AdoR on the apical surface (2) is approximately
10-12 h, this 15-h incubation with cytoskeletal disrupting agents
provided optimal opportunity to explore the impact of these agents on
nascent as well as pre-existing receptor populations. Previous studies
have demonstrated that overnight incubations like those we utilized in
the present studies do not lead to nonspecific effects of these drugs;
e.g. although the maximal effects of colchicine often are
noted after only 4 h of incubation with 10 µM
colchicine (26), others have shown that treatment with nocodazole
requires at least 6 h for complete microtubule depolymerization
(27). Each experiment in the present study was preceded with an
assessment of transepithelial leak to confirm the integrity of the
polarized MDCK II cell monolayer under control and treated conditions
(see "Experimental Procedures"). Similarly, the impact of these
agents on endogenous polarized "marker" proteins in MDCK II cells
also was examined (Fig. 3).
As shown in Fig. 2, the most dramatic effect of the cytoskeletal
disrupting agents is seen for the steady-state localization of the
A1AdoR, where treatment with both of the
microtubule-disrupting agents, colchicine and nocodazole, altered the
apical staining pattern characteristic of the wild-type receptor under
control conditions to a basolateral staining profile. This change in
A1AdoR polarization is evident in both the xy
plane confocal images and the xz plane images, which
effectively provide a laser "slice" through the polarized MDCK II
cells (Fig. 2). In contrast to the dramatic changes in
A1AdoR polarization, microtubule disruption by either
colchicine or nocodazole did not alter the steady state basolateral
polarization of any of the three 2AR subtypes.
Although the microtubule-directed agents, colchicine and nocodazole,
did not redirect the steady-state localization of the 2AR subtypes, they nonetheless altered the relative
surface:internal compartment localization of the
2BAR
(Fig. 2). The intracellular localization of the
2BAR is
interpreted to result from ligand-induced internalization (28) or from
catecholamines in serum-containing medium.2
The fluorescence intensity of the lateral staining pattern for the
2BAR in the presence of colchicine was increased
dramatically concomitant with a decrease in intracellular staining of
this receptor. These findings suggest that in the presence of
colchicine, and to a lesser extent in the presence of nocodazole (Fig.
2), the fraction of the
2BAR population that existed in
the cell interior was now enriched on the lateral surface. In contrast, no change in relative lateral membrane staining of the
2AAR or
2CAR was observed following
colchicine or nocodazole treatment. The intracellular enrichment of the
2CAR subtype at steady state has been reported
previously in HEK293 cells (47) and, in later studies, in polarized
MDCK II cells (3). The
2CAR internal compartment can be
distinguished functionally and morphologically from endocytosing cell
surface GPCR (47). The relative surface:internal distribution of the
2CAR subtype was resistant to treatment with microtubule
disrupting agents. The existence of internal
2BAR and
2CAR in independent compartments may explain why
colchicine or nocodazole influences
2BAR redistribution
without impact on the intracellular
2CAR pool.
The apparent increase in 2BAR density at the lateral
surface detected morphologically was paralleled by an increase in
functional
2BAR density detected using radioligand
binding assays (Fig. 4). Saturation binding analyses
revealed that the increase in functional
2BAR binding
was due to an increase in maximal receptor density
(Bmax = 1 ± 0.06 pmol/mg of protein for
control versus Bmax = 2.96 ± 0.67 pmol/mg
of protein for colchicine-treated cell lines; n = 3;
p < 0.05), with no change in
2BAR
receptor affinity for the radiolabeled antagonist
[3H]rauwolscine (Kd = 1.86 ± 0.18 nM for cells treated under control conditions
versus 2.23 ± 0.16 nM in
colchicine-treated cells). This effect depended on the
microtubule-disrupting properties of colchicine, since
-lumicolchicine, a chemical analog of colchicine that does not
disrupt microtubules, had no effect on radioligand binding density
(data not shown). This marked increase in binding density was uniquely
observed for this
2AR subtype; although effects of
colchicine on
2CAR binding were observed, they were significantly smaller (approximately a 20% increase in binding; Fig.
4). The absence of any loss of binding for any of the GPCR following
treatment with these microtubule-disrupting agents provides additional
evidence that these agents are not having nonspecific deleterious
effects on the MDCK II cells under the conditions of our experiments.
We also evaluated the effect of microtubule-disrupting agents on the localization of endogenous MDCK II proteins that serve as "markers" of the polarized state: EGF receptor (localized laterally) (25), gp135 (localized apically) (7), and E-cadherin (a quintessential marker of cell polarity, localized laterally and involved in the development of polarity in MDCK cells) (29). Neither colchicine nor nocodazole had a significant effect on the localization of the basolateral markers (Fig. 3). Interestingly, the apical marker gp135 also remained enriched apically, but its distribution on the apical surface was modified in about 50% of the cells, creating an apparently "black" cellular interior. Earlier reports have noted a similar clustering of gp135 occurs in MDCK cells treated with cytochalasin D (7). It is unclear why, however, nocodazole but not colchicine caused this "black hole" phenomenon; perhaps it reflects differing kinetic effects of those agents.
Microtubule-disrupting Agents Cause Enriched Delivery of Nascent A1AdoR to the Basolateral SurfaceAs shown in Fig. 2,
disruption of microtubules with colchicine or nocodazole led to
relocalization of A1AdoR from a predominantly apical
localization to a predominantly lateral expression pattern. To
determine whether this A1AdoR relocalization results from
altered delivery to or retention on the basolateral versus
the apical surface, we examined the effects of colchicine and
nocodazole on delivery and turnover of this receptor in metabolically
labeled MDCK II cells (Fig. 5). Nocodazole treatment
reversed the predominant apical delivery (70:30, apical:basolateral) of
the A1AdoR such that a major portion of the apical receptor
population is delivered basolaterally (40:60, apical:basolateral) after
treatment (Fig. 5A). This reversal of the targeting pattern
for the A1AdoR seen with nocodazole treatment also is seen
with colchicine, where the apical:basolateral ratio of the
A1AdoR delivery went from 63:37 under control conditions to
45:55 following colchicine treatment (Fig. 5B). In the
autoradiogram shown, the apparent reversal appeared more dramatic for
nocodazole than colchicine. As mentioned earlier, these subtle
differences in quantitative effects may reflect kinetic differences in
colchicine versus nocodazole action (10). The studies shown
in Fig. 5A follow a 90-min pulse labeling with
[35S]Cys/Met-containing medium, but indistinguishable
findings were observed following a 60-min pulse. The colchicine or
nocodazole-induced basolateral enrichment of the A1AdoR was
retained over time, as evidenced following pulse-chase studies (data
not shown). These data suggest that microtubule disruption alters the
delivery of the direct apically targeted A1AdoR and that
the morphological manifestation of the steady-state redistribution to
the lateral subdomain is due, at least in part, to an enhanced delivery
to the basolateral surface.
Nocodazole and colchicine reverse the
polarity of apically delivered A1AdoR. A, MDCK
II cells expressing epitope-tagged A1AdoR grown in
Transwell culture in the presence of A1AdoR antagonists (see "Experimental Procedures") were metabolically labeled with 1 µCi/µl of [35S]Met/Cys protein labeling mix (150 µl) for 1 h in the absence or presence of nocodazole (33 µM) and then harvested and processed using sequential
immunoprecipitation and streptavidin-agarose chromatography as
described under "Experimental Procedures." The autoradiogram shown
is for an SDS-polyacrylamide gel exposed for 3 weeks to Kodak X-Omat
film between two Quanta III screens at 70 °C. Gel slices
corresponding to the position of the A1AdoR signal on
autoradiograms were excised and counted in 10 ml of NEN-963
scintillation fluid in a
-scintillation counter. In this representative autoradiogram, under control conditions, the
apical:basolateral ratio for A1AdoR was 70:30; in the
presence of nocodazole, it was 40:60. The findings from multiple
individual experiments are summarized as follows. The mean
apical:basolateral ratio under control conditions was 68:32, and for
nocodazole treatment it was 44:56 (n = 5).
B, MDCK II cells expressing epitope-tagged A1AdoR grown in Transwell culture were metabolically
labeled in the absence or presence of colchicine (10 µM)
with 1 µCi/µl of [35S]Met/Cys protein labeling mix
(150 µl) for 1 h and then processed as in Fig. 5A. In
this representative autoradiogram, the apical:basolateral ratio
A1AdoR under control conditions was 63:37; in the presence of colchicine, it was 45:55. The mean apical:basolateral ratio under
control conditions was 62:38, and for colchicine treatment it was 42:58
(n = 2). C, the retention time of the
A1AdoR on the apical surface in the presence of
microtubule-disrupting agents was not significantly different from
control. The loss of the receptor from the apical surface in the
presence or absence of colchicine or nocodazole was assessed by means
of "pulse-chase" experiments. MDCK II cells expressing
epitope-tagged A1AdoR grown in Transwell culture were
metabolically labeled in the absence or presence of colchicine (10 µM) or nocodazole (33 µM) with 1 µCi/µl
[35S]Met/Cys protein labeling mix for 1 h, chased for 8 (t0) and then
20 h (t1), and then processed as in Fig.
5A. This was done twice for each drug at these time points.
The ranges of the percentage remaining on the apical surface at 20 h were as follows: 50-67% (average, 57.3%; n = 3)
for untreated (control) conditions, 51-79% (average, 65%;
n = 2) in the presence of colchicine, and 57-58% (average, 57.5%; n = 2) in the presence of
nocodazole.
Since steady-state localization reflects a balance of the delivery or targeting to a surface domain and the retention on that surface domain, we examined the effects of colchicine or nocodazole on the retention of the A1AdoR on the apical surface. Neither colchicine nor nocodazole altered the half-life of this receptor on the apical surface (Fig. 5C). These findings suggest that the relative basolateral enrichment of the A1AdoR following treatment with microtubule-disrupting agents is not due to accelerated apical turnover of the A1AdoR. However, pulse-chase studies undertaken to evaluate whether the altered apical delivery profile was detected at multiple time points also revealed that the A1AdoR was retained on the basolateral surface for longer periods of time (t1/2 > 30 h) when compared with the apical retention in the presence of nocodazole (t1/2 = 13 h); the t1/2 of the A1AdoR in the presence of nocodazole is indistinguishable for the t1/2 in control, untreated cells (2). Previous studies of aminopeptidase N and sucrose-isomaltase in polarized Caco-2 intestinal epithelial cells also have noted a preferential retention of proteins on the basolateral versus apical surface in the presence of microtubule-disrupting agents (26).
The Apical Component of the Random Delivery of NascentOur observation that the A1AdoR was apparently
rerouted from the apical to the basolateral surface by
microtubule-disrupting agents raised the question of whether the apical
component of the random delivery of the 2BAR also might
be modified following similar treatments. As shown in Fig.
6A, the equivalent delivery of the
2BAR to both the apical and basolateral surfaces is
maintained in the presence of nocodazole: 47:53 apical to basolateral
under control conditions versus 42:58 in the presence of
nocodazole. Unexpectedly, the [35S]Cys/Met signal for
both the apical and basolateral
2BAR is increased 1.5-2
times that of the signal under control conditions in the presence of
nocodazole (Fig. 6A). Similarly, a 2-3-fold increase in
metabolically labeled
2BAR at both the apical and basolateral surfaces was observed in the presence of colchicine (Fig.
6B). This unexpected and unexplained increase in receptor delivery of the
2BAR to the cell surface in the presence
of microtubule-disrupting agents may be due to a generalized increase
in protein synthesis. Consistent with this hypothesis is our finding
that a 15-h (but not 3-h) treatment with cycloheximide (data not shown)
inhibited the colchicine-induced increase in radioligand binding,
directly correlating with the time-dependent effects of
colchicine to increase
2BAR radioligand binding, as
described earlier in Fig. 4. The direct targeting of the
2AAR and the
2CAR to the basolateral surface was not altered by the microtubule-disrupting drugs, nocodazole and colchicine (data not shown).
Impact of Cytochalasin D-mediated Disruption of the Actin Cytoskeleton on the Localization of GPCR and Proteins That Serve as Markers of the Polarized State in MDCK II Cells
We have shown
previously that the 2AAR subtype is delivered directly
to the basolateral domain (1). This apparent direct basolateral
delivery, however, might instead result from active exclusion of this
receptor subtype from the apical surface due to, for example, a
cytoskeletal "screen" (30). Consequently, we examined the effect of
cytochalasin D, which disrupts the actin-based cytoskeleton, on the
localization of the
2AAR and other GPCR. Cytochalasin D
did not alter the lateral localization of the
2AAR, suggesting that the direct basolateral delivery of this subtype is not
due to exclusion from the actin-rich sub-apical surface. Interestingly,
however, the
2BAR acquired a random localization at
steady state following exposure to cytochalasin D, revealed by
morphological detection of the
2BAR subtype on both the
apical and basolateral surfaces (Fig. 2). Morphological analysis
demonstrated that cytochalasin D also caused a redistribution of the
basolaterally expressed endogenous EGF receptor and E-cadherin.
Following cytochalasin D treatment, the EGF receptor appeared to be
distributed on all surfaces as well as in the cell interior, and
E-cadherin was apically as well as laterally localized (cf.
x-z scans, Fig. 3). Thus, it would appear that the
2BAR, albeit not the
2AAR, mirrors the
cytochalasin D-effected redistribution pattern of endogenous MDCK II
proteins that typically are enriched on the lateral subdomain. Cytochalasin D also caused the apical marker protein gp135 to localize
in a punctate, aggregated form on the apical surface (Fig. 3), as
reported previously (7), without significantly altering the
distribution of the apically targeted and localized A1AdoR.
Since cytochalasin D appeared to alter the steady-state localization of
a variety of endogenous MDCK proteins, we did not pursue this agent
further as a tool to reveal distinct mechanisms for delivery and
retention of GPCR in polarized renal epithelial cells. However, it
should be noted that we evaluated the effects of cytochalasin D on GPCR
radioligand binding and found that binding was not altered; the
percentage of control binding for the GPCR following treatment with
cytochalasin was 94 ± 7% for 2AAR, 115 ± 29% for
2BAR, 82 ± 25% for
2CAR,
and 115 ± 19% for A1AdoR (n = 3 for
all GPCR). Thus, despite the impact of cytochalasin D on the
morphological localization of both heterologously expressed GPCR and
endogenous proteins that serve as markers for the polarized state, this
agent did not have deleterious effects on the functional binding
properties of the
2AR subtypes or of the
A1AdoR.
Table I summarizes the morphological consequences of treatment with the microtubule-disrupting agents, colchicine and nocodazole, and of the actin cytoskeleton-disrupting agent, cytochalasin D, on the localization of GPCR and of polarized surface marker proteins in MDCK II cells.
|
Monensin
is a Na+/H+ electroneutral ionophore that
dissipates transmembrane pH gradients. This agent has been demonstrated
to inhibit receptor-mediated endocytosis (31) or the recycling (20-23)
of cell surface receptors and has been used as a tool to prevent
resensitization of GPCR (23). We evaluated the effects of monensin (1.4 µM) on the steady-state localization of GPCR to determine
whether receptor endocytosis/recycling played a demonstrable role in
their ultimate polarization (4-fold higher concentrations of this agent
had similar effects on receptor localization; data not shown). Monensin
altered the relative surface:internal compartment distribution of the
2BAR (Fig. 7A) without any
appreciable effect on the localization of the other three GPCR examined
(data not shown). Strikingly, there is an increased lateral staining
intensity for the
2BAR in the presence of monensin
compared with the control. As we observed with colchicine, these
effects of monensin on lateral surface enrichment of the
2BAR are specific for this receptor subtype and are not
seen for the other two
2AR subtypes. The localization of
the
2BAR in the absence of 0.1% Triton X-100 (surface
receptors only) demonstrates that the
2BAR is enriched solely on the lateral surface at steady state, both in the absence and
presence of monensin. However, monensin dramatically increased the
fraction of
2BAR detected intracellularly, as seen in
confocal images of cells incubated with anti-epitope antibody in the
presence of 0.1% Triton X-100 (Fig. 7A), a manipulation
that permeabilizes cells and allows the entire receptor population to
be evaluated (surface and intracellular). Thus, there is receptor
"trapped" inside the cells following treatment with this drug.
These findings are analogous to the effects of monensin on the VSV
G-protein trafficking in MDCK cells reported previously (24). Monensin did not interfere with the polarization of the endogenous proteins that
serve as surface markers for MDCK II cells: the EGF receptor, gp135,
and E-cadherin (Fig. 7B). Also, a 4-fold higher
concentration of monensin (5.6 µM) similarly did not
alter the polarization of these marker proteins (data not shown).
It is possible that the intracellular accumulation of the
2BAR subtype following monensin treatment results from a
blockade of recycling (as a potential first step in apical to
basolateral transcytosis and thus rerouting) of the randomly delivered
2BAR following its rapid removal
(t1/2 = 15 min; Ref. 3) from the apical surface. As
shown in Fig. 8, monensin did not alter the random
delivery of the
2BAR, although it did increase the
amount of nascent receptor delivered to both cell surfaces, again
similar to findings with colchicine (and to some extent, nocodazole). Longer pulse time points were also evaluated; however, there was no
change in the random targeting of
2BAR other than
increased signal intensity delivered to both surfaces of polarized MDCK cells (data not shown). Monensin also did not alter the targeting of
the A1AdoR, as assessed in metabolic labeling and nascent
receptor delivery studies (data not shown). We were not successful in
monitoring the surface delivery of nascent
2AAR and
2CAR in the presence of monensin, perhaps because the
drug interfered with the glycosylation of these two receptor subtypes
(13), thereby rendering it incapable of reaching the cell surface (the
2BAR subtype is not glycosylated (32)). Monensin had no
significant effect on radioligand binding for the GPCR: percentage of
control binding for the GPCR following treatment with monensin was
86 ± 14% for
2AAR, 126 ± 15% for
2BAR, 156 ± 22% for
2CAR, and
110 ± 20% for A1AdoR (n = 3-4 for
all GPCR).
Brefeldin A inhibits vesicular transport and secretion by selective
disruption of the Golgi apparatus. It often has been used to explore
trafficking itineraries in MDCK cells, and the results have been
conflicting. BFA perturbs basolateral to apical transcytosis (14,
16-18, 33) but also enhances apical endocytosis (14) and transcytosis
(14, 15). We were interested in determining whether or not the apical
trafficking of the A1AdoR might result from transcytosis,
thus accounting for the different consequences of microtubule
disruption on apical delivery of the A1AdoR
versus the 2BAR. A concentration of 3.5 µM BFA did not change the steady-state localization of
any of the GPCR or any of the "marker" proteins (data not shown).
As MDCK cells have been reported to be relatively resistant to BFA, we
also tried a 10-fold higher concentration. At 35 µM BFA,
the findings for the A1AdoR,
2AAR, and
2BAR were similar to those in the presence of 3.5 µM BFA. Only the
2CAR localization was
slightly altered following exposure to 35 µM BFA, such
that the intracellular compartment described previously (3, 47) was no
longer as visible (data not shown). Although we do not know the
significance of the ability of BFA to eliminate detection of the
intracellular compartment of the
2CAR, the data suggest
that the intracellular pool of the
2CAR either is in a
different vesicular compartment than internal pools of
2BAR (see "Discussion") or is delivered to or
maintained in the cell interior via a different, BFA-sensitive
mechanism. BFA had no significant effect on radioligand binding for the
GPCR.
Table I summarizes the morphological consequences of treatment with the agents that interfere with vesicular traffic, monensin and brefeldin A.
It is important to understand the molecular mechanisms that govern
polarized expression of epithelial cell proteins, since this polarity
is an intrinsic part of the vectorial functioning of these cells. The
importance of receptor localization in signal transduction is inferred
from the number of pathophysiologic states that result from
mislocalized proteins. In autosomal polycystic kidney disease and renal
ischemia, the Na+,K+-ATPase and epidermal
growth factor receptors of the basolateral plasma membrane are
redistributed from the lateral surface to the apical membrane (34). In
familial hypercholesterolemia, the low density lipoprotein receptor is
unable to internalize and subsequently cannot transport cholesterol for
storage as cholesterol esters (35). In cystic fibrosis, the most common
F508 mutation restricts the CFTR chloride ion transporter to the
endoplasmic reticulum and prevents its delivery to the cell surface
(34). Mislocalized GPCR also contribute to disease. For example, one form of retinitis pigmentosa results from intracellular trapping of the
G-protein-coupled receptor for light, rhodopsin (51). In addition,
diabetes insipidus can result from a variety of mutations in the
V2 vasopressin receptor, many of which lead to receptor misfolding and lack of surface delivery (36). Understanding the
mechanisms that govern the trafficking of receptors and signal transducing proteins should provide insights into the operation of
receptor-mediated signal transduction events in polarized cells under
physiologic conditions and reveal potential molecular culprits in
pathophysiologic states.
We are interested in elucidating the mechanisms conferring the
basolateral or apical polarization of GPCR. Our previous studies have
shown that the basolaterally localized 2AR subtypes have differing trafficking itineraries that lead to their ultimate shared
polarity at steady state (1, 3); in contrast, the A1AdoR is
apically targeted and resides principally on the apical surface at
steady state (2). In this study, we used experimental manipulations
that previously have been demonstrated to disrupt the cytoskeleton or
to interfere with vesicular transport of proteins as tools to explore
possible differences in apical versus basolateral polarization mechanisms in more detail.
The present studies have revealed that the apical delivery of the
A1AdoR and 2BAR is differentially affected
by microtubule-disrupting agents. Thus, apical delivery of the
A1AdoR, but not of the
2BAR, is inhibited by
microtubule disrupting agents, leading to enriched delivery of the
A1AdoR to the basolateral surface (Fig. 5, A and B) and localization there at steady state (Fig. 2). Whether
or not this relocalization is due to the direct rerouting of
A1AdoR delivery from the apical to the basolateral surface
or, alternatively, to blockade of basolateral to apical transcytosis of
the A1AdoR is not known. However, it should be noted that previous
studies examining the delivery of metabolically labeled
A1AdoR to the apical surface are consistent with the
interpretation that this is directly delivered to the apical surface
(2). The ability of colchicine and nocodazole to disrupt apical
delivery of the A1AdoR is consistent with previous reports
describing the requirement of an intact microtubule matrix for apical
delivery of a number of proteins (12, 37-39) and lipids (40) in
epithelial cells of the intestine, another absorptive epithelium. In
polarized MDCK cells, apical but not basolateral membrane proteins (18, 41, 42) and apically but not basolaterally secreted proteins (43) were
reported to be missorted to the basolateral surface following exposure
to microtubule-disrupting drugs.
It is reasonable to postulate that disruption of A1AdoR
delivery to the apical surface by microtubule-directed agents reveals a
role for these cytoskeletal elements in directing
A1AdoR-containing vesicles to that surface, by analogy with
previous reports for apically targeted membrane proteins in MDCK cells
described above (12, 37-39). In contrast, the apical component of the
random delivery of the 2BAR must utilize a distinct,
microtubule-independent mechanism to achieve apical trafficking.
Although a formal possibility is that the apical delivery of the
2BAR is due to transcytosis from the basolateral
membrane, two lines of evidence argue against this hypothesis. First,
the half-life of the
2BAR on the basolateral surface is
10-12 h, whereas its half-life on the apical surface is 10-15 min.
These data are more consistent with a random delivery, followed by an
internalization of the apical receptor and possible (although there is
no direct evidence) rerouting to the basolateral surface (3), as has
been noted for Na+/K+-ATPase in certain clones
of MDCK II cells (44). Second, we observed no effect of brefeldin A on
GPCR polarization, previously shown either to interfere with (45) or to
enhance (14, 15) transcytosis in MDCK cells. The mechanism contributing
to the apical component of the random delivery of the
2BAR remains to be elucidated. However, these findings
are of particular importance in establishing that there is a greater
complexity of direct vesicular traffic to a given surface of polarized
cells than simply the distinction between apical and basolaterally
destined vesicles emerging from the trans-Golgi network (46) or,
alternatively, transcytosis (42).
Another important contribution of the present studies is that they
provide an additional line of evidence (47) that the intracellular
compartment of the 2BAR is distinct from that occupied by the
2CAR. As indicated under "Results," the
fraction of the
2BAR detected in the cell interior (Fig.
2) is thought to result from the continuous endocytosis/recycling of
this receptor subtype in response to exogenous agonist or to
catecholamines in serum-containing medium2; consistent with
this interpretation are the present findings that treatment with
monensin, an agent that blocks receptor recycling to the surface
(20-22), increased the intracellular localization of the
2BAR subtype. In a seemingly reciprocal manner,
colchicine enhanced the basolateral surface labeling of the
2BAR with a concomitant reduction in the fraction of
detectable intracellular receptor (Fig. 2). This surface enrichment
detected morphologically was paralleled by dramatic increases in the
density of functional
2BAR binding sites (Fig. 4).
Neither of these colchicine-induced phenomena occurred with the
2CAR subtype. A similar discordance of
2BAR localization in comparison with that of an
endocytosing GPCR, the
2-adrenergic receptor, was
observed in HEK 293 cells (47). It should be noted that nocodazole is
quantitatively less effective than colchicine in eliciting the apparent
redistribution of
2BAR to the lateral surface (Fig. 2)
and in enhancing receptor binding density (Fig. 4), consistent with
earlier reports that these agents, although both disrupting
microtubules, can elicit quantitatively different effects on cells
(12). The unique increase in signal intensity in the autoradiograms of
metabolically labeled
2BAR-expressing cells seen in the
presence of nocodazole and colchicine, as well as in the presence of
monensin, are intriguing. However, at present, we have no data that
reveal a molecular basis for these findings.
Although we evaluated the effects of the actin cytoskeleton-disrupting
agent, cytochalasin D, on the polarization of A1AdoR, 2AR subtypes, and endogenous markers for polarization in
MDCK cells, we chose not to pursue these studies beyond our
morphological characterization, since we noted that the localization of
E-cadherin, which initiates and then maintains polarization of
epithelial cells (4), is perturbed by treatment with cytochalasin D
(Fig. 3). Interestingly, the staining profile of the
2BAR is the one most altered by cytochalasin D, and this
is the only receptor of the four studied known to undergo extensive
ligand-induced endocytosis. Previous studies have shown that
cytochalasin D selectively inhibits apical but not basolateral
endocytosis of VSV G-protein, cationized ferritin, and Lucifer yellow
in MDCK cells (48). It also is interesting that cytochalasin D-induced
relocalization of the basolateral GPCR evaluated in these studies
parallels that of the basolateral markers, EGF receptor and E-cadherin,
whereas the morphological changes in distribution of the apically
targeted and polarized A1AdoR are paralleled by those of
the gp135 apical marker. These findings with the actin-disrupting agent
cytochalasin D are in distinct contrast to the consequences of
disruption of the microtubule-dependent cytoskeleton by
colchicine and nocodazole, which led to differential effects on the
localization of the GPCR with no remarkable changes in the localization
of gp135, E-cadherin, or the EGF receptor.
The present studies provide the first evidence that multiple pathways
exist for the direct targeting of G-protein-coupled receptors of the
Gi/Go-coupled structural family, the
A1AdoR and 2BAR, to the apical surface of
polarized renal epithelial cells. These findings extend earlier
demonstrations that apical and basolateral proteins are packaged into
different vesicular carriers following egress from the trans-Golgi
network (49) and suggest that there must be at least one other
bifurcation at the level of apically directed vesicles. The ability of
microtubule-disrupting agents to discriminate between the apical
delivery mechanisms for the A1AdoR and the
2BAR suggest that these drugs will be powerful tools in delineating the molecular mechanisms underlying these diverse
apical delivery pathways. Elucidation of the molecular mechanisms
responsible for microtubule-dependent
versus -independent apical delivery may provide novel
therapeutic advantage in settings where apical delivery defects
contribute to human pathology, as in cystic fibrosis (50).
We thank Carol Ann Bonner for technical assistance in the development and maintenance of MDCK II cell lines and for the assessment of receptor binding density in MDCK II cells. We thank Dr. Peter Dempsey for the gift of the gp135 and EGF receptor monoclonal antibodies. We also thank Dr. Leigh B. MacMillan for critical reading of this manuscript. We are also grateful to Dr. Tom Jetton for kind assistance with the use of the confocal microscope and discussions concerning immunocytochemical techniques. We are also very appreciative to the members of the Limbird laboratory for helpful discussions and assistance during these experiments. Thanks also go to Dr. James Nelson for a careful critique and suggestions for these studies. Finally, we thank Dr. Enrique Rodriguez-Boulan for providing our laboratory with the parental MDCK II cells.