Meeting of the Apical and Basolateral Endocytic Pathways of the Madin-Darby Canine Kidney Cell ...

14 Pages·2002·5.51 MB·English

Meeting of the Apical and Basolateral Endocytic syawhtaP of the
Madin-Darby Canine Kidney Cell in Late Endosomes
Robert .G Parton, Kristian Prydz, Morgane *,lesmoB iaK ,snomiS and Gareth Griitiths
European Molecular Biology Laboratory, Cell Biology Program, Post fach .01 2209, D-6900 Heidelberg, Federal Republic of Germany
and ¢ ER64-Centre National de la Recherche Scientifique, Etats Li6s Mol6culaires, 60057-F Paris, France
Abstract. Electron microscopic approaches have been antibodies against the cation-independent mannose-6-
used to study the endocytic pathways from the apical phosphate receptor (MPR) on cryosections. With in-
and basolateral surface domains of the polarized epi- creasing incubation times, markers passed from these
thelial cell, MDCK strain I, grown on polycarbonate MPR-positive structures into a common set of MPR-
filters. The cells were incubated at 37°C in the pres- negative lysosomes that were mainly located in the ap- D
o
ence of two distinguishable markers administered ical half of the cell. wn
lo
separately to the apical or the basolateral domain. Ini- A detailed quantitative analysis of the endocytic ad
e
d
tially each marker was visualized within distinct apical pathways was carded out using stereological tech- fro
m
or basolateral peripheral endosomes. However, after 51 niques in conjunction with horseradish peroxidase and h
min at 37°C, both markers were observed within com- acid phosphatase cytochemistry. This enabled us to es- ttp://ru
mon perinuclear structures. The compartment in timate the absolute volumes and membrane surface pre
which meeting first occurred was shown to be a late areas of the endocytic organelles involved in apical ss.o
rg
endosome (prelysosome) that labeled extensively with and basolateral endocytosis. /jcb
/a
rticle
-p
d
f/1
S E: OEXC.FL plasma membrane components of mammalian structures containing lysosomal proteins; the latter are most 09/6
/3
cells are continuously taken into the cell by endocytosis likely functional lysosomes (Brown et al., 1986; Geuze et al., 25
9
(Steinman et al., 1983; Goldstein et al., 1985). After 1988; Griffiths et al., 1988). /10
5
5 min at 37°C, internalized markers enter tubulovesicular In the polarized epithelial cell, where the surface mem- 89
5
elements, termed early or peripheral endosomes, before be- brane is divided into two domains of distinct composition, 3/3
2
ing routed to lysosomes or recycled back to the plasma mem- continual endocytosis and recycling also occur (Simons and 59
.p
brane (Storrie, 1988; Courtoy, 1989; Gruenberg and Howell, Fuller, 1985). Moreover, there is extensive membrane traffic df b
1990). Ligands destined for lysosomes pass from early en- between the two domains (transcytosis) (Abrahamson and y g
u
dos1983; omes Hopkins, to juxtanuclear 1986), also late ernedfeorsroedm es to (Healse nius prelysosomes et al., RMoodsetwoavl d, and I981; Simister, Limet 1985; et alv.on, Bons1985; dorffet Hoppe al., et 1985). al., 1985; De- est on 0
(Kornfeld and Mellman, 1989), with kinetics that appear to spite this process, the cell maintains its polarity with high 9 F
e
depend on both the cell type and the ligand used (Hubbard, efficiency (Fuller and Simons, 1986). bru
a
1989). Late endosomes are enriched in both the cation- We have studied endocytosis in polarized monolayers of ry 2
independent mannose-6-phosphate receptor (MPR) t and MDCK cells grown on polycarbonate filters. In the preced- 02
3
lysosomal proteins as compared with early endosomes ing paper (Bomsel et al., 1989), it was shown that markers
(Geuze et al., 1988; Grittiths et al., 1988) and have been pro- internalized from the two surface domains of the MDCK cell
posed to be the target of igloG-snart network (TGN)-derived were initially located in distinct early endocytic structures
vesicles containing newly synthesized lysosomal enzymes on but that meeting of the markers occurred within 51 min after
their way to lysosomes (Kornfeld and Mellman, 1989). Late the onset of internalization. In this study, we have character-
endosomes are more acidic than both early endosomes ized and quantitated the endocytic pathways from the apical
(Schmid et al., 1989; Fuchs et al., 1989) and the TGN and basolateral domains of the MDCK cell by an ultrastruc-
(Griffiths et al., 1988). The low pH may facilitate dissocia- tural approach and have demonstrated that the two pathways
tion of lysosomal enzymes from MPRs, allowing the latter meet in late endosomes (or prelysosomes) enriched in the
to recycle back to the TGN for further rounds of transport, MPR.
while the lysosomal enzymes pass to lysosomes (Goda and
Pfeffer, 1988). In agreement with this model is the finding
Materials and Methods
that at later times endocytic markers reach MPR-negative
.1 Abbreviations used in this paper: HRP, horseradish peroxidase; MPR, Cells
mannose-6-phosphate receptor; Sv, surface density; TGN, trans-Golgi
network. MDCK strain I cells were grown on 0.4-#m-pore polycarbonate filters
© The Rockefeller University Press, 0021-9525189/12/3259/14 $2.00
The Journal of Cell Biology, Volume 109 (No. 6, Pt. 2), Dec. ! 989 3259-3272 3259
(Costar Corp., Cambridge, MA) for 3-4 d (Bomsel et al., 1989). The elec- cycloid with the apical, basal and lateral surface domains and the points over
trical resistance across the bilayer was routinely assayed as described previ- the cell were counted and Sv was calculated as described previously (Grifliths
ously (Fuller et al., 1984) and was always >2,000 f/.cm .2 et al., 1989b).
The volume density of the cytoplasm (the ratio of the cytoplasmic volume
srekraM of sisotycodnE to the cell volume) was estimated using the same negatives and a square dou-
ble lattice grid to relate the number of points over the nucleus to the number
Cells were washed twice with the incubation medium (MEM, containing of points over the cell. The mean cell height across the monolayer was
0.2% BSA, 350 mg/liter Na HCO3, and 01 mM Hepes). They were in- directly measured on the same negatives at three different fixed positions
cubated for various times at 37°C or 20°C, in the incubation buffer With across'the negative.
horseradish peroxidase (HRP, 01 mg/ml), added to the apical medium • Magnification level 2: a primary magnification of 10,500× was used to
(Sigma Chemical Co., St. Louis, MO) and BSA-coated gold particles obtain estimates for the volume density of HRP or acid phosphatase-labeled
o25DO( ,'~20) in the basolateral medium. 13-14 nm gold was prepared by compartments. 25-35 random micrographs were analyzed using a double-
the tannic acid method (Slot and Geuze, 1985) and bound to 200 ~g/ml BSA square lattice grid. Points over HRP or acid phosphatase-labeled structures
at pH 7.0. were related to the points over the cytoplasm of the cell.
Due to the relatively low endocytic rate from the apical surface of the Magnification level 3:20-30 micrographs were taken at a primary
MDCK cell (yon Bonsdorff et al., 1985), colloidal gold-BSA did not give magnification of 33,000× after systematically searching for HRP or acid
a sufficiently high signal when administered apically. For this reason, in phosphatase-labeled structures. The Sv of the labeled structures was esti-
those experiments in which HRP was applied basolaterally, a membrane mated by using square lattice grids of various sizes and relating the number
marker was used to follow endocytosis from the apical surface. An anti- of points over the structures of interest to the intersections the grid lines
MDCK rabbit polyclonal antibody against MDCK strain II cells (a gift from made with the limiting membrane. The Sv of these structures was calculated
J. Gruenberg, European Molecular Biology Laboratory) was bound to the as described by Weibel (1979).
apical cell surface (30 min on ice). After washing three times over 30 min,
the cells were incubated with protein A-gold for 30 min on ice and then ymotorcimoyrC and gnilebalonumml
warmed to 37°C for various times while 01 mg/ml HRP was administered
basolaterally. Alternatively, cationized ferritin (0.1 mg/ml; Sigma Chemical Filter-grown MDCK cells were processed for EM after fixation either as D
o
Co.) was bound to the apical surface (5 min on ice) and then warmed to described above for plastic sections or alternatively in %1 glutamldehyde wn
37°C for various times with HRP in the basolateral medium. in 250 mM Hepes, pH 7.4. After washing, the filters were cut into segments loa
d
To terminate the incubations, cells were either fixed without a washing and incubated in 10% gelatin in PBS for 10 min at 37°C. Two pieces of filters ed
step or washed twice with ice-cold incubation buffer before fixation, as de- were then laid on top of each other and the gelatin was allowed to set at 4°C. fro
m
scribed below. The gelatin was cross-linked with 8% paraformaldehyde in 250 mM Hepes, h
pH 7.4, for 1 h at room temperature. The samples were then mounted on ttp
nortcelE ypocsorciM copper stubs with the filter perpendicular to the plane of sectioning and sec- ://ru
p
tioned on a ultramicrotome (model OMU4; Reichert, Vienna, Austria) with re
Cells on polycarbonate filters were fixed with %1 glutaraldehyde in 1.0 M cryoattachment. Labeling was performed as described previously (Griffiths ss.o
cacodylate buffer, pH 7.4 (cacodylate buffer) for 30 rain. To visualize HRP, et al., 1983a, 1984). The afffinity-purified anti-HRP was a gift from J. rg
the fixed cells were rinsed with cacodylate buffer, incubated in cacodylate Gruenberg (European Molecular Biology Laboratory). The antiserum /jcb
/a
buffe1 r rain acnodn taining then H202 0.1% was diamino-benzidine added to a final conce(nDtABr; ation Sigma of Chemical 0.01% to Co.) initiate for awgaasi nst prepared chicken by K. 215K Rtmisch cation-independent and B. Hoflack mannose-6-ph(Eousrpohpaetane Molecular recBeiopltoogyr rticle-p
threem oved reaction. and Afttheer cells a 2-min were incuwbaatsihoend with in catchoed ylate dark, the buffer DAB (Marsh solution et was al., pLeanbdoreantto ry). MPR This by immunoblottinagn tiserum shows of speciMfiDcCitKy cell only homogenates. for the cation-inde- df/10
9
1986). The cells were postfixed for 60 min in %1 OsO4 containing 1.5% /6/3
potassium ferricyanide, or with 2% aqueous OsO4 and block-stained for 25
9
1 h with %1 uranyl acetate in 50 mM maleate buffer, pH 5.2. Cells were Results /10
dehydraaxl in an alcohol series, and embedded in Epon. Sectioning was per- 58
9
formed perpendicular to the plane of the filter to obtain vertical sections as 5
3
required for the stereological analysis. The cells showed no endogenous per- gniteeM of the citycodnE syawhtaP from the /32
5
oxidase activity. owT ecafruS sniamoD of MDCK slleC 9
.p
Acid phosphatase was visualized by the lead capture method as described d
previously (Grifliths et al., 1983b) with a 75-min incubation with the reac- HRP was administered independently to the two domains of f by g
tion mixture. When the same sample was to be stained for acid phosphatase MDCK I cells grown on polycarbonate filters. At early times ue
and HRP, the staining for acid phosphatase was performed before the stain- of internalization (5-10 min at 37°C), apically administered st o
ing with DAB. HRP was visualized close to the apical surface of the cell, n 09
F
cirtemohproM sisylanA whereas basolateral HRP was observed in profiles along the eb
ru
lateral and basal surfaces. A similar distribution of HRP was a
Mean Cell Volume. The absolute volume of the MDCK I cell was estimated shown by yon Bonsdorff et al. (1985) and is in good agree- ry 20
by measuring the mean area of filter covered by one cell (124 pm )2 and the ment with the confocal microscopic observations of Bomsel 23
mean cell height (9.38 ~tm) across the monolayer (Griffiths et al., 1984,
et al. (1989). These peripheral structures correspond to the
1989b). The former was estimated by counting the number of cells per area
of filter on light micrographs after staining the nuclei with Hoechst dye, and early endosomes described in the companion paper.
the latter was measured as outlined below. By EM the apical and basolateral early endosomes ap-
Stereology. For all the sterological analysis Epon blocks were sectioned peared morphologically similar (Fig. ,1 a, b, and d). Both
perpendicular to the plane of the filter to obtain the necessary vertical sec-
had tubular and vesicular portions as observed in other cells
tions (Baddeley et al., 1986). Photographs were taken at three different
magnification levels as described below and all the measurements were (Geuze et al., 1983; Marsh et al., 1986). HRP reaction prod-
made on negatives enlarged 4.1 times on an EMBL-designed projector uct often appeared to be limited to the periphery of such
system. structures and to surround an electron-lucent lumen (Fig. .)1
The three magnification levels used were as follows: Level I: a primary
Such images have been observed by other workers and as-
magnification of 2,800× was used to gain estimates for several morphologi-
sumed to represent spherical vesicles (Steinman et al., 1976;
cal features of the MDCK I cell. Sections were cut from a total of six Epon
blocks after embedding two filters of MDCK I cells. 35 random micro- de Chastellier et al., 1987). However, in some of our images
graphs were taken along the monolayer by using the translation controls of an inner membrane is visible (Fig. ,1 a and b), suggesting
the microscope. that a tubule, or more likely, cisterna, containing HRP may
The surface to volume ratio (Sv) of the MDCK I cell was estimated using
surround an electron-lucent area. In addition to these struc-
the method of Baddeley et al. (1986). A cycloid lattice grid (Cruz-Orive and
tures, a small number of noncoated vesicles directly underly-
Hunziker, 1986) was laid over the projected negatives. Intersections of the
The Journal of Cell Biology, Volume 109, 1989 3260
tions of the cell and gold bound to the apical membrane via
an anti-MDCK antiserum provided the highest signal of all
the markers used (results not shown). After 5-10 min inter-
nalization at 37°C, there was very little evidence of colocali-
zation of the two endocytic markers within the same organ-
elle. HRP and gold particles were predominantly located in
peripheral structures lying close to the basolateral and apical
surfaces, respectively (Fig. 3). However, after 51 min of in-
ternalization, colocalization of HRP and gold particles was
apparent within structures predominantly in the supranuclear
portion of the cell close to the Golgi stacks (Fig. 4). Identical
results were obtained with cationized ferritin as a marker of
the apical surface and HRP as a basolateral marker (see Fig.
5 d). These results are therefore in good agreement with the
biochemical and confocal microscopic results presented in
the companion paper (Bomsel et al., 1989). The organelles
containing the two markers had numerous vesicular inclu-
sions and resembled the multivesicular bodies previously de-
scribed in these cells (von Bonsdorff et al., 1985) and in
other cell types (McKanna et al., 1979; Wall et al., 1980;
D
Hopkins and Trowbridge, 1983). ow
n
Further experiments were designed to test whether any loa
d
meeting could be detected within early endosomes. The pas- ed
sage of ligands from endosomes to lysosomes is blocked at from
20°C (Dunn et al., 1980; Marsh et al., 1983) and Grittiths http
et al. (1988) showed that this block occurs between periph- ://ru
p
eral early endosomes and an intermediatep relysosomal com- re
partment (late endosome). We therefore investigated whether ss.o
rg
any meeting of apical and basolateral markers occurred at /jcb
this temperature. No colocalization of the markers could be /a
erugiF .1 Early apical and basolateral endocytic structures in the demonstrated despite the presence of gold labeled structures rticle
cell. Examples MDCK of with labeled structures HRP after 5-min a -p
apical (a, b, and c) or basolateral (d) incubation at a 37°C. and b and HRP-labeled structures in close proximity (results not df/1
structures the located in show apical region of the cell. HRP reac- shown). These results confirm that MDCK cellsp ossess dis- 09/6
tion product appears to be limited to a membrane-enclosed tubule tinct apical and basolateral early endosomes. Meeting occurs /32
5
or, more cisterna likely, whose inner membrane )sdaehworra( en- in a later compartment, after a lag >5-10 min at 37°C and 9/1
closes an electron-lucent lumen, a c typical shows small uncoated subsequent to the 20°C block in the pathway. 05
8
structure lying close to the apical surface of the cell d (A). swohs 95
3
a basolateral early endosomal structure. Again an electron-lucent Characterization of the Meeting Compartment /32
5
area is surrounded by HRP reaction product. Bars, 1.0 /zm. To characterize the compartment in which the basally ap- 9.pd
f b
plied and apically applied ligands colocalized, ultrathin fro- y g
u
zen sections were prepared. The thawed sections were e
double-labeled with antibodies to HRP and to the cation- st on
ing the apical surface were labeled after short incubations independent MPR. 09 F
with HRP in the apical medium (Fig. 1 c; see below). Cryosections were prepared after a 15-min internalization eb
faces With appeared longer incubatiwointsh in perinuclear at 37°C, structures. HRP from To the determine two sur- aobofve ), apically a 15-min bound internalization gold and basally of apically administered bound catioHnRizPed (see ruary 20
2
whether the pathways shared late endocytic compartments, ferritin and basal HRP, or after a continuous incubation for 3
two distinguishable markers were applied independently to 2 h at 37°C with BSA-gold in the basal medium and HRP in
the surfaces of the MDCK cell. BSA-gold was internalized the apical medium. After 51 min at 37°C, gold internalized
from the basolateral surfacaen d HRP from the apical surface from the apical surface could be visualized within peripheral
of the cell. After continuous internalization of these markers HRP-negative structures showing negligible labeling with
for a period of 51 min or longer at 37°C, colocalization of anti-MPR antibodies (Fig. 5 a). Basolaterally internalized
HRP and colloidal gold was observed within profiles close HRP was visualized within structures in close proximity to
to the nucleus (Fig. 2). No colocalization was observed in the lateral and basal surfaces (Fig. 5 b). Labeling for MPR
peripherally located structures lying close to the apical and was very low but consistently observed in these basolateral
basolateral surfaces of the cell. structures. In contrast, structures containing both the inter-
To establish the precise kinetics of the meeting of the two nalized gold and showing labeling for HRP were more heavily
pathways and define the compartment where meeting first oc- labeled with anti-MPR antibodies (Fig. 5 c). Similar results
curred, HRP was internalized basally and membrane-bound were obtained with the two other sets of markers. Cationized
gold was internalized apically. The use of HRP as a basolat- ferritin bound to the apical surface of the cell colocalized
eral marker prevented any possible problems with access of with basolateral HRP after 51 min at 37°C in MPR-enriched
a particulate marker along the lateral spaces to apical por- structures (Fig. 5 d). After a 2-h continuous incubation with
Parton et al. Endocytic Pathways in MDCK Cells 3261
D
o
w
n
lo
a
d
e
d
fro
m
h
ttp
://ru
p
re
ss.o
rg
/jcb
/a
rticle
-p
d
f/1
0
9
/6
/3
2
5
9
/1
0
5
8
9
5
3
/3
2
5
9
.p
d
f b
y g
u
e
st o
n
0
9
F
e
b
ru
a
ry 2
0
2
3
erugiF .2 Meeting of apical and basolateral fluid-phase markers in the MDCK ceil. MDCK cells were incubated with HRP apically and
BSA-gold basolaterally for 2 h at 37°C. HRP is present in small structures under the apical surface and in larger structures deeper in the
cell .)sdaehworra( HRP and gold particles colocalize in structures (asterisks) close to the Golgi stacks (G). Higher magnification views
of these two structures showing the internalized gold particles )sworra( are shown in the insets. The multivesicular nature of the structures
where meeting occurs is apparent. A, apical surface. L, lateral space. Bars, main figure, 0.5/zm; insets, 0.2/~m.
BSA-gold in the basolateral medium and HRP in the apical varied in structure, often contained intralumenal tubulove-
medium, meeting was again detected within structures that sicular membranes, and were predominantly in a perinuclear
labeled extensively with the anti-MPR antibody (Fig. 5 e) but location in the cell.
not within peripheral structures showing low labeling for These results, using three different marker systems, have
MPR. The MPR-enriched organelles where meeting occurred shown that MDCK cells possess two distinct sets of early en-
'lhe lanruoJ of lleC ,ygoloiB emuloV 109, 9891 2623
1988). Therefore, we investigated whether ligands would
pass from the late endosomes to MPR-negative structures in
this cell type. For this analysis, BSA-gold was used asa n eas-
ily quantifiable and nondegradable marker of bulk flow. Note
that our aim was not to define the kinetics of this step of the
pathway; BSA-gold appears to be slower in its movement to
the end station of the endocytic pathway as compared with
HRP (Bomsel et al., 1989).
BSA-gold was internalized from the basolateral surface of
the cell for 40 min at 3"/°C and then the cells were washed
and reincubated in ligand-free medium for 40 rain, or over-
night at 37°C. The 40-rain incubation in ligand-free medium
was used to chase the gold from early endosomes to late en-
docytic compartments, whereast he overnight incubation was
intended to chase gold to the end station of the endocytic
pathway.
We first determined the percentage of the internalized gold
particles present in MPR-negative and MPR-positive struc-
tures after the two chase times by immunolabeling of frozen
sections. As shown in Table I, after a 40-min chase, % 75 of
D
the total internalized gold was within MPR-positive struc- o
w
tures (Fig. 6). After the overnight chase, however, the per- nlo
a
centage of gold particles within the MPR compartment had de
d
decreased considerably and represented only % 27 of the total fro
m
internalized gold. The remainder of the gold particles was h
present in large aggregates in MPR-negative, electron-dense ttp://ru
structures (Fig. 7), which we presume to be lysosomes. pre
To investigate in more detail the accumulation of gold ss.o
rwietahcihne s the these lysosomes organelles and via to the determine MPR-enriched whether the late marker endo- rg/jcb/a
somes, we counted the absolute number of gold particles in rticle
the various compartments after the two incubation times. -pd
The total number of gold particles within the cell after a f/10
9
40-min uptake followed by a 40-rain chase was approxi- /6/3
2
mately equal to that found in the cell after the overnight 59
/1
chase. This indicates that negligible exit of gold from the 0
5
8
cells occurred during this period due, for example, to trans- 95
3
cytosis or release of lysosomal contents. Using the absolute /32
5
erugiF .3 Distribution of apical and basolateral markers in the number of gold particles per cubed micron of cytoplasm, to- 9.p
d
5 fixed MDCK min cell after after the onset 5 min at 37°C. Epon section of internalization of apical ofmembrane- an MDCK cell tgieotnsh er (above), we with the results could from show the that MPR the labelinngu mber ofo f frogzoelnd parti- sec- f by gu
e
bound gold and basolateral 9 HRP. nm protein A-gold saw bound cles in the MPR-positive compartment decreased by m2,300 st o
4°C. via the antibody at an anti-MDCK surface apical to The cells per cell during the overnight chase, whereas gold in the n 0
9
warmed then were to 37°C for 5 the min in presence of basolateral MPR-negative compartment (lysosomes) increased by 2,600 F
e
particles Gold are apparent HRP. tubular a in structure underlying gold particles per cell during the same time period. These bru
a
surface apical the ,)worra( structures HRP-labeled whereas -worra( results show clearly that most, if not all, of the golpda rticles ry 2
)sdaeh lie close to the lateral space (L). A possible HRP-labeled 0
that reach the lysosomes must pass through the late endo- 23
coated vesicle si indicated by a small arrowhead. Note that some
somes.
basolaterally labeled structures are close located to the apical sur-
The labeled profiles after the overnight chase clearly had
face (A). Bar, 0.2/xm.
a greater mean number of gold particles than the structures
labeled after the 40-min chase (Table I). As only half as
many labeled profiles per cytoplasmic volume were observed
after the overnight chase, this suggested that the gold parti-
dosomes showing low or negligible labeling for MPR and cles were chased from a relatively large compartment to one
that markers from these compartments meet in MPR-en-
of a smaller volume.
riched late endosomes (Kornfeld and Mellman, 1989).
Our qualitative observations indicated that the location of
the late endosomes and the lysosomes differed in that the lat-
Accumulation of lnternalized Ligands within Putative
ter are predominantly in the apical part of the cell whereas
Lysosomes
the former are fountdo be above and around the nucleus. We
In nonpolarized cells ligands have been shown to be trans- therefore counted the percentage of the internalized gold par-
ported from MPR-containing endosome structures to MPR- ticles in the basal region of the cell (arbitrarily defined as be-
negative lysosomes (Griffiths et al., 1988; Geuze et al., low the midpoint of the nucleus). After the 40-min chase,
no~taP te .la citycodnE syawhtaP ni KCDM slleC 3623
D
o
w
n
lo
a
d
e
d
fro
m
h
ttp
://ru
p
re
ss.o
rg
/jcb
/a
rticle
-p
d
f/1
0
9
/6
/3
2
5
9
/1
0
5
8
9
5
3
/3
2
5
9
erugiF .4 Meeting of apical and basolateral markers in the MDCK cell after 51 min at 37°C. Epon section of MDCK cell fixed 51 min .pd
f b
after the onset of internalization of and apical gold basolateral membrane-bound HRP (for details see Fig. 3). HRP product reaction fills y g
a number of vesicles )sdaehworra( the alongside lateral space (L). administered Basolaterally HRP and apically applied gold colocalize ue
in structures )sksiretsa( close to the Golgi area (G) of the cell as shown at higher magnification arid with a lighter printing in the two st o
n
insets sworra( gold indicate particles). A gold particle is also evident in an structure ill-defined (possibly a grazing section of a vesicle) 09
underlying the apical surface .)worra( A, apical surface. Bars, main figure, 5.0 ;mLt insets, 1.0 ttm. Fe
b
ru
a
ry 2
0
2
3
,o31% of the gold particles were present in the basal region. structures after short incubations at 37"C (5-10 min) but af-
After the overnight chase, however, when the labeled struc- ter longer incubations, passed into acid phosphatase-posi-
tures were almost exclusively lysosomal by the criteria de- tive structures (Fig. 7). After the 40-rain chase of BSA-gold,
scribed above, only % 4 of the gold particles were in basaUy "090% of the internalized gold particles were within acid
located structures. Thus, in the MDCK cell the lysosomes phosphatase-positive structures and after the overnight chase
are predominantly (or exclusively) located in the apical re- this had increased to almost 100% (results not shown). Thus,
gion of the cell, whereas the late endosomes are also found both the MPR-enriched late endosomes and the lysosomes
in basal regions. in MDCK cells contain detectable amounts of acid phos-
Acid phosphatase has been shown to be present in the phatase.
lysosomes and the late endosomes of normal rat kidney and
Volume of the Endocytic Compartments in the
baby hamster kidney cells but absent from early endosomes
MDCK Cell
(Gruenberg et al., 1989; Grittiths, G., R. Matteorti, R. Back,
and B. Hoflack, manuscript in preparation). We observed a To gain a quantitative view of the endocytic pathways in the
similard istribution of this enzyme in the MDCK cell. Fluid- MDCK cell, HRP was administered independently to each
phase markers were observed in acid phosphatase-negative of the two surface domains and internalized for various times
lanruoJ ehT of lleC ,ygoloiB emuloV ,901 9891 4623
D
o
w
n
lo
a
d
e
d
fro
m
h
ttp
://ru
p
re
ss.o
rg
/jcb
/a
rticle
-p
d
l~gure .5 Characterization of compartments containing apically and basolaterally internalized markers on cryosections. Cryosections were f/1
0
double-labeled with antibodies to MPR and to HRP. Three different sizes of gold are shown in the figure; the largest gold (13-14 nm) 9/6
corresponds to the basolaterally internalized marker, the medium-sized gold (9 nm) (or in Fig. 4, d, cationized ferritin) corresponds to /32
5
the apically administered marker and the smallest (6 nm) represents labeling for MPR. a, b, and c show cryosections of cells after binding 9/1
protein A-gold to the apical surface and internalization for 51 min at 37°C with HRP in the basolateral medium (see legend to Fig. .)1 05
8
In a, gold (arrows) internalized from the apical surface (A) is evident within a putative early apical endosome, which shows no labeling 95
3
with anti-HRP or anti-MPR, b shows a typical basolateral endosome, in closper oximity to the lateral surface (L) which labels with anti-HRP /32
5
egral( gold) but has no apically internalized gold. This compartment shows low, but significant labeling with anti-MPR (small gold, -worra 9.p
head), c shows an example ofa compartmen! where the basolateral marker (HRP, labeled with anti-HRP and 31 nm protein A-gold, small df b
arrows) colocalizes with the apical marker (BSA-goid, 9 nm; large arrows). The compartment labels significantly With antibodies against y g
u
tahned cation-inidnteeprennadleiznetd for 51 mannroaisne at 3"/0C, 6-phosphate with HRP receptor in the (6 basolateral nm gold, mediumarrowheads). . HRP (d) (labeled Cationized with anti-HfeRrPr itin was and bound large gold, to the apicalla rge )sworrasurface est on
0
colocalizes with cationized ferritin (small arrows indicate an area where a small cluster of the small ferritin particles are apparent) in a 9
F
structure containing MPR (arrowheads). (e) HRP was internalized from the apical surface and BSA-gold 41( nm) from the basolateral sur- eb
face for 2 h at 37"C. HRP (labeled with anti-HRP and 9 nm protein A-gold, small arrows) colocalizes with internalized BSA-gold egral( rua
arrows) in compartments containing significant levels of MPR (small gold, arrowheads). Anti-HRP labeling is low compared with the ry 2
0
labeling from the basolateral surface (compare with Fig. 4, c and d) but significantly above background. Bars, 1.0 /~m. 23
at 37°C. Stereological analysis was then performed to deter- A 10-min incubation resulted in a similar labeling pattern
mine the volume density (or volume fraction) of the labeled with no significant increase in the volume density of the la-
compartments (Table II). beled compartment (0.40 + 0.04%). To determine whether
After 5 min of basolateral uptake of HRP, reaction product any HRP had reached late endosomes at this time and to
was visualized within tubulovesicular organdies located define more accurately the volume of the basolateral early
close to the lateral and basal surfaces of the cells, as de- endosome, we stained cells for acid phosphatase after inter-
scribed above. The volume density of the labeled structures nalization of HRP for 10 min at 37°C. The diffuse reaction
at this time was 0.34 -1- 0.05% of the cytoplasmic volume. 2 product in the HRP-labeled structures was readily distin-
guishable from the electron-dense, acid phosphatase-posi-
2. The electron-lucent parts of these structures were assumed to be HRP tive structures. The HRP-labeled acid phosphatase-negative
negative and only the HRP-reaetive area surroundingt hese regions was in- compartment, which we defined as the early endosome, had
cluded in the analysis (Fig. 1 a and Grittiths et al., 1989a). However, as
a volume density of 0.37 + 0.05 %, not significantly smaller
shown in the legend to Table HI, our basicc onclusions are unchanged if the
than the total HRP-labeled compartment at this time.
electron-lucent areas of the structures are also assumed to be part of the
early endosomal compartment. After a 5-min incubation with HRP in the apical medium,
Parton et al. citycodnE in Cells MDCK Pathways 3265
Table L Distribution of Colloidal Gold Particles within the underlying the apical surface were apparent. These struc-
MDCK Cell after a 40-min Basolateral Incubation and a tures appear to be distinct from apical early endosomes
40-rain or Overnight Chase at 37~C (results not shown). Such structures have been described
previously (von Bonsdorff et al., 1985; Hoppe et al., 1985)
40-Min chase Overnight chase
and may be involved in transcytosis.
Estimated No. of gold particles To ascertain the volume of the later compartments in the
per cell (+ *)MES 4,950 + 1,390 5,270 + 2,040 two endocytic pathways, cells were incubated either apically
Percentage (and estimated no. or basolaterally with HRP for 2 h, washed, and incubated for
per cell) of internalized gold a further 2 h in HRP-free medium. After the chase period,
particles~ in (a) MPR +ve strs 75 (3,710) 27 (1,420)
the typical labeling of the apical and basolateral early endo-
(b) MPR -ve strs 25 (1,240) 73 (3,850)
somes was no longer apparent. The structures labeled from
No. of labeled structures per unit
the two surfaces appeared similar, although labeling of the
cytoplasmic volume (+ SEM)§ 9 + 1 3 + 1
subapical structures described above was only evident after
Mean no. of gold particles in
labeled profile (5: SEM)§ 7 + 2 24:1:7 basolateral addition of HRP. The quantitative data showed
Percentage of gold particles in that the structures labeled from the basolateral side occupied
basal region of cell (5: SEM)§ 31 5:14 4 ± 3 approximately twice the volume of those labeled after apical
administration (1.78 + 0. 23 vs. 0. 93 + 0.13 % of the cytoplas-
* Data obtained from analysis of 27 micrographs of Epon sections, final mag-
nification 41,000×. Sections were assumed to be of a uniform thickness of 60 mic volume).
nm to gain estimates for the number of gold particles internalized per cell (vol- To investigate further the volume of the late compartments
ume 1,163/~m3).
:~ Data obtained by labeling frozen sections of the same experiment with anti- accessible to apical and basolateral ligands, two approaches D
MPR antibodies. Any profiles containing internalized gold were scored as were taken. Firstly, we measured the volume density of the ow
MPR positive (i.e., containing three or more MPR-directed gold particles) or acid phosphatase-positive compartment (i.e., late endo- nlo
MPR negative. 120 structures were examined after the 40-min chase and ',,60 ad
min after the overnight chase (i.e., for each time point the distribution of over somes and lysosomes) in the MDCK cell. Secondly, we ad- ed
1,600 gold particles was assessed) and the experiment was performed three ministered HRP simultaneously to the two surfaces of the fro
m
ttiumreess. varied The percentage by (cid:127)15% of between internalized these egoxlpde riments. particles From within the MPR estimated positive number struc- MDCK cell. We then investigated whether the volume den- http
of gold particles per cell in these compartments (shown in parentheses), it is sity of structures labeled in this way was equal to the sum ://ru
em(vi=ed2ne,tn6t 0 0). (=2,300) that the is decrease similar to in the the increase number in of the gold number particles of particleins the MPR in lysosomes compart- osufrf aces the volume independently. densities of structures labeled from the two press.o
ba§ sal Estimates region froomf the analysis cell was of defined micrographs as the portion of Epon of the sections cell basal as above. to the mid- The evident After within staining mostly for acid circular phosphatase, or oval-profiled reaction structures product twahsa t rg/jcb/a
point of the nucleus. were predominantly positioned around the cell nucleus. In- rticle
-p
ternal membrane structures were sometimes apparent within d
f/1
these profiles, which appeared very similar to those seen in 09
/6
few HRP-labeled structures were evident (a volume density the MPR-enriched late endosomes and lysosomes. The vol- /32
5
ume density of acid phosphatase-reactive structures was 9
of 0.02 + 0.01%). As described above, a number of small un- /1
1.44 + 0.21% of the cytoplasmic volume, not significantly 05
coated vesicles or tubules of a similar size to coated vesicles 8
9
different from the volume of structures labeled by basolateral 5
directly underlying the apical surface were labeled (Fig. 1 HRP after a 2-h incubation followed by a 2-h chase. The total 3/32
c). Serial-section analysis showed that these structures were 59
combined volumes of the late endosomes and lysosomes, .p
not part of the larger tubulovesicular early endosomes d
defined as acid phosphatase-positive, thus appeared to be f b
(results not shown). After 10 min of HRP internalization at similar to the volume of those late compartments accessible y gu
3"/°C, more of the larger tubulovesicular structures were la- e
beled (Fig. I a) and the volume density increased fivefold to to basolateral ligands. When HRP was internalized from st on
both surfaces for 2 h and then chased for a further 2 h, the 0
0.11 + 0.03 %. The volume density of the HRP-labeled acid total volume density of the structures labeled in this way 47.1( 9 Fe
phosphatase-negative compartment at this time was 0.10 + b
0.03 %, showing that little HRP had been delivered to acid the + 0.21%) compartment was, again, labeled not only significantly from the badsioflfaetreernatl from side. that This of ruary 2
phosphatase-positive compartments. 0
2
makes it unlikely that there is a major class of late structures 3
After 51 min of endocytosis at 37°C from the basolateral
accessible only to apically endocytosed ligands, but a sub-
surface, HRP could be detected in juxtanuclear acid phos-
class of structures only labeled by basolaterally administered
phatase-positive structures (results not shown) and the vol-
markers may exist.
ume density increased to 0.72 + 0.08 % of the cytoplasmic
Estimates of the absolute volumes and surface areas of the
volume. Similarly, between 10 and 51 min of uptake from the
endocytic compartments per MDCK cell, as well as some
apical surface, the volume density almost doubled from 11.0
morphometric features of the cell, are given in Table III.
to 0.18 + 0.05% of the cytoplasmic volume.
From these estimates it is apparent that the basolateral early
At later times of internalization the volume density of
endosomes are ,x,3.5 times larger in volume and in surface
structures labeled from both surfaces gradually increased,
area than the apical early endosomes. This is similar to the
reaching 2.28 + 0.21% of the cytoplasmic volume after 2 h
uptake from the basolateral surface, and 1.07 + 0.17% after ratio of the areas of the two surface domains (3.8 + 0.5:1). 3
the same period of uptake from the apical surface. After 2 h
3. This ratio is significantly lower than that estimated by yon Bonsdortt et
of uptake, the number of labeled structures surrounding the
al. (1985; 7.6:1). This difference can be explained by variations in the type
nucleus had visibly increased, and in the case of basolateral
of filter used as well as in the precise growth conditions, both of which can
uptake, a row of small tubular or vesicular structures directly affect the final appearance of the MDCK cells (unpublished observations).
The Journal of Cell Biology, Volume 109, 1989 3266
D
o
w
n
lo
a
d
e
d
fro
m
h
ttp
://ru
p
re
ss.o
rg
/jcb
/a
rticle
-p
d
f/1
0
9
/6
/3
2
5
9
/1
0
5
8
9
5
3
/3
2
5
9
.p
d
f b
y g
u
e
st o
n
0
9
Figure .6 Late endosomes (prelysosomes) of the MDCK cell. a and b show Epon sections of typical structures labeled with internalized Fe
b
BSA-gold after a 40-min incubation in the basolateral medium at 370C and a 40-min chase in marker-free medium. The labeled structures ru
a
are located close to the Golgi (G) and nucleus (N). Numerous internal vesicular profiles are evident. These may not all be free vesicles ry 2
but simply infoldings of the limiting membrane as indicated tbhye arrowhead b. in c, d, and e show cryosections from the same experiment 02
3
as above labeled with anti-MPR and protein A-gold. Internalized basolaterally administered gold (large particles, arrowheads) are located
within structures containing relatively high levels of MPR (small gold, arrows). Internal membranes are evident within all the labeled struc-
tures and appear as circular profiles (e.g., c and d) or as myelin-like arrays of parallel membranes (d and e). In e, the MPR and internalized
gold appear to be segregated to different areas of the same structure. N, nucleus. In c, an asterisk indicates an area of the cell that most
lickoenltya ined glycogen deposits. These are invariably not retained in our routine frozen sections leaving large"empty" areas. Bars, 1.0 #m.
Discussion carry out a detailed characterization of the enaocytic path-
ways and in particular to identify the compartment where
In the preceding paper (Bomsel et al., 1989) a combined bio-
meeting occurs. In addition, we have obtained estimates for
chemical and confocal microscopic approach was used to in-
the volumes and surface areas of the endocytic compart-
vestigate the spatial organization of the endocytic apparatus
ments. We were able to distinguish three distinct endocytic
in polarized monolayers of MDCK cells and to demonstrate
structures in the MDCK cell: peripherally located early en-
meeting of the apical and basolateral pathways of endocyto-
dosomes, perinuclear late endosomes, and lysosomes.
sis. In this paper, we used an ultrastructural approach to
Parton et al. Endocytic Pathways in MDCK Cells 3267
D
o
w
n
lo
a
d
e
d
fro
m
h
ttp
://ru
p
re
ss.o
rg
/jcb
/a
rticle
-p
d
f/1
0
9
/6
/3
2
5
9
/1
0
5
Figure .7 Putative lysosomes of the MDCK cell. a, b, and c show typical structures labeled with BSA-gold after a 40-min incubation in 89
5
the basolateral medium at 370C and an overnight chase in marker-free medium. Large aggregates of gold particles are present within the 3/3
2
electron-dense lumen of the labeled structures (a and b) (compare with Fig. 7, a and b). Parallel arrays of membranes were often observed 5
9
(e.g., b, arrowheads), c shows an unstained section of a parallel experiment in which the cells were reacted for acid phosphatase after .pd
e fixation. show cryosectioDness pite some labeled background with anti-MPR cytoplasmic antibodies. reaction The product, electron-dense the gold-labeled structures structure containing is clearly internalized acid phosphatase-positive, gold (large gold, arrows) d and f by gu
e
do not have significant levels of MPR (small gold, arrowheads). In d, an adjacent profile is evident that labels significantly for MPR but st o
n
is free of the internalized gold. G, Golgi. Bars, 1.0 /~m. 0
9
F
e
Table II. Summary of Volume Density Measurements bru
a
Volume ytisned ry 2
0
2
3
Time medium Apical laretalosaB medium
m/n % of cimsalpotyc emulov ~= MES
HRP 5 0.02 -I- 0.01 0.34 + 0.05
10 .O 11 + 0.03 0.40 + 0.04
10 0.10 + 0.03 0.37 6- 0.06
})-PA(
51 0.18 + 0.05 0.72 + 0.08
30 0.32 + 0.06 1.22 + 0.10
60 0.66 + 0.09 1.68 -I- 0.22
120 1.07 + 0.17 2.28 + 0.21
120 plus 120 esahc 0.93 + 0.13 1.78 + 0.23
1.74 + 0.14"
Acid phosphatase 1.44 + 0.21,
HRP * in apical dna medium. basolateral
* stnemerusaeM were to restricted vesicular stnemtrapmoc > (diameter 50 to the urn) Golgi exclude stacks dna TGN, which variable acid showed esatahpsohp
as activity in other cell types (see Hand dna Oliver, Griffiths 1984; dna Simons, .)6891
structure. acid phosphatase-negative HRP-labeled
The Journal of Cell Volume Biology, 109, 9891 3268