Kinetics of endocytosis in renal proximal tubule studied

with ruthenium red as membrane marker

HENRIK BIRN, ERIK ILSQ) CHRISTENSEN, AND SBREN NIELSEN
Department of Cell Biology, Institute of Anatomy, University of Aarhus, DK-8000 Aarhus C, Denmark

Birn, Henrik, Erik Ilsgl Christensen, and Sgren Niel- as small, membrane-bounded structures within the
sen. Kinetics of endocytosisin renal proximal tubule studied cytoplasm.
with ruthenium red as membranemarker. Am. J. Physiol. 264 Other studies have indicated that invaginations con-
(Renal Fluid Electrolyte Physiol. 33): F239-F250, 1993.-This stitute a considerable part of small membrane-bounded
study was performed to determine the membranearea of the structures in the apical cytoplasm of rat kidney proximal
different compartments involved in endocytosis in rat kidney tubule (6, 32), but so far no study has established the
proximal tubule. This enablesa direct estimation of the kinetics morphological and quantitative distribution of mem-
of the membraneturnover in the subprocesses of endocytosis
and recycling of membraneconstituents. To accomplishthis, brane confined to invaginations, small endocytic vesi-
cross-sectionedinvaginations have to be distinguishedfrom en- cles, and large endocytic vacuoles. Such data are vital to
docytic vesicles.To obtain a continuous heavy staining of the establish the dynamics of endocytosis and internal-
apical plasmamembrane,we have developeda sandwich-stain- ization of membranes constituents. Several membrane
ing technique involving initial staining with ruthenium red, proteins have been located by immunocytochemistry
osmium, and thiocarbohydrazide functioning as a molecular or immunohistochemistry to vesicles in the apical cyto-
bridge between osmium molecules.The distribution of invagi- plasm, e.g., H+-adenosinetriphosphatase (H+-ATPase)
nations, noncoated and coated endocytic vesicles,aswell asthe (3), Na+ cotransporter (13), glycoprotein gp330 (18),
surfacedensity of invaginations, small endocytic vesicles,large and folate-binding protein (15). The interpretation of
endocytic vacuoles,lysosomes,and denseapical tubules are de- such apparently vesicular localization of membrane pro-
termined in segment1 (Sl) and 2 (S2) of the proximal tubule. tein with respect to distribution between the plasma
This has shown that invaginations constitute most [54% (SI) membrane and endocytosed membrane requires knowl-
and 62% (S2)] of the membranein small membrane-bounded
edge of the quantitative relationship between invagina-
structures in the apical cytoplasm. No morphologicalcharacter-
istics enableddirect differentiation between cross-sectionedin- tions and endocytic vesicles. Furthermore, it has been
vaginations and smallendocytic vesicles,but a method for cor- suggested that endocytosis and recycling are fast pro-
rect identification of 72% of all invaginations and 82% of all cesses occurring within minutes (2, 4). However, no
smallendocytic vesiclesis presented.Using the surfacedensities study has described the kinetics of the intracellular
of various compartments together with previous data on the transport of endocytosed membranes to thereby give the
internalization of membrane markers, we developed a kinetic possibility of calculating the turnover of membrane in
model enabling calculation of the velocity of membraneinter- each compartment.
nalization and subsequentrecycling. The velocity of internal- To determine the distribution of endocytic invagina-
ization is 6.4 x 10d3 pm2 prne3 s-l, correspondingto internal-
l l

tions and endocytic vesicles in the apical cytoplasm of

ization of a membranearea equivalent to the entire surface of rat kidney proximal tubule cells and to calculate the
invaginations within 78 s. The velocity of membranetransfer
surface density of each structure, we have developed a
from the vacuolar compartment to denseapical tubules and of
recycling from denseapical tubules to the luminal membraneis method of intense ruthenium red membrane staining.
6.2 x 10m3pm2.pm-3 l s- l. The entire membranearea equiva- This method involves sandwich staining by ruthenium
lent to the surface of the vacuolar compartment and denseapi- red and osmium using thiocarbohydrazide as a molecu-
cal tubules is transferred in 43 and 92 s, respectively. The lar bridge. By staining luminal membranes including
amount of membranethat surroundslysosomesis transported endocytic invaginations connected to the surface, we are
from the vacuolar compartment within 23 min. able to determine the surface density of invaginations,
electron microscopy; histochemistry; kidney; recycling; mem- endocytic vesicles, large endocytic vacuoles, dense apical
brane transport tubules, and lysosomes. Using these parameters, we have
developed a kinetic model for the transfer of membranes
from endocytic invaginations through an endocytic
THE ENDOCYTIC PATHWAY in proximal tubule cells in- intracellular compartment to lysosomes and to dense
volves transfer of membrane from the plasma membrane apical tubules for recycling. Furthermore, by observing
through endocytic invaginations into endocytic vesicles micrographs of proximal tubules stained by ruthenium
and large endocytic vacuoles followed by transport into red, we have developed a procedure for estimating the
lysosomes. A substantial part of the endocytosed mem- distribution of endocytic invaginations and small
brane is recycled from the vacuolar compartment to the endocytic vesicles in nonstained rat kidney proximal
plasma membrane through dense apical tubules (6, 24). tubules.
In the kidney proximal tubule both invaginations and
endocytic vesicles are present in large numbers in the MATERIALSAND METHODS
apical cytoplasm (21). Some invaginations are directly Animals and fixation. Nine male Wistar rats (weighing 225
seen to be connected to the luminal membrane. How- 250 g) allowed free accessto food and water were anesthetized
ever, the ultrastructural appearances of endocytic in- with pentobarbital sodium(60 mg/kg body wt), and the kidneys
vaginations with no apparent connection to the cell sur- were fixed by retrograde perfusion through the abdominalaorta
face and endocytic vesicles are very similar. Both appear using 1% glutaraldehyde in 0.1 M sodium cacodylate buffer,
0363-6127/93 $2.00 Copyright 0 1993 the American Physiological Society F239

Downloaded from www.physiology.org/journal/ajprenal by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 11, 2019.

Downloaded from www.physiology.org/journal/ajprenal by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 11, 2019.
 KINETICS OF ENDOCYTOSIS STUDIED WITH RUTHENIUM RED F241

Table 1. Ruthenium red staining procedure A coherent lattice squaretest system was usedfor intersec-
tion counting. The distance between lines for intersection
1) Ruthenium red + 1% glutaraldehyde in cacodylate buffer counting was 10 mm for counting invaginations and small en-
for 1 h docytic vesicles (diam smaller than 0.5 pm) and 20 mm for
2) Ruthenium red + 1% 0~0~ in 0.1 M cacodylate buffer for counting large endocytic vacuoles (diam greater than 0.5 pm)
2h and lysosomes.The test area included the area of the corre-
3) Ruthenium red in 0.1 M cacodylate buffer for 3 periods of spondingproximal tubule cellsas determinedby point counting
5 min using a lattice squarewith the distanceof 30 mm betweenlines.
4) Ruthenium red + 1% thiocarbohydrazide in 0.1 M
cacodylate buffer for 1 h The surface density (Sv) of all invaginations as well as only
5) Ruthenium red + 1% 0~0, in 0.1 M cacodylate buffer for membrane-boundedinvaginations with no apparent connection
lh to the surface,smallendocytic vesicles,largeendocytic vacuoles,
6) Ruthenium red in 0.05 M maleate buffer for 3 periods of and lysosomeswas calculated as Sv = 21/L,, in which I is the
5 min, pH 5.2 numberof intersectionsand LT is the length of the test line (34).
8) Ruthenium red + 0.5% uranyl acetate in 0.05 M maleate To reduce the Holmes effect (effect of section thickness) the
buffer for 1 h, pH 5.2 surface densitiesof invaginations and small endocytic vesicles
9) Ruthenium red in 0.05 M maleate buffer for 3 periods of were corrected by multiplication with the correction factor
5 min, pH 5.2 Jwv) = 7&r + 4g), in which g is the ratio between section
Small blocks of perfusion-fixed renal cortical tissue were immersed in thickness and the diameter D of invaginations and small en-
various solutions according to the scheme described above. docytic vesicles, respectively (34). Average section thickness
was 0.05 pm, asjudged by interference color during sectioning,
pH 7.2. The kidneys were removed, cut into small blocks of and D was calculated as describedbelow.
cortical tissue,and immersionfixed in the samefixative for 3.5 The surfacedensity of denseapical tubules wascalculatedby
h. Thin slices(loo-150 pm) were cut by Vibratome sectioning counting the number of profiles and using the formula Sv =
and kept overnight in the 1% glutaraldehyde sodiumcacodylate 2 *NA7rd,in which NA is the number of profiles per test area and
solution. d is the diameter of denseapical tubules (12). The number of
Ruthenium red staining. Ruthenium red (Sigma Chemical) denseapical tubules wascounted in a lattice squaresystemwith
was purified by separation from ruthenium violet and ruthe- a distance between lines of 20 mm, counting profiles in every
nium brown according to Luft (20). seventh square.The averagediameter of 129 denseapical tu-
The ruthenium red membranestaining wasperformed by the bules selectedat random is 0.074 pm. There was no difference
immersion of renal cortical tissue into ruthenium red, 0~0~ between the diameter of dense apical tubules in Sl and S2
(Johnson Matthey), and thiocarbohydrazide (Sigma Chemical) proximal tubules. The surfacedensity was converted to surface
as described in Table 1. To enhance staining, the procedure area (pm2) per millimeter tubule length by multiplying the
involves sandwich binding of ruthenium red to the cell mem- surface density with the test area and lo3 pm. The average
branes. Initial binding of osmium and ruthenium red to cell cross-sectionedareawas1,230t 250pm2for Sl proximal tubule
membranesis followed by the immersion of the tissue into (n = 9 tubulesfrom 7 different animals)and 1,180t 130pm2for
thiocarbohydrazide and a secondimmersion into osmium and S2 proximal tubule (n = 6 tubules from 5 different animals).
ruthenium red. Assumingthat endocytic vacuolesare sphericalstructures,we
The sandwich-stainingprocedure was evaluated by compar- calculated the averagediameter D of small endocytic vesicles
ing the resultswith preliminary staining experimentsperformed and largeendocytic vacuolesasD = D’~/T (34). D’ is the average
without thiocarbohydrazide incubation. diameter of all observedprofiles of the relevant structure and is
The ruthenium red staining procedurewas followed by dehy- calculated on the basisof boundary length B(a) as D’ = B(a)/
dration in graded ethanols and embeddingin Epon Polarbed
812 (Bio-Rad) in flat molds.Ultrathin sections(50 nm) were cut 0T.d = uw4N4 )9in which I is the number of intersections
and h is the distancebetweenlines in the squarelattice usedfor
on a LKB Ultratome III and studied in a JEOL 100 B, JEOL intersection counting. The average diameter of membrane-
100 CX, or ZeissEM 902A electron microscope. boundedinvaginations with no apparent connection to the sur-
Morphometric analysis. Electron micrographs covering one- face was calculated by the sameprinciple.
fourth to one-third of a cross-sectioned proximal tubule The data are meansrt SD of either nine (SI) or five animals
were photographed at a primary magnification of ~15,000 or (S2). Statistical differences were tested using Student’s t test.
~16,000. The micrographswere obtained from heavily stained
tubules at the periphery of the tissue blocks including both RESULTS
segment1 (SI) proximal tubule (9 animals)and segment2 (S2)
proximal tubule (5 animals) and enlargedto a final magnifica- Ultrastructural evaluation. When ruthenium red sand-
tion in the range of X45,800-51,600, determined by using a wich staining using thiocarbohydrazide was compared
carbon replica (2,160 lines/mm). with the results of preliminary experiments without
The number of profiles were counted for the following struc- thiocarbohydrazide, we found an increase in staining
tures: 1) invaginations defined as ruthenium red-stainedmem- intensity that enabled us to clearly distinguish between
brane-boundedstructures with no apparent connection to the membrane-stained invaginations and nonstained small
luminal surface, 2) noncoated small endocytic vesiclesdefined
endocytic vesicles (Fig. 1).
as nonstained, noncoatedmembrane-boundedstructures in the
apical cytoplasm with a diameter smaller than 0.5 pm, and 3) The extent to which ruthenium red has gained access
coated endocytic vesiclesdefined as nonstained, coated mem- to the tubular lumen depends very much on the distance
brane-boundedstructures. The relative distribution of each of from the surface of the tissue block to the observed tu-
the above compartments was then calculated. bule. The more peripheral that the tubule was located, the

Fig. 1. Apical part of proximal tubule cells after staining with ruthenium red using conventional staining with ruthenium
red and 0~0, (A) or the sandwich-staining technique involving thiocarbohydrazide (B). Insets: low-magnification
micrographs showing location of tubules (arrowhead) in tissue blocks (x240). Both tubules are from the surface of the
tissue block, and both blocks were sectioned in such a way that tubular lumina were open to the surface during staining.
A more intense, thicker staining of brush border and invaginations is observed when using sandwich-staining technique.
Magnification, ~2,700.
Downloaded from www.physiology.org/journal/ajprenal by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 11, 2019.
F242 KINETICS OF ENDOCYTOSIS STUDIED WITH RUTHENIUM RED

better membrane marking by ruthenium red, and the ob- ruthenium red-stained, cross-sectioned Sl proximal tu-
served tubules were chosen from the outer part of the bule cell. The membranes of brush border and invagina-
tissue blocks. This is due to the fact that only tubules tions are heavily stained by ruthenium red. Wherever
open to the surface of the tissue block are stained. both leaflets of the apical cell membrane were visible,
Figure 2, A and B, shows the typical appearance of a staining is located to the outer leaflet only; staining was

Fig. 2. Cross-sectioned proximal tubule cells of Sl segment sandwich stained by ruthenium red as described in Table 1.
Membranes of brush border (BB) and invaginations (I) are clearly stained by ruthenium red, whereas no staining is
observed in noncoated endocytic vesicles (V) or coated endocytic vesicles (CV, B). No staining of dense apical tubules
is observed (small arrows in B). Some invaginations tended to have a more angular appearance than vesicles (notice
especially B). Ruthenium red staining has not penetrated the tight junctions into the intercellular space (large arrow in
B). LYS, lysosome.

Downloaded from www.physiology.org/journal/ajprenal by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 11, 2019.

 KINETICS OF ENDOCYTOSIS STUDIED WITH RUTHENIUM RED F243
never observed within the cells. Also, no staining of dense A and B, shows a ruthenium red-stained, cross-sectioned
apical tubules was observed. There were no morphological S2 proximal tubule cell. The distribution of invaginations
characteristics enabling direct differentiation between in- and small endocytic vesicles appears similar to the dis-
vaginations and vesicles, although some invaginations tribution in Sl, although the total number of small mem-
tended to be more angular than vesicles (Fig. 2B). Invag- brane-bounded structures appears to be smaller than in
inations are confined to the most apical cytoplasm, and Sl.
the smallest diameter never exceeded 0.5 pm. Figure 3, The amount of both invaginations and small endocytic

Fig. 3. Cross-sectioned proximal tubule cells of S2 segment sandwich stained by ruthenium red as described in Table 1.
Membranes of brush border (BB) and invaginations (I) are clearly stained by ruthenium red, whereas no staining is
observed in noncoated endocytic vesicles.(V) or large endocytic vacuoles (LEV, B). No staining of dense apical tubules
is observed (small arrows in B). Some invaginations tended to have a more angular appearance than vesicles (notice
especially A). Notice lighter appearance of lysosomes (LYS) and longer brush-border microvilli in S1 (Fig. 2) compared
with S2 proximal tubule cells shown here. Magnification, X2,300 (A) and ~43,000 (B).

Downloaded from www.physiology.org/journal/ajprenal by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 11, 2019.

F244 KINETICS OF ENDOCYTOSIS STUDIED WITH RUTHENIUM RED

vesicles in a given section very much depends on section coated endocytic vesicles. There is no significant differ-
angle relative to the tubular axis. If tubules were not ence between the distribution in Sl and S2 segments,
cross-sectioned, then the amount of small membrane- although the relative number of invaginations tends to be
bounded structures was significantly increased, as illus- higher in Sl than S2. Invaginations with no apparent
trated in Fig. 4, which shows an obliquely sectioned prox- connection to the lumen constitute between 40 and 50%
imal tubule. of the total number of small membrane-bounded struc-
Quantitative analyses. Quantitative analyses were per- tures in the apical cytoplasm.
formed on micrographs from cross-sectioned tubules as The average diameters of invaginations, small en-
assured by a longitudinally sectioned brush border. Table docytic vesicles, and large endocytic vacuoles are shown
2 shows the total number and relative distribution of in Table 3. Invaginations are significantly wider than
invaginations, noncoated small endocytic vesicles and small endocytic vesicles, whereas no difference is found

Fig. 4. Obliquely sectioned proximal tubule cells stained by ruthenium red. Number of small membrane-bounded
structures in apical cytoplasm is very large, and most of are invaginations as indicated by ruthenium red staining.
Invaginations are present up to 3.7 am down into the apparent cytoplasm of the cell. Magnification, ~30,000.

Downloaded from www.physiology.org/journal/ajprenal by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 11, 2019.

 KINETICS OF ENDOCYTOSIS STUDIED WITH RUTHENIUM RED F245

Table 2. Total number and relative distribution Surface density (p&m3)

of invaginations, noncoated small endocytic vesicles, 0.8
and coated endocytic vesicles in Sl and S2 proximal
tubule segments

Noncoated
Coated
Endocytic Small
Endocytic
Invaginations Endocytic 0.6
Vesicles
Vesicles

Total no. of profiles

Sl segments
766 689 88
0.5 - 1
Relative no. of profiles, 49t5 45t5 5t2
% of total 0.4
S2 segments
Total no. of profiles 339 422 77 0.3
Relative no. of profiles, 42t8 49t7 9t2
% of total
Data are means t SD of 9 (Sl) or 5 (S2) animals. 0.2

Table 3. Average diameter of invaginations, 0.1 -

small endocytic vesicles, and large endocytic vacuoles

Small Large
0.0
Endocytic MB-INV INV SEV LEV DAT LYS
2P Endocytic Endocytic
Invaginations
Vesicles Vacuoles Fig. 5. Surface density of membrane-bounded, nonconnected invagina-
tions (MB-INV), all invaginations (INV), small endocytic vesicles
Sl 0.33t0.03 <O.OOl 0.21t0.03 1.2kO.2 (SEV), large endocytic vacuoles (LEV), dense apical tubules (DAT), and
s2 0.30t0.05 co.01 0.20t0.02 l.lt0.2 lysosomes (LYS) in Sl (white bars) and S2 (black bars). Error bars show
All data are average diameters in pm and are means t SD of 9 (Sl) SD.
or 5 (S2) animals. Average diameter of invaginations is significantly
different from average diameter of small endocytic vesicles in both Sl
and S2. Table 4. Surface area of invaginations, small endocytic
vesicles, large endocytic vacuoles, dense apical tubules
for the same structures between Sl and S2 segments. and lysosomesin Sl and S2 proximal tubule segments
The surface densities of invaginations with no appar-
ent connection to the surface, all invaginations, small Small Large
Dense
Endocytic Endocytic Endocytic
endocytic vesicles, large endocytic vacuoles, dense apical Invaginations Vesicles Vacuoles
Apical Lysosomes
Tubules
tubules, and lysosomes are illustrated in Fig. 5. There is (CO.5 pm) (>0.5 pm)
no significant difference between Sl and S2 segments in Sl 616t116 227t68 114k51 701+261 319k89
the amount of membrane of each of the described struc- s2 447k117 223~31 89zk42 720t136 31Ok52
tures, although the total amount of membrane of invag- Data are in lo3 pm2/mm tubule length and are means & SD of 9 (Sl)
inations, small endocytic vesicles, and large endocytic or 5 (S2) animals. Surface area was calculated as the surface density of
vacuoles tends to be higher in Sl than S2. The surface the relevant structure multiplied by lo3 pm and the average cross sec-
density of invaginations with no apparent connection to tioned tubule area (1,230 pm2 for Sl and 1,180 pm2 for S2, see also
MATERIALS AND METHODS).
the surface accounts for 62 and 54% of the total surface
density of all small membrane-bounded (diameter smaller
than 0.5 pm) structures in SI and S2 segments, respec- average distance of 0.63 pm (3 animals). On micrographs
tively. from three different animals a line was then drawn
When calculating the surface area per millimeter tubule through the apical cytoplasm longitudinally to and 0.63
length of invaginations, small endocytic vesicles, large pm from the apical cell surface (Fig. 6). When determin-
endocytic vacuoles, dense apical tubules, and lysosomes ing the distribution of invaginations with no apparent
(Table 4), there was no significant difference between Sl connection to the surface and small endocytic vesicles on
and S2 in the surface area of invaginations, small en- each side of the line drawn, it was found that apical to
docytic vesicles, large endocytic vacuoles, dense apical this line 82% of the surface density of all small mem-
tubules, and lysosomes. brane-bounded structures consisted of invaginations and
To estimate the distribution of invaginations and small 18% consisted of small endocytic vesicles, whereas basal
endocytic vesicles within the apical cytoplasm, we di- to the line 75% of the surface density of all small mem-
vided the apical cytoplasm of Sl proximal tubule cells brane-bounded structures consisted of small endocytic
into a basal and an apical part. This was performed by vesicles and 25% consisted of invaginations. To evaluate
randomly measuring the distance between the apical cell the predictive value of the procedure, all small mem-
surface and the basal part of the deepest invagination for brane-bounded structures with no apparent connection to
every 2.1 pm (equal to 10 cm on the enlarged micro- the surface, apical to the line, were defined as invagina-
graphs) along the apical membrane. This resulted in an tions, and all similar structures basal to the line were

Downloaded from www.physiology.org/journal/ajprenal by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 11, 2019.

F246 KINETICS OF ENDOCYTOSIS STUDIED WITH RUTHENIUM RED

the polysaccharides and reduction of 0~0~. As second-

ary ruthenium red is added during sandwich staining,
reduction/oxidation reactions between secondary 0~0~
and ruthenium red enhances staining intensity.
The above staining method has enabled us to clearly
distinguish between invaginations with no apparent con-
nection to the tubular lumen and small endocytic vesicles
in the apical cytoplasm. Such invaginations normally in-
distinguishable from small endocytic vesicles account for
49 (Sl) and 42% (S2) of all membrane-bounded profiles
in the apical cytoplasm with a diameter smaller than
0.5 pm. The surface density of these invaginations repre-
sents 62 and 54% of the surface density of all small mem-
brane-bounded structures in Sl and S2 segments, respec-
tively. Small endocytic vesicles represent 66 and 71% of
the surface density of all vacuoles (small and large) in Sl
and S2, respectively. This shows that most of all the small
membrane-bounded structures and most of the corre-
sponding membrane in the apical cytoplasm are in fact
cross-sectioned invaginations with no apparent connec-
tion to the tubular lumen.
Dense apical tubules are almost never observed to be
connected to the luminal cell surface in perfusion-fixed
tubular cells unless tubular hydrostatic pressure is re-
duced as shown by Cui et al. (7) by microinfusion of oil
into the tubule. This is confirmed by the fact that no
ruthenium red-stained dense apical tubules were observed
at the luminal surface, suggesting that fusion of dense
apical tubules is followed by instant dilatation and inte-
gration into the luminal membrane.
When studying endocytosis and internalization of
Fig. 6. Cross-sectioned Sl proximal tubule cells. A line is drawn through
apical cytoplasm longitudinally to and 0.63 Km from apical cell surface. membrane, receptors, or ligands by morphological meth-
This distance was determined as the average distance from the apical ods, it is of crucial importance to consider that local-
cell surface to most basal part of most basal invaginations on 3 different ization of markers in small membrane-bounded struc-
animals. Distribution of invaginations and small endocytic vesicles on tures with no apparent connection to the lumen is no
each side of line was calculated. Structures located on the line were
allocated to the apical or basal side depending where most of apparent evidence of internalization. This raises the question of
membrane of relevant structure was located. Magnification, ~36,000. how to distinguish between invaginations and small en-
docytic vesicles in conventional sections without mem-
defined as small endocytic vesicles. As a result 82% of the brane staining. Although invaginations were significantly
total surface density of small endocytic vesicles was wider, often more angular (Fig. 2B), and generally located
counted as such, and 72% of the total surface density of more apically than small endocytic vesicles, they were
apparently “nonconnected” invaginations was counted as often mixed, and no morphological characteristics en-
such. abled us to distinguish between the two structures.
Rather than simply counting all small membrane-
DISCUSSION bounded structures as vesicles, we have developed a
This study presents a sandwich technique for ruthe- method for better estimation of the distribution of
nium red staining of cell membranes. The use of this the two structures. Testing the predictive value of this
technique resulted in an increase in the staining intensity procedure, we are able to correctly identify 72% of non-
of cell membranes within the rat kidney nephron. The connected invaginations and 82% of small endocytic
technique is based on the use of thiocarbohydrazide, an vesicles. The procedure can be used for studies on cross-
osmiophilic agent capable of binding osmium at both sectioned tubules where better differentiation between
ends of the molecule (30). We propose that the sandwich- invaginations and vesicles is desirable (2). Also, observ-
staining reaction involves binding of one end of the thio- ing the distribution of invaginations and small en-
carbohydrazide molecule to the osmium initially bound to docytic vesicles could be a valuable procedure in the
the tissue. This is followed by binding of additional os- study of internalization. Conclusions concerning inter-
mium to the other, free end of the thiocarbohydrazide nalization are more reliable if intensive labeling of
molecule. Luft (20) suggested that the staining of mem- membrane, receptors, or ligands is observed in small
branes with ruthenium red and 0~0~ includes initial membrane-bounded structures basal to the described
binding of ruthenium red to acidic polysaccharides on line in cross-sectioned tubules. The importance of using
the luminal cell membranes followed by oxidation of cross-sectioned tubule cells as the basis for these eval-
ruthenium red by OsOl resulting in a final oxidation of uations is illustrated by Fig. 4.

Downloaded from www.physiology.org/journal/ajprenal by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 11, 2019.

 KINETICS OF ENDOCYTOSIS STUDIED WITH RUTHENIUM RED F247
The ultrastructural appearance of the ruthenium red- lysosomes or by the recycling of membrane to the luminal
stained proximal tubule is similar to the findings of other cell surface through dense apical tubules as originally
investigators using microinfusion of a lanthanum solu- suggested by Maunsbach (23) and later substantiated
tion (32) or ruthenium red (25) into rat proximal tubule. in the rat proximal tubule (2,4,7, 14,17). Micropuncture
Other studies have shown extensive penetration of lan- studies based on the labeling of luminal membrane with
thanum nitrate or ruthenium red into the basolateral cationized ferritin (4) or antibodies against receptors
space by means of different staining procedures (19, 33). within the membrane (2) have suggested that this is a
This was not observed in this study. However, the stain- rapid process occurring within minutes, but so far no
ing in these different studies cannot be directly com- studies have investigated the kinetics of this dynamic
pared, as the staining technique differs either by the process. On the basis of the described model of endocy-
chemical used for staining or by the staining procedure. tosis and on the established surface densities of various
Studies on the permeability of tight junctions using tu- compartments, we have determined the kinetics of the
bular microinfusion of lanthanum show some penetration intracellular transfer of endocytosed membrane. The
into the most apical parts of the intercellular space (32). model illustrated in Fig. 7 involves transfer of membrane
In this study ruthenium red was only very occasionally from invaginations to a vacuolar compartment followed
found within the lateral intercellular space. by transfer to either lysosomes for degradation or dense
Other studies have determined the surface density
apical tubules for recycling to the luminal membrane.
of large endocytic vacuoles, lysosomes, dense apical tu-
bules, and small membrane-bounded profiles (usually de- The velocity (u) of membrane transport expressed as
fined as small endocytic vesicles only) (e.g., 5, 7). The the area of membrane transport per second per cell
results of these studies are very similar to the results of volume (equal to the transferred surface density per
this study. A study by Cui et al. (7) on Sl proximal second) is assumed to be proportional to the surface
tubules shows similar surface densities of large endocytic density of the compartment from which it is transported,
vacuoles, lysosomes, and dense apical tubules, and, fur- i.e., v = kSv, in which k is the velocity constant. In
thermore, the surface density of small endocytic vacuoles the steady-state situation the velocity (ul) of membrane
is similar to the total surface density of both invagi- transported from invaginations to the vacuolar compart-
nations and small endocytic vesicles in Sl segment of ment must be equal to sum of the velocity of membrane
our study (0.41 t 0.07 vs. 0.48 t 0.11 pm2/pm3). This transport from the vacuolar compartment to dense apical
suggests that structures previously defined as small tubules ( u2) and to lysosomes ( u3). If r represents the
endocytic vacuoles include both invaginations and en- percentage of membrane transported to dense apical
docytic vesicles. tubules, i.e., r = uzl(u2 + u3), then
The endocytic pathway in proximal tubule involves
4 = v2 + v3 = u2 + u2(1 - r)/r = u2(l/r)
internalization of membrane from invaginations into a
vacuolar compartment. This is followed by transport to kS
1 v Inv IS VVac = k2(11r)

Dense apical tubules

Li
\ \\
)_
v, = k,. Svlnr

Endocytic vesicles and vacuoles Fig. 7. Model of endocytosis and recycling on which calcula-
tions of membrane transport kinetics are based. Membrane is
transported from invaginations into a vacuolar compartment
involving both small and large endocytic vacuoles. Further
transfer involves either recycling through dense apical tubules

.‘.‘.‘.‘..
,.:.:::y:.::
,,.. .,’..::.‘.,
.::.
‘,.:.;
.;. .:.:..,;..
‘.‘.‘.‘..:‘..‘.‘.
.::
.0:.‘..
......
..‘:‘.:.
..:.
::. ,...
v3= k,. Svv,
back to the luminal membrane or transport into lysosomes.
Velocity u of membrane transport is assumed to be propor-

.:.;:::.
Lysommes
/

::;;y:-
.
‘.;
‘. :,
_...,’
_.,.
...‘;:
..;,,.:,..:..
:.‘,‘;.
:.:;
.:
‘:;i,:
,I.:
i ;I:
ir:I’:::.
tional to surface density (Sv) of the compartment from which
it is transported. See DISCUSSION for further description of

0.I,
.::‘.‘:’
.,. displayed equations.

Downloaded from www.physiology.org/journal/ajprenal by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 11, 2019.

F248 KINETICS OF ENDOCYTOSIS STUDIED WITH RUTHENIUM RED

in which Sv rnv and Sv vat represent the surface density of In a study by Nielsen (26) only 3% of total membrane
all invaginations and the vacuolar compartment, respec- label was confined to lysosomes after 30 min, suggesting
tively. Finally the velocity (Q) of recycling of membrane that 97% is recycled through dense apical tubules equiv-
from dense apical tubules to the luminal cell surface must alent to r = 0.97.
be equal to the velocity (Q) of membrane transport from Using that
the vacuolar compartment into dense apical tubules kS1 V Inv IS VVac = Iz,(llr)
kS2 VVac= v2= v4 = Iz,SvDAT and
k4 = k2&7 Vac/SV DAT k4 = kS2 VVac IS V DAT = rklSV Inv/& DAT

To establish the velocity constants kl, k2, k3, and k4 for we can determine the velocity constant kI using bisec-
each of the described processes, we applied the results of tion analysis as r = 0.97 and corresponding values
a study by Christensen (4) based on microinjection of and T obtained from the study of
of SV fer-DAT(T)
cationized ferritin into rat proximal tubule. It was shown Christensen (4) are inserted. The velocity constants k2,
that 15, 30, and 50% of all dense apical tubules were k3, and k4 can then be calculated according to the above,
labeled with ferritin when observing tubules fixed 20, 60, i.e., k2 = rklSVInv/SVVac9 k3 = k2[(1 - r)lr19 and k4 =
and 90 s after microinjection, respectively. Consequently lzeSv Vat /Sv DAT. The results of these calculations for T
the amount of membrane transported from invaginations equal to 20, 60, and 90 s are shown in Table 5. The
to dense apical tubules equals 0.15, 0.30, and 0.50 times velocity of membrane transport between the various com-
the surface density of dense apical tubules, respectively. partments is calculated by multiplying the velocity con-
The surface density Sv fer-vac(t) of ferritin-labeled mem- stant with the surface density of the corresponding
brane in the vacuolar compartment at time t after micro- compartment.
injection can be calculated as the amount of ferritin- The velocity constants from data of 20 s are about
labeled membrane transferred from invaginations minus twofold the velocity constants from data of 60 s and 90 s.
the sum of ferritin-labeled membrane removed to dense This variation in results is to be expected as the ferritin-
apical tubules and to lysosomes, i.e. injected tubules were fixed by dripping the fixative on
S V fer-Vac( the surface of the kidney. Studies have shown that the
t)
diffusion time of the fixative into the kidney is of the
t order of 33 pm/min (16), implying that the tubules were
=
[k 1 s VInv - (‘2’V fer-Vat(t) + k,SV fer-Vac(t))l dt not fixed immediately at the time of dripping. The addi-
s 0 tional time before fixation due to penetration is similar
t at 20, 60, and 90 s and consequently relatively more im-
SV portant at shorter fixation times, causing T to be too
fer-Vat(t) = [k 1 s V Inv - k2(11r) ‘V fer-Vac(t)l dt
s 0
small and thereby the calculated kI to be too high. Fol-
lowing this, the kI values can be corrected by adding the
in which Sv Inv is the surface density of invaginations penetration time of the fixative to each of the T variables.
which are fully labeled with ferritin. Solving this we get On micrographs the minimum distance from the kidney
surface to the lumen of the most superficial tubule is - 10
‘lsV Inv IzlsV Inv
S V fer-Vac( t) = k,o - k,o l exPb-Mll~)tl pm, equal to a penetration time for the fixative of -20 s.
The resulting velocity constants and velocities of mem-
The surface density Sv fe&AT(T) of ferritin-labeled mem- brane transport calculated after correcting all T variables
brane in dense apical tubules at time T can then be cal- by using T + 20 s are shown in Table 5. Velocity constant
culated as the amount of ferritin-labeled membrane are now similar at all three times of fixation. The average
transferred from the vacuolar compartment minus the velocity constants and membrane transport velocities cal-
sum of ferritin-labeled membrane recycled to the luminal culated on the basis of these corrected values of T for all
surface, i.e. three times of fixation are shown in the last column of
Table 5. A further improvement in the estimation of fix-
ation time would require experiments involving microin-
S V fer-DAT(T) = T(k 2 S V fer-Vac( t ) - k4SV fer-DAT(t )) dt jection of fixative and thereby instant fixation shortly
s 0 after the microinjection of cationized ferritin.
T The amount of membrane equal to the entire surface
kS kS
S V fer-DAT(T) =
density of invaginations is internalized within 78 s, and
k2 m-m
Sr( 0 membrane equivalent to the total surface density of the
vacuolar compartment is transported to dense apical tu-

Solving this we get

l exp[-k2(1/r)t1
)
- k4SVfer-DAT(t)
1
dt
bules or lysosomes within 43 s. The amount of membrane
equal to the surface density of dense apical tubules is
recycled within 92 s. The membrane equivalent to the
surface density of lysosomes is transferred from the
KS
1 V Inv 1 exP~-k,(l/Wl vacuolar compartment within -23 min. This does not
SV fer-DAT( T) include supplement of membrane to the lysosomes from,
= - l/r 1 IF4 + k,(l/r) - k,
e.g., the Golgi region. Thus the internalization and
- 1
K+ k2(llr)-
1
Fz,
1 ' exp(-k4T)

Downloaded from www.physiology.org/journal/ajprenal by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 11, 2019.

recycling process is very fast, and return to the luminal
membrane is possible within a few minutes. The surface
 KINETICS OF ENDOCYTOSIS STUDIED WITH RUTHENIUM RED F249

kl, lo-” s-l 28 15 16 14 11 13 13&l

k2, 1o-3 s-l 49 27 27 25 20 22 22&Z
k3, lo-” s-l 1.5 0.82 0.84 0.76 0.62 0.69 0.64t0.07
kq, lo-” s-l 24 13 13 12 9.7 11 lltl
vl, lo-” ~m2.~rn-“s-l 14.0 7.6 7.8 7.0 5.7 6.4 6.4t0.7
v2 = v4, 10-3 pm2.pmv3.s-1 13.6 7.3 7.6 6.8 5.5 6.2 6.2~10.6
v3, lo-” pm2 pm-“. s-l
l 0.4 0.2 0.2 0.21 0.17 0.19 0.19t0.02
Equation v = km S, is calculated on basis of ferritin-labeled membrane transport into dense apical tubules after 20, 60, and 90 s (4). v, velocity;
k, velocity constants; S v, surface density. Results are given before and after correction for diffusion time during dripping fixation. See DISCUSSION
for descriptions of k,-k, and vl-v4. Values of Sv are obtained from Fig. 5 (see also Fig. 7).

density of dense apical tubules relative to the surface ing by ruthenium red (11). Other studies suggest that
density of the endocytic vacuolar compartment (small coated structures remain surface bound by narrow necks
and large endocytic vacuoles) varies from 2.1 (Sl) to 2.3 and that noncoated receptosomes are formed from the
(S2), and, accordingly, the velocity constant k4 for the closed coated pits (27, 35-37). This is partly based on
recycling from dense apical tubules to the luminal mem- either membrane staining and serial sectioning (36) or
brane is smaller than k2 for the transfer of membrane membrane staining by ruthenium red (37). However,
from the vacuolar compartment to dense apical tubules. only very rarely are necks connecting invaginations
This shows that membrane and recycling membrane pro- with the luminal surface observed in the proximal tubule
teins are contained for a considerably longer time in the cell. Furthermore, the diameter of the ruthenium red mol-
dense apical tubules than in the vacuolar compartment. ecule is estimated to be 1.13 nm (20), implying that
The vacuolar compartment includes both small (~0.5 ruthenium red should be capable of penetrating even
pm) and large endocytic vacuoles, and studies have shown very narrow passages, thereby staining all surface-con-
that endocytosis involves transfer from small endocytic nected structures. This strongly suggests that the ob-
vesicles into larger vacuoles (4, 21, 22). Dense apical tu- served nonstained, coated vesicular structures are in fact
bules can be observed connected to both size vacuoles, coated vesicles.
showing that recycling occurs from both. This supports In conclusion, the sandwich-staining procedure involv-
the presented model of endocytosis. This model does not ing ruthenium red, osmium, and thiocarbohydrazide as
exclude the possibility that the kinetics of recycling are a molecular bridge increased staining of luminal mem-
different depending on whether dense apical tubules are branes in renal proximal tubule. By staining luminal
formed early (i.e., from small endocytic vesicles) or late membranes including endocytic invaginations connected
(i.e., from large endocytic vacuoles) during endocytosis. to the surface, we have determined the distribution of
However, calculations based on such a refined model of invaginations and noncoated and coated endocytic vesi-
endocytosis requires more detailed information on the cles in addition to the surface density of invaginations,
dynamics of endocytosis than can be obtained from endocytic vesicles, large endocytic vacuoles, dense apical
present studies. tubules, and lysosomes. Using these parameters we have
A study on rat fibroblasts has shown that the half-life developed a kinetic model for the transport of membranes
of plasma membrane constituents in lysosomes is ~48 from endocytic invaginations through an endocytic intra-
min (8). In the present study the entire membrane area cellular compartment to dense apical tubules and lyso-
equivalent to the surface of lysosomes is transported from somes as well as recycling through dense apical tubules.
vacuoles in 23 min. Consequently the endocytic and deg- This has established that the velocity of membrane trans-
radational process is significantly faster in the renal prox- port is high, as the surface density of invaginations is
imal tubule. This is consistent with results showing that
internalized within 78 s. Furthermore, we have developed
fibroblast-derived cells (L-cells) internalized their entire a technique for predicting the distribution of endocytic
surface membrane within 125 min (31) compared with invaginations and small endocytic vesicles in nonstained
the endocytosis of a membrane area equivalent to the
rat kidney proximal tubule.
surface of invaginations within 78 s in renal proximal
tubule. Macrophages internalize the equivalent of their
total surface membrane within 33 min (31). We appreciate the skillful technical assistance by Helle Bergman,
Hanne Sidelmann, Inger Kristoffersen, Poul Boldsen, and Jytte Krage-
The findings of this study included unstained, coated lund. We thank Dr. N. C. Nielsen (Institute of Chemistry, Univ. of
membrane-bounded structures generally located very Aarhus) for valuable help concerning mathematical evaluations.
close to the luminal membrane representing coated vesi- This work was supported by the Danish Medical Research Council,
cles in the rat kidney proximal tubule. The relative num- the University of Aarhus, Kong Christian den Tiendes Foundation, the
Biomembrane Research Center University of Aarhus, Aarhus Univer-
ber of apparently coated vesicles varied from 11% (Sl) to sity Research Foundation, Novo Nordisk Foundation, Danish Founda-
15% (S2) of all small vesicles. Coated vesicles have been tion for the Advancement of Medical Science, Diabetesforeningen, Ruth
shown to be part of the endocytic pathway in many stud- I. E. K@nig Petersen Foundation, Leo Nielsen and Margrethe Nielsen
ies using different cell types (1, 9-l 1, 28, 29). Some of Foundation, and the Helen and Ejnar Bjtirnows Foundation.
these studies are based on serial sections (9,28), and some Address reprint requests to H. Birn.
are based on both serial sectioning and membrane stain- Received 5 March 1992; accepted in final form 28 May 1992.
Downloaded from www.physiology.org/journal/ajprenal by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 11, 2019.
F250 KINETICS OF ENDOCYTOSIS STUDIED WITH RUTHENIUM RED
REFERENCES 19. Larsson, L. Ultrastructure and permeability of intercellular con-
1. Anderson, R. G. W., M. S. Brown, and J. L. Goldstein. tacts of developing proximal tubule in the rat kidney. J. Ultra-
struct. Res. 52: 100-113, 1975.
Role of the coated endocytic vesicle in the uptake of receptor-
20. Luft, J. H. Ruthenium red and violet I. Chemistry, purification,
bound low density lipoprotein in human fibroblasts. Cell 10: 351-
364, 1977. methods of use for electron microscopy and mechanism of action.
Anat. Rec. 171: 347-368, 1971.
2. Birn, H., J. Selhub, and E. I. Christensen. Internalization
21. Maunsbach, A. B. Absorption of 1125-labeled homologous albu-
and intracellular transport of folate-binding protein in rat kidney
proximal tubule. Am J. Physiol. 264 (Cell Physiol. 33): C302-C310, min by rat kidney proximal tubule cells. J. UZtrastruct. Res. 15:
197-241, 1966.
1993.
22. Maunsbach, A. B. Absorption of ferritin by rat kidney proximal
3. Brown, D., S. Hirsh, and S. Gluck. Localization of a proton-
pumping ATPase in rat kidney. J. CZin. Invest. 82: 2114-2126, tubule cells. J. Ultrastruct. Res. 16: 1-12, 1966.
1988. 23. Maunsbach, A. B. Cellular mechanisms of tubular protein trans-
4. Christensen, E. I. Rapid membrane recycling in renal proximal port. In: International Review of Physiology. Kidney and Urinary
tubule cells. Eur. J. Cell BioZ. 29: 43-49, 1982. Tract Physiology II, edited by K. Thurau. Baltimore, MD: Uni-
5. Christensen, E. I., and Proteinuria
A. B. Maunsbach. versity Park, 1976, vol. 11, p. 145-167.
induced by sodium maleate in rats: effects on ultrastructure and 24. Maunsbach, A. B., and E. I Christensen. Functional ultra-
protein handling in renal proximal tubule. Kidney Int. 17: 77l- structure of the proximal tubule. In: Handbook of Physiology. Re-
787, 1980* naZ PhysioZogy. Bethesda, MD: Am. Physiol. Sot., 1992, sect. 8,
6. Christensen, E. I., and S. Nielsen. Structural and functional vol. 1, chapt. 2, p. 41-107.
features of protein handling in the kidney proximal tubule. 25. Molitoris, B. A., S. A. Falk, and R. H. Dahl. Ischemia-in-
In: Seminars in Nephrology, edited by C. C. Tisher and K. M. duced loss of epithelial polarity. Role of the tight junction. J. CZin.
Madsen. Philadelphia, PA: Saunders, 1991, vol. 11, p. 414- Invest. 84: 1334-1339, 1991.
439. 26. Nielsen, S. Recycling of insulin receptors in renal proximal tu-
7. Cui, S., E. I. Christensen, and Membrane traffic
S. Nielsen. bules (Abstract). Eur. Workshop Endocytosis 2nd Paris 1990, p.
after inhibition of endocytosis in renal proximal tubules. J. Struct. 91.
BioZ. 107: 201-210, 1991. 27. Pastan, I. H., and M. C. Willingham. Journey to the center of
8. Draye, J.-P., P. J. Courtoy, J. Quintart, and P. Baudhuin. the cell: role of the receptosome. Science Wash. DC 214: 504-509,
A quantitative model of traffic between plasma membrane and 1981.
secondary lysosomes: evaluation of inflow, lateral diffusion and 28. Petersen, 0. W., and B. Van Deurs. Serial-section analysis of
degradation. J. CeZZ BioZ. 107: 2109-2115, 1988. coated pits and vesicles involved in adsorptive pinocytosis in cul-
9. Fan, J. Y., J.-L. Carpentier, P. Gorden, E. V. Obberghen, tured fibroblasts. J. Cell BioZ. 96: 277-281, 1983.
N. M. Blackett, C. Grunfeld, Receptor-mediated
and L. Orci. 29. Roth, T. F., and K. R. Porter. Yolk protein uptake in the
endocytosis of insulin: role of microvilli, coated pits, and coated oocyte of the mosquito Aedes aegypti L. J. CeZZ BioZ. 20: 313-332,
vesicles. Proc. NatZ. Acad. Sci. USA 79: 7788-7791, 1982. 1964.
10. Friend, D. S., and M. G. Farquhar. Functions of coated ves- 30. Seligman, A. M., H. L. Wasserkrug, and J. S. Hanker. A
icles during protein absorption in the rat vas deferens. J. Cell BioZ. new staining method (OTO) for enhancing contrast of lipid-con-
35: 357-376, 1967. taining membranes and droplets in osmium tetroxide-fixed tissue
11. Geoffroy, J. S., and R. P. Becker. Endocytosis by endothelial with osmiophilic thiocarbohydrazide (TCH). J. Cell BioZ. 30: 424-
phagocytosis: uptake of bovine serum albumin-gold conjugates in 432, 1966.
bone marrow. J. Ultrastruct. Res. 89: 223-239, 1984. 31. Steinman, R. M., S. E. Brodie, and 2. A. Cohn. Membrane
12. Gundersen, H. J. G. Notes on the estimation of the numerical flow during pinocytosis. A stereological analysis. J. Cell BioZ. 68:
density of arbitrary profiles: the edge effect. J. Microsc. 111: 219- 665-687, 1976.
223, 1977. 32. Tisher, C. C., and
W. E. Yarger. Lanthanum permeability of
13. Haase, W., and H. Koepsell. Electron microscopic immunohis- the tight junction (zonula occludens) in the renal tubule of the rat.
tochemical localization of components of Na+-cotransporters Kidney Int. 3: 238-250, 1973.
along the rat nephron. Eur. J. Cell BioZ. 48: 360-374, 1989. 33. Walski, M., and K. Olszewska. Electronmicroscopic his-
14. Hatae, T., M. Fujita, H. Sagara, Forma-
and K. Okuyama. tochemical investigation of the surface of the cell membrane of the
tion of apical tubules from large endocytic vacuoles in kidney renal tubules with the use of ruthenium red and lanthanum ni-
proximal tubule cells during absorption of horseradish peroxidase. trate. FoZia Histochem. Cytochem. 16: 19-25, 1978.
Cell Tissue Res. 246: 271-278, 1986. 34. Weibel, E. R. Stereological methods. In: Practical Methods for
15. Hjelle, J. T., E. I. Christensen, F. A. Carone, and J. Sel- Biological Morphometry. New York: Academic, vol. 1, 1979.
hub. Cell fractionation and electron microscope studies of kidney 35. Willingham, M. C., and I. Pastan. The receptosome: an inter-
folate-binding protein. Am. J. Physiol. 260 (Cell Physiol. 29): mediate organelle of receptor-mediated endocytosis in cultured
C338-C346, 1991. fibroblasts. Cell 21: 67-77, 1980.
16. Jaremko, G., 6. Kallskog, and S.-O. Bohman. In vivo drip- 36. Willingham, M. C., and I. Pastan. Formation of receptosomes
ping fixation of the superficial renal cortex (Abstract). Proc. Int. from plasma membrane coated pits during endocytosis: analysis
Congr. Nephrol. 11 th Tokyo 1990, p. 365A. by serial sections with improved membrane labeling and preser-
17. Kawai, Y ., and T. Hatae. Three-dimensional representation vation techniques. Proc. NatZ. Acad. Sci. USA 80: 5617-5621,
and quantification of endosomes in the rat kidney proximal tubule 1983.
cell. J. Electron Microsc. 40: 411-415, 1991. 37. Willingham, M. C., A. V. Rutherford, M. G. Gallo, J. We-
18. Kerjaschki, D., L. Noronha-Blob, B. Sacktor, and M. hland, R. B. Dickson, R. Schlegel, Recep-
and I. H. Pastan.
Farquhar. Microdomains of distinctive glycoproteins in the kid- tor-mediated endocytosis in cultured fibroblasts. Cryptic coated
ney proximal tubule brush border. J. Cell BioZ. 98: 1505-1513, pits and the formation of receptosomes. J. Histochem. Cytochem.
1984. 29: 1003-1013, 1981.

Downloaded from www.physiology.org/journal/ajprenal by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 11, 2019.