Taken together, our results indicate that the GD3–R24 complex is endocytosed in GD3-expressing cells, accu-mulates in the recycling endosome, and is transported back to the plasma membra
Trang 1endocytosed and recycled to the plasma membrane via the endocytic recycling compartment
Inhibitory effect of brefeldin A and monensin
Ramiro Iglesias-Bartolome´, Pilar M Crespo, Guillermo A Gomez and Jose L Daniotti
Centro de Investigaciones en Quı´mica Biolo´gica de Co´rdoba, CIQUIBIC (UNC-CONICET), Departamento de Quı´mica Biolo´gica,
Universidad Nacional de Co´rdoba, Argentina
Gangliosides are complex glycosphingolipids
contain-ing one or more sialic acid residues, which are mainly
located at the outer leaflet of the plasma membrane of
eukaryotic cells They participate in cell-surface events
such as modulation of growth factor receptors and
cell-to-cell and cell-to-matrix interactions [1–5] They
are synthesized in the lumen of the Golgi complex by
a complex system of membrane-bound glycolipid
acceptors, glycosyltransferases, and sugar nucleotide transporters [6] After their synthesis, gangliosides leave the Golgi complex via the lumenal surface of transport vesicles We recently demonstrated that the ganglioside NeuAca2,8NeuAca2,3Galb1,4Glc ceramide (GD3) is transported from the trans-Golgi network (TGN) to the plasma membrane via a Rab11-independent, brefeldin A (BFA)-insensitive exocytic
Keywords
endocytic recycling; gangliosides; glycolipid
antibodies; intracellular transport; R24
antibody
Correspondence
J.L Daniotti, CIQUIBIC (UNC-CONICET),
Departamento de Quı´mica Biolo´gica,
Facultad de Ciencias Quı´micas, Universidad
Nacional de Co´rdoba, Ciudad Universitaria,
5000 Co´rdoba, Argentina
Fax: +54 3514334074
Tel: +54 3514334171
E-mail: daniotti@dqb.fcq.unc.edu.ar
(Received 15 January 2006, revised 14
February 2006, accepted 20 February 2006)
doi:10.1111/j.1742-4658.2006.05194.x
Gangliosides are sialic acid-containing glycosphingolipids present on mam-malian plasma membranes, where they participate in cell-surface events such as modulation of growth factor receptors and cell and cell-to-matrix interactions Antibodies to gangliosides have been associated with
a wide range of clinically identifiable acute and chronic neuropathy syn-dromes In addition, antibodies to tumor-associated gangliosides are being used as therapeutic agents Their binding to and release from cell mem-branes and intracellular destinations have not so far been extensively exam-ined In this study, we characterized in both GD3 ganglioside-expressing Chinese hamster ovary (CHO)-K1 and SK-Mel 28 melanoma cells the intracellular trafficking and subcellular localization of the mouse monoclo-nal antibody to GD3, R24 By biochemical techniques and detailed confo-cal microscopic analysis, we demonstrate that the GD3–R24 antibody complex is rapidly and specifically internalized by a dynamin 2-independent pathway and then accumulates in the endocytic recycling compartment In addition, we show that the R24 antibody exits the recycling compartment
en route to the plasma membrane by a dynamin 2-dependent pathway sen-sitive to brefeldin A and monensin Taken together, our results indicate that the GD3–R24 complex is endocytosed in GD3-expressing cells, accu-mulates in the recycling endosome, and is transported back to the plasma membrane via a route that involves clathrin-coated vesicles
Abbreviations
BFA, Brefeldin A; CHO, Chinese hamster ovary; DMEM, Dulbecco’s modified Eagle’s medium; GFP, green fluorescent protein; GalNAc-T, UDP-GalNAc–LacCer ⁄ G3 ⁄ GD3 N-acetylgalactosaminyltransferase; GD3, NeuAca2,8NeuAca2,3Galb1,4Glc ceramide; GM3,
NeuAca2,3Galb1,4Glc ceramide; Sial-T2, CMP-NeuAc–GM3 sialyltransferase; TGN, trans-Golgi network; YFP, yellow fluorescent protein.
Trang 2pathway [7] After their arrival at the plasma
mem-brane, gangliosides undergo endocytosis Once
inter-nalized, glycosphingolipids can be: (a) recycled to the
plasma membrane directly from early endosomes; (b)
sorted from endosomes to the Golgi complex, where
they can be reglycosylated; or (c) degraded at the
lyso-somal level [8]
Antibodies to more than 20 different glycolipids,
including gangliosides, have been associated with a
wide range of clinically identifiable acute and chronic
neuropathy syndromes, including Guillain–Barre´
syn-drome (anti-GM1, anti-GM2, anti-GQ1b), chronic
idiopathic ataxic neuropathy (anti-GM3, anti-GD1b,
anti-GD3, anti-GQ1), Miller–Fisher syndrome
(anti-GQ1b) and multifocal motor neuropathy (anti-GM1)
[9,10] In addition, antibodies to tumor-associated
gangliosides are being used as therapeutic agents, for
example, GD2 for neuroblastoma [11] and
anti-GD3 for melanoma [12,13] Several targeted therapies
need the antibody to remain at the cell surface to
mediate antibody-dependent and
complement-depend-ent cytotoxicity However, rapid internalization of the
antibodies to intracellular compartments is desired for
cytotoxic effects of toxins or cytotoxic drugs
conju-gated with the antibody [14] Neither the binding and
release of ganglioside antibodies from cell membranes
nor their intracellular destinations have so far not been
extensively tested
In this study, we characterized in both
GD3-expressing Chinese hamster ovary (CHO)-K1 and
SK-Mel 28 melanoma cell lines the binding, intracellular
trafficking, and subcellular localization of the mouse
monoclonal antibody to GD3, R24(IgG3) By
bio-chemical techniques, immunofluorescence and confocal
microscopic analysis, we demonstrate that the
GD3–R24 complex was rapidly and specifically
inter-nalized and accumulated mainly in a perinuclear
compartment This compartment was positive for
expression of the GTPase Rab11 and internalized
transferrin, which are two recycling endosome
mark-ers, but not for UDP-GalNAc–LacCer⁄ G3 ⁄ GD3
N-acetylgalactosaminyltransferase (GalNAc-T), a TGN
marker In addition, we show that R24 exited the
recycling compartment by a dynamin 2-dependent
pathway sensitive to BFA and monensin Interestingly,
after 60 min of endocytosis of the R24 antibody, most
of the antibody was recovered from the culture
medium Taken together, our results indicate that
the GD3–R24 complex is rapidly endocytosed in
CHO-K1 and SK-Mel 28 melanoma cells, accumulates
in the recycling endosome, and is transported back
to the plasma membrane via a route involving
clathrin-coated vesicles
Results
GD3–R24 antibody complex is rapidly internalized
in GD3-expressing CHO-K1 cells
We established an antibody-binding technique to track the fate of surface GD3–R24 after its internalization in
a GD3-expressing CHO-K1 cell clone (clone 2), already established in our laboratory by stable expres-sion of CMP-NeuAc–GM3 sialyltransferase (Sial-T2) cDNA [7,15,16] Briefly, cells from clone 2 were incu-bated for 10 min on ice to inhibit intracellular trans-port and then with R24 [17] for 45 min on ice Then, cells where extensively washed with cold buffer to remove unbound antibody, and the temperature chan-ged to 37C to restore transport and thereby allow endocytosis of GD3–R24 for different times Confocal microscopic analysis revealed that R24 bound to live cells at 4C had a plasma membrane punctate distri-bution (Fig 1A), as previously demonstrated in this cell line [18] After induction of endocytosis by chan-ging the temperature to 37C, GD3–R24 was found in vesicles all over the cytoplasm, and at 15 min it began
to show a perinuclear distribution After 30 min at
37C, the intracellular pool of R24 became more con-centrated in the perinuclear region and the plasma membrane mark almost disappeared (Fig 1A) At
60 min after the beginning of endocytosis, almost all cells were negative for R24 staining unless the anti-body was present constantly in the cell culture med-ium It should be mentioned that, at the steady-state, most of the GD3 ganglioside was found to be present
in the plasma membrane in cells from clone 2, although a fraction was also observed in endosomal structures [18] Endocytosis of R24 seems to be specif-ically mediated by GD3, as wild-type CHO-K1 cells, which only express the GM3 ganglioside, did not bind (Fig 1B, 45 min at 4C) and internalize R24, even with antibody in the culture medium for 2 h at 37C (Fig 1B)
Next, we performed a western blot to quantify the proportion of endocytosed R24 at different times (Fig 1C) Thus, cells were incubated with R24 at 4C
to label the surface pool of GD3 ganglioside, and then shifted to 37C to restore transport Cells at 15, 30,
60 and 90 min were acid stripped to remove mem-brane-bound antibody, harvested, and the presence of internalized R24 analyzed by western blot To evaluate the efficiency of surface stripping by acid washing, one sample was acid-treated directly after incubation at
4C; under these conditions the antibody was com-pletely removed from the cell surface (result not shown) About 50% of the antibody bound to the cell
Trang 3surface at 4C was efficiently endocytosed after
15 min at 37C After 30 min, the pool of intracellular
antibody had decreased to 25%, and, at 60 and 90 min
after internalization, the antibody concentration was
below the limit of detection by western blot These
results confirm the immunofluorescence and confocal
microscopic findings Taken together, the results show
that R24 is rapidly internalized in GD3-expressing
CHO-K1 cells, intracellularly accumulated, and later
degraded and⁄ or depleted from subcellular
compart-ments
R24 antibody is not targeted and degraded in
lysosomes
To assess whether the R24 antibody was targeted to
lysosomes during endocytic transport, we performed
a colocalization analysis of internalized antibody
with the acidotropic probe LysoTracker Red Results
shown in Fig 2A clearly indicate that endocytosed
R24 did not significantly colocalize with the
lyso-some marker at 30 min Comparable results were
also obtained at 15 and 45 min after endocytosis (results not shown) R24 was found colocalized to a minor extent with the acidotropic probe in small acidic organelles, which probably represents endo-somes These results suggest that little R24 anti-body enters lysosomes or that the epitope for the R24 antibody is rapidly lost on transport into lyso-somes
To evaluate further whether the antibody is degra-ded in lysosomes, internalization and intracellular transport of R24 was performed in the presence of
NH4Cl and chloroquine, which are inhibitors of lyso-some degradation Cells from clone 2 were allowed to internalize R24 in the absence (control) or presence
of 30 mm NH4Cl or 60 lm chloroquine over 90 min
at 37C Cells were then harvested and the presence
of R24 analyzed by western blot As clearly shown in Fig 2B, the inhibitors could not prevent cellular depletion of R24 antibody Taken together, these results indicate that most of the R24 antibody is not targeted to and degraded in lysosomes after its internalization
A
Fig 1 R24 is rapidly and specifically endocytosed in GD3-expressing CHO-K1 cells (A) CHO-K1 cells (GD3+) were incubated with R24 for
45 min at 4 C After washing of the cells, the temperature was shifted to 37 C to allow endocytosis of the complex GD3–R24, and cells were fixed at 15, 30 and 60 min R24 antibody was detected by using anti-mouse IgG conjugated with Alexa 488 Single confocal sections of 0.7 lm were taken parallel to the coverslip Cell boundaries (white lines) are indicated at 30 and 60 min The perinuclear region is also indic-ated (arrows) at 15 and 30 min The fluorescence micrographs shown are representative of five independent experiments (B) Wild-type CHO-K1 cells (wt, GD3–) were incubated with R24 for 45 min at 4 C Then the temperature was shifted to 37 C, and cells were fixed at
2 h R24 antibody detection was carried out as indicated in (A) (C) Cells from clone 2 (GD3+) were incubated with R24 at 4 C to label the surface pool of GD3 ganglioside, and then shifted to 37 C to restore transport Cells at 15, 30, 60 and 90 min were acid stripped to remove membrane-bound antibody, harvested, and the presence of internalized R24 analyzed by western blot The expression of Sial-T2–hemagglut-inin (GD3 synthetase) in the same membrane was analyzed as a control of protein loading IgG (L), Light chain of IgG Scale bars: 20 lm.
Trang 4The internalized R24 antibody is targeted and
transiently accumulated at the recycling
endosome
After internalization, a significant fraction of R24 was
located in a perinuclear region that resembles the
Golgi complex or the recycling endosomes as these
or-ganelles have a pericentriolar distribution in CHO-K1
cells [19] In an effort to identify the juxtanuclear
com-partment where the endocytosed R24 antibody is
con-centrated, we performed extensive colocalization with
markers of both recycling endosome and Golgi
com-plex (Fig 3) After 30 min of internalization, no
colo-calization was observed between R24 and GalNAc-T
fused to yellow fluorescent protein (YFP), a TGN
marker However, we observed extensive colocalization
between R24 and the GTPase Rab11, an established
recycling endosome marker [19] In addition, we also
found substantial overlapping of R24 with
coendocyto-sed Alexa647-transferrin in a perinuclear compartment,
demonstrating that a significant fraction of
endocyto-sed R24 in CHO-K1 cells was present in the recycling
endosome The fraction of perinuclear labeling of R24
that does not overlap with transferrin and Rab11 would be associated with different recycling endosome membranes, as it has been suggested, on the basis of cellubrevin and endocytosed transferrin juxtanuclear localization, that these compartments may be subdivi-ded into distinct populations [20]
The R24 antibody is recycled to the plasma membrane and released into the culture medium
As shown previously, we found that most of the endo-cytosed R24 was not targeted to lysosomes but transi-ently accumulated at the recycling endosome After 60
or 90 min of R24 endocytosis, we could not detect the internalized antibody by biochemical and immunologi-cal techniques An explanation for these results is that R24 may be recycling from the pericentriolar endocytic compartment to the plasma membrane where it is released into the culture medium
To address this issue, cells from clone 2 were incuba-ted for 10 min on ice to inhibit intracellular transport and then with R24 for 45 min on ice Afterwards, cells were allowed to internalize the antibody for 20 min at
A
B
Fig 2 R24 antibody is not degraded in lysosomes (A) CHO-K1 cells (GD3+) were incubated with R24 for 45 min at 4 C After washing of the cells, the temperature was shifted to 37 C to allow endocytosis of the complex GD3–R24, and cells were fixed at 30 min R24 antibody was detected by using anti-mouse IgG conjugated with Alexa488 For lysosome staining, cells were incubated with 0.2 l M acidotropic probe LysoTracker Red DND-99 for 15 min at 37 C before fixation Single confocal sections of 0.7 lm were taken parallel to the coverslip Cell boundaries (white lines) are indicated Scale bar: 20 lm (B) GD3-expressing CHO-K1 cells were incubated with R24 for 45 min at 4 C (0 min) After washing of the cells, the temperature was shifted to 37 C to allow internalization of R24 antibody in the absence or presence
of 30 m M NH4Cl or 60 l M chloroquine over 90 min at 37 C Then cells were harvested and the presence of R24 antibody analyzed by west-ern blot The expression of Sial-T2–hemagglutinin (GD3 synthetase) in the same membrane was analyzed as a control of protein loading.
Trang 537C, and then the temperature was changed again to
4C The cell surface was then stripped of any
remain-ing antibody with acid wash At this point, cells only
contained R24 in intracellular compartments
Subse-quently, prewarmed culture medium was added to the
cells, which were maintained at 37C to restore
intra-cellular transport Cells and culture medium were
recovered at different times, and the presence of R24 in
both samples was analyzed by western blot As shown
in Fig 4, at the beginning of the time-course
experi-ment (stripped cells, 0 min) the antibody was present
only in the cell fraction At 15 min, it was detected in
both fractions (cells and culture medium), and at
60 min most of it was recovered from the culture med-ium The antibody recovered from the culture medium was found to have the expected molecular mass (whole molecule) in gels run under nonreducing conditions (results not shown) Together, these results indicate that R24, once internalized, is recycled to the plasma membrane and released into the culture medium
R24 antibody recycling is sensitive to BFA and dependent on dynamin activity
It has previously been shown that transferrin receptor recycling as well as the formation of clathrin-coated
Fig 3 The internally accumulated R24 antibody colocalizes with recycling endosome markers but not with the Golgi marker GalNAc-T CHO-K1 cells (GD3+) transiently expressing both GalNAc-T-YFP (upper row, pseudo colored green) and wild-type GFP-Rab11 (GFP-Rab11
wt, middle row) were incubated with anti-GD3 IgG for 45 min at 4 C After washing of the cells, the temperature was shifted to 37 C to allow endocytosis of the complex GD3–R24 for 30 min R24 antibody was detected by using rhodamine-conjugated goat anti-mouse IgG.
In another set of experiments, uptake of Alexa 647 -transferrin (Alexa 647 -Tf, pseudo colored green) was monitored simultaneously with R24 endocytosis (lower row) Expression of Rab11 and GalNAc-T was detected by the intrinsic fluorescence of GFP and YFP, respectively Cell boundary (white line) is indicated in the upper row Insets in merge panels show details at higher magnification The fluorescence micro-graphs shown are representative of three independent experiments Scale bars: 10 lm.
Trang 6pits on the recycling endosome is inhibited in the
pres-ence of BFA [21–23] If clathrin-coated vesicles play a
role in R24 recycling, BFA should interfere with this
pathway To test this hypothesis, GD3-expressing
CHO-K1 cells were incubated on ice with R24 for
45 min The cells were then incubated on ice in the
absence (control) and presence of 2 lgÆmL)1 BFA for
15 min At the end of this period, cells where washed
to remove unbound antibody, and then prewarmed
culture medium supplemented with BFA, when
neces-sary, was added, and the cells were transferred to
37C to allow endocytosis for different times
Results shown in Fig 5A indicate that BFA did not
affect R24 internalization, because the fraction of
internalized and accumulated antibody at 30 min was
similar in both control and BFA-treated cells As
shown previously, 90 min after internalization we
could not detect the presence of intracellular antibody;
however, in BFA-treated cells a significant fraction of the antibody remained accumulated in a perinuclear compartment, consistent with the requirement of clath-rin-coated vesicles for efficient R24 recycling The intracellular accumulation of R24 at 90 min was also detectable by western blot experiments (Fig 5B) Under these conditions, we also found in BFA-treated cells substantial overlapping of R24 with co-endocyto-sed Alexa647-transferrin and with GalNAc-T-YFP, a TGN marker (Fig 5C, first row) These results clearly demonstrate a fusion between the recycling endosomal system and the TGN in the presence of BFA, as previ-ously described [24]
Dynamins function in the pinching off of clathrin-coated vesicles It has previously been demonstrated that the transferrin receptor egresses recycling endo-somes, at least in part, by endosome-derived clathrin-coated vesicles in a dynamin-dependent manner [22]
To study the requirement for dynamin in R24 depar-ture from recycling endosome, GD3-expressing CHO-K1 cells were transiently transfected to express both wild-type (wtDyn2) and the dominant-negative form of dynamin-2 (Dyn2K44 A), a mutant defective in GTP loading and hydrolysis After 24 h, cells were incuba-ted with R24 at 4C for 45 min and then induced to internalize the antibody by shifting the temperature to
37C Cells were fixed at 30, 60 and 90 min, and the presence of R24 was visualized by immunofluorescence and confocal microscopic analysis Results shown in Fig 6 demonstrate that wild-type dynamin-2 did not affect R24 internalization (30 min) and later depletion (60 and 90 min) On the other hand, the dominant-negative version of dynamin-2 did not have much effect on R24 internalization (30 min), but it signifi-cantly affected the departure of R24 from the recycling endosome (60 and 90 min) Taken together our data indicate that recycling endosome-derived, clathrin-coa-ted vesicles play a role in the endosomal recycling pathway of the R24 antibody
R24 antibody recycling is also sensitive to monensin
It was previously proposed that endosomal acidifica-tion is a prerequisite for the actual formaacidifica-tion of the carrier structures that move the TGN proteins (TGN38 or furin) from the endosome back to the TGN [25] In an attempt to learn more about the mechanism of endosomal recycling of R24, we examined the effect of monensin, an ionophore that dissipates pH gradients across organelle membranes [26], on R24 recycling
A
B
Fig 4 R24 antibody is recycled to the plasma membrane and
released into the culture medium (A) GD3-expressing CHO-K1 cells
were incubated with R24 antibody for 45 min on ice Afterwards,
cells were allowed to internalize the antibody for 20 min at 37 C,
and then the temperature was shifted again to 4 C The cell
sur-face was then stripped of any remaining antibody with acid wash
(0 min) Cells were then incubated at 37 C to restore intracellular
transport, and cells and culture medium were recovered at 15 and
60 min The presence of R24 antibody in both samples was
ana-lyzed by western blot as indicated in Experimental procedures (B)
The relative contribution of bands in each condition was calculated
using the computer software SCION IMAGE on the scanned film
shown in (A) Ponceau S staining was used to normalize levels of
proteins seeded in each lane The band intensity for R24 antibody
at 0 min (cellular fraction) was arbitrarily taken as 1 Results are
representative of four independent experiments.
Trang 7GD3-expressing CHO-K1 cells were incubated with
the antibody for 45 min on ice Then cells were
incu-bated on ice for another 15 min in the absence
(con-trol) or presence of 10 lm monensin At the end of this period, the unbound antibody was removed by washing, and internalization was allowed to continue
A
B
C
Trang 8at 37C for different times Results shown in Fig 5A
indicate that monensin did not affect R24
internalizat-ion, as it was similar in both control and treated cells
at 30 min As described above, at 90 min after
inter-nalization we could not detect intracellular antibody in untreated cells; however, in monensin-treated cells a significant fraction of the antibody remained accumu-lated in a perinuclear compartment The intracellular
Fig 6 R24 antibody recycling is dependent
on dynamin activity GD3-expressing
CHO-K1 cells were transiently transfected to
express both wild-type (wtDyn2) and the
dominant-negative form of dynamin-2
(Dyn2K44 A) After 24 h, cells were
incub-ated with the R24 antibody at 4 C for
45 min and then allowed to internalize the
antibody by shifting the temperature at
37 C Cells were fixed at 30, 60 and
90 min, and the presence of R24 analyzed
by using rhodamine-conjugated goat
anti-mouse IgG Single confocal sections of
0.7 lm were taken parallel to the coverslip.
Arrows indicate dynamin-transfected cells.
Cell boundaries (white lines) are indicated at
30, 60 and 90 min Scale bars: 10 lm.
Fig 5 R24 antibody recycling is sensitive to BFA and monensin (A) GD3-expressing CHO-K1 cells were incubated on ice with the R24 anti-body for 45 min Then cells were incubated on ice in the absence (control) and presence of 2 lgÆmL)1BFA or 10 l M monensin for 15 min Cells where washed to remove unbound antibody and then prewarmed culture medium supplemented with BFA or monensin, when neces-sary, was added and cells transferred to 37 C to allow endocytosis for 30 and 90 min R24 antibody was detected by using anti-mouse IgG conjugated with Alexa 488 Single confocal sections of 0.7 lm were taken parallel to the coverslip Cell boundaries (white lines) are indic-ated at 30 and 90 min (B) GD3-expressing CHO-K1 cells were incubindic-ated with R24 for 45 min at 4 C (0 min) After washing of the cells, the temperature was shifted to 37 C to allow internalization of R24 antibody in the absence (C) or presence of 2 lgÆmL)1BFA or 10 l M monen-sin (Mon) over 30 and 90 min at 37 C Then cells were harvested and the presence of R24 antibody was analyzed by western blot in sam-ples containing equal amounts of total proteins (C) Same experiment as in (A) except that GalNAc-T-YFP was transiently expressed 24 h before R24 and Alexa 647 -transferrin (Alexa 647 -Tf) internalization Insets in merge panels show details at higher magnification Triple color imag-ing of simag-ingle fixed CHO-K1 cells were taken with filters for rhodamine, Cy5 and YFP (first, second, and third panels from left, respectively) The fourth panel is a merging of these images Triple colocalization is visualized in white, and colocalization between R24 antibody and Alexa 647 -transferrin is visualized in pink Scale bars: 10 lm.
Trang 9accumulation of R24 in monensin-treated cells was
also observed in western blot experiments at 30 and
90 min (Fig 5B) We also found in monensin-treated
cells overlapping of R24 antibody with co-endocytosed
Alexa647-transferrin However, in contrast with the
effect of BFA, in monensin-treated cells we only
observed a slight colocalization between R24 antibody
and GalNAc-T-YFP, a TGN marker (Fig 5C, second
row) Our initial interpretation of these results is that
they indicate that, like TGN38 and the TGN protease
furin, endosomal acidification is probably required for
R24 antibody to exit the recycling endosome
R24 antibody is also internalized, recycled to
plasma membrane, and released into culture
medium in SK-Mel 28 melanoma cells
To investigate if the internalization and recycling
path-way of R24 antibody is a common feature that occurs
in other cell types, we analyzed the endocytic transport
of this antibody in SK-Mel 28 melanoma cells The
radioactive labeling pattern of gangliosides from
SK-Mel 28 cells is shown in Fig 7A As previously
reported [17], GD3 and GM3 were the major
ganglio-sides expressed in this cell line, which run as doublets because of differences in ceramide structure Melan-oma cells were incubated with R24, the unbound anti-body was removed by washing, and internalization was induced by transferring the cells to 37C for dif-ferent periods of time As shown in Fig 7B, R24 was efficiently internalized in SK-Mel 28 cells at 15, 30 and
60 min Colocalization between endocytosed R24 and transferrin was observed after 30 min (results not shown) In this cell line, tubules from endocytic recyc-ling are distributed more widely throughout the cyto-plasm Compared with in CHO-K1 cells, the antibody showed a lower rate of intracellular disappearance In western blot experiments we observed that, at 60 and
90 min, a fraction of the antibody still remained asso-ciated with intracellular structures whereas a significant fraction was found in the culture medium (Fig 7C)
Discussion
We have followed the entire endocytic itinerary of the mouse monoclonal antibody to GD3, R24 in GD3-expressing CHO-K1 and SK-Mel 28 cells We found that endocytosed R24 first appears in a diffuse,
C
Fig 7 R24 antibody is internalized, recycled to the plasma membrane, and released into culture medium in SK-Mel 28 melanoma cells (A) SK-Mel 28 melanoma cells were metabolically labeled from [14C]galactose for 24 h Lipid extracts were prepared, purified, chromatographed and visualized as indicated in Experimental procedures The positions of cochromatographed glycolipid standards are indicated The asterisk indicates the position of an unidentified lipid (B) SK-Mel 28 melanoma cells were incubated with R24 for 45 min at 4 C After washing of the cells, the temperature was shifted to 37 C to allow endocytosis of the complex GD3–R24, and cells were fixed at 15, 30 and 60 min R24 antibody was detected by using anti-mouse IgG conjugated with Alexa 488 Single confocal sections of 0.7 lm were taken parallel to the coverslip (C) SK-Mel 28 melanoma cells were incubated with R24 antibody for 45 min on ice Cells were allowed to internalize the antibody for 20 min at 37 C, and then the temperature was shifted again to 4 C The cell surface was then stripped of any remaining antibody with acid wash (0 min) Then cells were incubated at 37 C to restore intracellular transport, and cells and culture medium were recovered at 30,
60 and 90 min The presence of R24 antibody in both samples was analyzed by western blot as indicated in Experimental procedures Scale bars: 20 lm.
Trang 10punctate distribution in the cytoplasm that is
consis-tent with the sorting compartment of the early
endo-somes Subsequently, R24 appeared concentrated in a
pericentriolar distribution which we have characterized
as the recycling endosome After that, the antibody is
recycled to the plasma membrane and released into the
culture medium by a BFA⁄ monensin-sensitive,
dynam-in-dependent recycling pathway In addition, no
evi-dence was found for targeting and degradation of R24
antibody in lysosomes These observations suggest that
R24 follows an endocytic pathway typical of other
recycling proteins, such as the model recycling protein
transferrin receptor Nevertheless, in this study we
des-cribe for the first time the entire endocytic itinerary of
a glycolipid antibody
Antibody-binding techniques are extensively used
to follow endocytic transport of proteins such as
glycosylhoshatidylinositol-anchored proteins [27–29],
major histocompatibility complex class I protein and
interleukin 2 a-subunit receptor [30], integrin b1 [31]
and cation-independent mannose 6-phosphate receptor
[32] Antibodies tend not to have a significant effect
on the endocytic behavior of the proteins studied
[31,32] Lysosome and endosome compartments have
an acidic pH, which both promotes the dissociation of
ligands such as low-density lipoprotein from their
receptors and the proper function of hydrolytic
enzymes We demonstrate that the association of R24
with the GD3 ganglioside was unaffected even after
1 h at pH 5 or 6 (Fig S1), suggesting that the GD3–
R24 complex could not be disrupted in acidic
organ-elles Taking all these antecedents together, it is
plaus-ible that the itinerary of the R24 antibody reflects the
intracellular transit of the disialo ganglioside GD3 In
this vein, it has been demonstrated that exogenous
glycosphingolipids can be internalized and directed to
the Golgi apparatus, where they can be reglycosylated
and then delivered back to the cell surface [8,33] Also,
internalized sphingolipid analogs (BODIPY-and
NBD-labeled lipids) can be recycled to the plasma
mem-brane via endosomes or through the Golgi complex
[34,35] However, it should be taken into consideration
that the quantitative and qualitative behavior of the
analog lipids is quite different from long-chain cellular
lipids, as the ability of short fluorescent lipids to
diffuse spontaneously between different membranes is
not generally shared by their endogenous lipid
counterparts [36] Even the binding of antibodies and
toxins to endogenously synthesized lipids may perturb
the natural behavior of these molecules In a similar
way, it was reported that cholera toxin alters the
internalization mechanism of a fluorescent GM1
ganglioside [37]
It has been reported that transferrin receptor and its ligand transferrin recycle to the plasma membrane with the same kinetics as certain lipids [34] and independ-ently of the transferrin receptor cytoplasmic domain [38] These data suggest that recycling of molecules from endosome to plasma membrane can occur without active recruitment of cytosolic coat proteins (bulk flow process) However, more recently dynamin-dependent transferrin receptor recycling by endosome-derived clathrin-coated vesicles was reported [22] Supporting this observation, results obtained in this work using BFA and the dominant-negative form of dynamin indicate that recycling endosome-derived, clathrin-coated vesicles may play a role in the endosomal recyc-ling pathway of the R24 antibody Previous studies using a cross-linkable form of clathrin light chain indi-cated that, once internalized, the return of the transfer-rin receptor to the cell surface was largely insensitive
to clathrin cross-linking [39], consistent with the notion that clathrin does not play a role in trafficking mole-cules from the endosome back to the plasma mem-brane However, these results do not entirely exclude the possibility that clathrin-coated pits may be oper-ating in the transport of transferrin receptor back to the plasma membrane in the absence of the cross-linker Alternatively, clathrin independent recycling pathways may also be involved in the recycling of molecules to the plasma membrane, as discussed below If it is assumed that R24 is associated with the luminal membrane-bounded GD3 during the recycling pathway, the complex must be segregated into specialized domains to be sequestered by nascent clath-rin-coated vesicles Three properties are key to lipid sorting: headgroup interactions, lipid shape and mem-brane-order parameters [40] GD3–R24 interactions with a protein may cause the lipid to be sorted on the basis of the characteristics of the protein, and such a mechanism is important for the trafficking of gly-cosylphosphatidylinositol-anchored proteins [41] We recently demonstrated that GD3 ganglioside, like gly-cosylphosphatidylinositol-anchored proteins, is mainly expressed in glycosphingolipid-enriched microdomain (also called detergent-resistant membranes or rafts), dynamic assemblies of cholesterol, saturated phospho-lipids, and sphingolipids [16,18] The partition of GD3 into glycosphingolipid-enriched microdomains may represent a positive sorting signal for the correct sub-cellular trafficking, as previously described for other lipid-anchored proteins [42,43] Finally, it is known that dynamin participates in both clathrin-mediated endocytosis and caveolae internalization [37] Thus, the lack of effect of the dominant-negative form of dynamin on R24 internalization (Fig 6) suggests that