|
|
||||||||

*
Department of Biology, University of Oslo, Oslo, Norway; and
Laboratory of Immunogenetics, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Rockville, MD 20852
| Abstract |
|---|
|
|
|---|
| Introduction |
|---|
|
|
|---|
- and ß-chains are
important for presentation by mature recycling class II molecules
(11, 13). The various requirements for presentation of
different Ags indicate that class II peptide loading can occur in
several distinct endocytic compartments. In epithelial cells, MHC class II molecules, like MHC class I molecules (14, 15) and CD4 (16), are located at the basolateral surface (17, 18, 19, 20, 21). The basolateral membrane faces the vascular space, where the class II molecules may present Ag to emigrating T cells. Class II molecules have also been detected in intracellular vesicles in tissue epithelial cells of human and rodents in vivo (17, 18, 19, 20, 21), but little is known about the actual endocytic compartment to which class II molecules traffic and the signals involved in this transport in epithelial cells. Polarized cells may have a more complex sorting system than nonpolarized cells, as they have separate apical and basolateral plasma membrane domains, in addition to separate apical and basolateral early endosomal populations (22). However, cognate apical and basolateral pathways are also found in nonpolarized cells (23, 24).
In Madin-Darby canine kidney (MDCK) cells, apical and basolateral proteins are generally sorted in the trans-Golgi network for direct delivery to the respective plasma membrane domains (25), and signals for sorting to either the apical or the basolateral plasma membrane have been identified. Basolateral sorting signals are located in the cytoplasmic tail of several proteins and can generally be divided into two classes: 1) basolateral signals that are colinear with signals for internalization through clathrin-coated pits, either tyrosine- or leucine-based motifs; and 2) basolateral signals that are different from signals for coated pit localization, either tyrosine-dependent or -independent motifs (for reviews see Refs. 26, 27, 28). Some basolateral sorting signals can also mediate sorting in the endocytic/transcytotic pathways (26, 29).
We have previously shown that human class II molecules (HLA-DR1) and Ii are located at the basolateral surface and in endosomes when transfected stably in polarized MDCK cells (30). Here, we have extended these studies and show that an FL (single letter amino acid code) motif is required for localization of class II molecules at the basolateral plasma membrane and in early endosomes.
| Materials and Methods |
|---|
|
|
|---|
The mouse mAb L243 (American Type Culture Collection, Manassas,
VA; Ref. 31) recognizes the luminal domains of HLA-DRß
molecules. The IgG1 mAb BU45 (The Binding Site, Birmingham, U.K.; Ref.
32) and Vic Y1 (a gift from Dr. W. Knapp, Vienna, Austria)
recognize the C-terminal and the N-terminal domain of human Ii,
respectively. The mAb VP3 (a gift from Sreenivasan Ponnambalam, Dundee,
Scotland) recognizes the luminal domain of CD8. The cells were labeled
for immunofluorescence with FITC-conjugated goat anti-mouse (G
M)
IgG (Dianova, Hamburg, Germany) Abs. The cells were labeled for
immuno-electron microscopy (EM) with 15-nm and/or 20-nm gold particles
conjugated to G
M-IgG (G
M15/G
M20) Abs (British BioCell
International Cardiff, U.K.).
Cells and viruses
MDCK cells (strain II) were grown in DMEM (BioWhittaker,
Walkersville, MD) supplemented with 10% FCS (Life Technologies,
Rockville, MD). Cells (3.3 x
105/cm2) were cultured on
Transwell polycarbonate filter units with a pore size of 0.4 µm
(Costar, Cambridge, MA) for 45 days to form tight monolayers before
the experiments. Fresh medium was added every day after plating onto
filters. Transepithelial resistance was measured using a Millipore ERS
apparatus (Millipore, Bedford, MA). All cell lines exhibited resistance
of more than 200
cm2. The relative surface
area of the apical and the basolateral domains in MDCK II cells grown
under the present conditions has been measured to be 1.1
(apical/basolateral; Ref. 33).
The generation and maintenance of the HLA-DR1-restricted T cell clones C1.6, specific for the influenza virus A hemaglutinin H3 (peptide 307318), and the clone C3.5, specific for the influenza virus A matrix protein M1 (peptide 1831), have been described (34). Influenza virus A, strain X-31 (a gift from J. Yewdell, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD) was purified as described (10) and inactivated by exposure to 1200 µW/cm2 of UV light for 15 min on ice.
Plasmid constructions
MDCK cells stably expressing human MHC class II molecules,
HLA-DR1
- and ß-chain, were made as described (30).
cDNA-encoding truncated HLA-DR molecules, 
and
ß, lacking
the last 10 or 13 aa of the cytoplasmic tail, respectively, in the
vector CDM8 (11) were used. Point mutations in the
cytoplasmic tail of the ß-chain were introduced by site-directed
mutagenesis in a single-stranded M13 mp19 vector by the method of
Kunkel (35). The cytoplasmic tail region used for
mutagenesis was controlled by DNA sequencing. The mutant cDNA (DR
ßFL-AA) was then subcloned into the CDM8
vector. The CDM8-cDNA was cotransfected with the RSV.5(neo) vector
(36) for cell selection. Cells expressing full-length or
truncated HLA-DR molecules were cotransfected with cDNA encoding the
p33 form of human Ii (p33Ii) in the expression vector pMEP4 (Hygro;
Invitrogen, San Diego, CA).
Transfection and clone selection
MDCK cells were stably transfected by the DNA-calcium phosphate procedure as described (37). Clones expressing the neomycin or hygromycin resistance markers were selected in the presence of G418 (geneticin; 0.5 mg/ml active weight; Sigma, St. Louis, MO) or hygromycin B (0.3 mg/ml; Sigma) in the culture medium, respectively. Stably transfected cells were cloned by isolating resistant colonies using cloning cylinders. Colonies expressing the proteins of interest were identified by immunofluorescence, as described. To screen for expression of Ii, stable hygromycin-resistant colonies were induced with 10 µM CdCl2 for 16 h before the assays. The cell line used for transfection showed correct polarity, as confirmed by measuring methionine uptake (38) and secretion of the endogenous glycoprotein complex gp80 (Ref. 39 and data not shown).
Binding of radio-iodinated Abs
The mouse mAbs BU45 and L243 were purified from mouse ascites fluid by precipitation with sodium sulfate (0.18 g/ml Na2SO4) and purification with a protein A-Sepharose (Pharmacia, Piscataway, NJ) column at pH 8.0. The Abs were >90% pure IgG as determined by SDS-PAGE. Purified Ab molecules were labeled with Na125I using chloramine-T-catalyzed iodination and were purified by gel-filtration chromatography on a Sephadex column (Pharmacia). The spec. act. of the Abs and the amount of acid-soluble and -precipitable material was determined by TCA precipitation and counting in a Cobra Auto-Gamma counter (Downers Grove, IL). The amount of soluble radioactivity was generally <2%.
Binding assays were all performed on ice at 4°C. Cells were cooled on ice, and 125I-Ab, diluted in DMEM containing 1% FCS and 10 mM HEPES, was added to either the apical or the basolateral side of the monolayer for 2 h. Integrity of the monolayers was assessed by measuring the radioactivity in the medium of both chambers, and generally <0.1% of the added radioactivity diffused across the monolayer. Unbound Ab was removed by extensive washing (five times) with PBS supplemented with 1 mM CaCl2, 1 mM MgCl2, and 2% FCS. The filters were then excised, and bound radioactivity was measured in a gamma counter. Nonspecific binding was determined by quenching binding of the iodinated Ab with a 100-fold excess of nonlabeled Ab. These values, which generally represented <10% of the total binding, were subtracted from the total bound radioactivity to give the specific binding to expressing cells. Total binding in cpm x 103 ranged from 100 to 200 for HLA-DR molecules and from 5 to 20 for Ii. All given values represent the mean values ± SEM derived from at least three identical experimental setups, each performed in duplicates.
Immunofluorescence microscopy
Stably transfected MDCK cells were grown on polycarbonate
filters and processed for immunofluorescence by a variation of the
method described by Berod (40). Briefly, plasma membrane
staining was visualized by adding specific mAbs, diluted in DMEM/1%
FCS/10 mM HEPES, to the apical or the basolateral side of the monolayer
for 1 h at 4°C. The cells were then washed extensively with PBS
supplemented with 1 mM CaCl2, 1 mM
MgCl2, and 2% FCS and fixed by the pH shift
paraformaldehyde fixation procedure. The filters were then excised, and
bound IgG was visualized by incubation with an FITC-conjugated
G
M-IgG. To study internalization of the transfected proteins, cell
monolayers were incubated with Abs on the apical or the basolateral
side for 30 min at 37°C, before fixation, permeabilization by Triton
X-100, and staining with FITC-conjugated G
M-IgG. To visualize total
protein, cell monolayers were fixed and permeabilized before labeling
with primary and secondary Abs for 1 h each at 37°C in a
humidified chamber. Fluorescence was detected, and images were acquired
by a Leica TCS-NT digital scanning confocal microscope equipped with a
60/1.2 water immersion objective (Leica, Heidelberg, Germany). Images
were averaged four times during acquisition to reduce background noise
and processed for presentation with Adobe Photoshop (Adobe Systems,
Mountain View, CA).
Cytotoxic T cell assays
MDCK cells (2 x 105) were labeled for 15 h with sodium [51Cr] chromate (100 µCi in a six-well plate) in 1 ml DMEM supplemented with 2 mM glutamine, 10 mM HEPES, pH 7.4, 12 µg/ml gentamicin, and 10% heat-inactivated FCS. The adherent cells were then incubated for 4 h with either 10 µM synthetic peptide (H3 307318 or M1 1831) or with different dilutions of UV-inactivated influenza virus in 1 ml of DMEM containing 5% heat-inactivated FCS. The cells were then trypsinized, washed, counted, and plated at 5 x 103 cells in V-bottom 96-well plates containing the effector cells C1.6 (H3 specific) or C3.5 (M1 specific). After 4 h, supernatants were harvested and counted. The maximum lysis was measured by lysis of target cells with 2% Triton X-100. All cytotoxic assays included a titration of E:T cells, and all experiments were performed at least twice. The data are presented as percentage of maximal lysis after subtraction of the spontaneous lysis.
Electron Microscopy (EM)
To identify endosomal compartments, cell monolayers were
incubated with 5-nm and 10-nm colloidal gold particles (Zymed, San
Francisco, CA) coated with BSA (41) in the apical or the
basolateral medium, respectively, for 1 h at 37°C (procedure A).
Apical early endosomes (AP-EE) and basolateral early endosomes (BL-EE)
were defined as compartments containing only 5-nm or 10-nm gold
particles, respectively, whereas late endosomes (LE-1h) were defined as
compartments containing both endocytosed markers. As the late endosomal
fraction was morphologically heterogeneous, a different assay was used
to separate different late endosomal populations. The 10-nm gold
particles were internalized from the basolateral side for 3 h,
followed by an 18-h chase and a subsequent 1-h apical uptake of 5-nm
colloidal gold particles (procedure B). After washing in PBS, the cells
were fixed by immersing the filters in Sörensens phosphate
buffer, pH 7.4, containing 4% paraformaldehyde together with 0.1%
glutaraldehyde for 1 h at room temperature. After fixation, the
filters were incubated in 2.3 M sucrose for 1 h at room
temperature, and then cut into triangular segments and mounted on
silver pins perpendicular to the plane of sectioning. Mounted filters
were frozen and stored in liquid N2. The
specimens were sectioned on a Reichert Ultracut S ultramicrotome
(Vienna, Austria) with a Reichert FCS cryo attachment using Drukker
International Diamond knifes (Cuijk, The Netherlands). Single and
double immunocytochemical labeling of thawed cryosections was performed
mainly as described (42) using mouse mAbs followed by
15-nm and/or 20-nm gold particles coated with G
M-IgG
(G
M15/G
M20) Abs (British BioCell International). The sections
were examined on a JEOL 100CX and JEOL 1200EX transmission electron
microscope (Tokyo, Japan).
To estimate the distribution of a transfected protein in the endocytic pathway, the number of positively labeled compartments with a defined character was related to the total number of positively labeled compartments. The fraction of a defined endosomal population containing the transfected protein was estimated by relating the number of positively labeled compartments to the total number of endosomes having the defined character. Quantitation was performed on randomly acquired micrographs.
| Results |
|---|
|
|
|---|
- and the ß-chain
We have found that human MHC class II molecules (HLA-DR) are
located at the basolateral plasma membrane and in vesicles both when
expressed alone and together with human Ii in stably transfected MDCK
cells (30). The class II
and ß cytoplasmic tails
have been shown to jointly contribute a signal for internalization
(11), and to study whether the cytoplasmic tails were
equally important for basolateral targeting, stably transfected MDCK
cells expressing class II molecules having either one or both
cytoplasmic tails truncated (outlined in Fig. 1
) were eventually stably
supertransfected with Ii. Cells expressing truncated class II molecules
alone or together with Ii were analyzed to determine the apical vs the
basolateral cell surface distribution of
ß and Ii.
125I-labeled anti-Ii (BU45) or
anti-HLA-DR (L243) mAbs were added to the apical or the basolateral
medium of the cell monolayers at 4°C. After removal of unbound Ab,
filters were excised and bound radioactivity was determined in a gamma
counter. The percentage specific binding to either the apical or the
basolateral surface is shown in Table I
.
Full-length class II molecules were predominantly located basolaterally
(80%), whereas class II molecules having either tail truncated were
distributed in a nonpolarized fashion (
40% apical and 60%
basolateral), indicating that both
and ß cytoplasmic tails were
required for efficient class II basolateral targeting. Ii coexpression
had little or no effect on the steady-state polarized surface
distribution of wild-type or truncated class II molecules, although
>90% of surface-expressed Ii was detected basolaterally in these
cells (Table I
). This was expected because class II molecules in
complex with Ii only account for a few percent of total class II at the
plasma membrane at steady state (43). Together, these data
show that basolateral distribution of class II molecules depended on
the cytoplasmic tails of the
- and ß-chains but were independent
of Ii coexpression.
|
|
Efficient basolateral sorting of several transmembrane proteins
depends on sorting signals in their cytoplasmic domains. These signals
can be colinear with signals for internalization through
clathrin-coated pits and often depend on a critical tyrosine residue or
on two hydrophobic amino acids, one being a leucine (for reviews see
Refs. 26, 27, 28). The cytoplasmic tail of the DR1 ß-chain
contains a putative leucine motif (FL), which is conserved as LL or FL
in several species (13). To analyze whether the FL
residues were involved in internalization and basolateral targeting of
class II molecules, these residues were mutated to alanines (Fig. 1
)
and the mutant ß-chain (
ßFL-AA) was
cotransfected with wild-type
-chain. The polarized plasma membrane
distribution of
ßFL-AA was analyzed by
binding of 125I-L243 to the apical or the
basolateral side of the cell monolayer at 4°C.
ßFL-AA molecules were located at both apical
and basolateral plasma membrane domains in a nonpolarized fashion (50%
apical, 50% basolateral) (Table I
). Thus, the FL residues in the
ß-chain cytoplasmic tail were essential for efficient basolateral
distribution of the class II molecules.
The polarized surface distribution of wild-type and mutant class II
molecules found by binding of iodinated Abs on ice was confirmed by
confocal immunofluorescence microscopy (Fig. 2
A) and immuno-EM (Fig. 3
). Both wild-type and mutant class II
molecules were also detected in vesicles (micrographs not shown). To
determine whether vesicular class II localization was due to
internalization from the plasma membrane, cells expressing
ß or
ßFL-AA molecules were incubated with L243 at
the apical or the basolateral side of the cell monolayer for 30 min at
37°C. Vesicular staining was seen when Abs were added at the
basolateral, but not at the apical, surface of cells expressing
ß
(Fig. 2
B, upper panel). No vesicular localization
was observed upon the addition of Ab to cells expressing
ßFL-AA molecules (Fig. 2
B,
lower panel) or truncated class II molecules (micrographs
not shown). Thus, the FL motif in the ß-chain tail is required for
internalization of class II molecules at the basolateral plasma
membrane domain.
|
|
To further identify the intracellular localization of wild-type
and mutant class II molecules, cells were grown on filter supports and
processed for immuno-EM analysis. MDCK cells have been found to possess
separate apical and basolateral early endosomes, but common late
endosomes (22). To identify the different endosomal
compartments, cell monolayers were incubated with 5-nm and 10-nm
BSA-coated colloidal gold particles in the apical or the basolateral
medium, respectively, for 1 h at 37°C (procedure A). Apical
early endosomes (AP-EE) and basolateral early endosomes (BL-EE) were
defined as compartments containing only 5-nm or 10-nm gold particles,
respectively, and late endosomes-1h (LE-1h) as compartments containing
both endocytosed tracers. The compartments defined as LE-1h were
morphologically heterogenous and may also include a common recycling
compartment and transcytotic compartments. Sections were labeled with
different Abs against class II molecules (anti-DRß or L243)
and/or Ii (VicY1 or BU45). Both Abs used against class II and Ii gave
similar results. The endosomal distribution of class II and Ii was then
estimated by relating the number of positively labeled compartments
with a defined character to the total number of positively labeled
compartments, as shown in Table II
.
|
ß) or together
with Ii, 2025% of the class II-positive compartments were BL-EE and
correspondingly 7580% LE-1h (Table II
ßIi cells. No class II or Ii was detected in AP-EE. The class II
expression levels in these cells were similar as judged by
immunoprecipitation analysis (data not shown). Moreover, the fraction
of class II-positive BL-EE and LE-1h compared with the total number of
BL-EE and LE-1h was similar in
ß and
ßIi cells (data not
shown). Thus, the class II molecules were mainly located in
compartments accessible to both apical and basolateral tracers, both
when expressed alone and together with Ii. Class II and Ii molecules
were also located in BL-EE, probably due to internalization from the
basolateral plasma membrane.
In contrast, point-mutated (
ßFL-AA) or
truncated class II molecules (

ß) were only detected in LE-1h
and neither in AP-EE or BL-EE, both when expressed alone and together
with Ii (Table II
). Moreover, in 

ßIi cells 97% of the
Ii-positive compartments were LE-1h and only 3% BL-EE (Table II
),
suggesting that signals in the class II cytoplasmic tails influenced
the intracellular sorting of Ii. The fraction of class II-positive
LE-1h compared with the total number of LE-1h was similar in cells
expressing wild-type or mutant class II molecules (data not shown).
These data indicate that transport of class II molecules to BL-EE, but
not to late endosomes, requires the FL motif in the ß-chain
tail.
Class II molecules are localized in multivesicular late endosomes independently of the cytoplasmic tails
The late endosomal population (LE-1h) defined above was
morphologically heterogeneous. To further elucidate the precise
intracellular class II localization, we used a different assay to
separate distinct late endocytic populations. Cell monolayers were
incubated with 10-nm BSA-gold particles at the basolateral side for
3 h at 37°C followed by an 18-h chase and a subsequent 1-h
apical uptake of 5-nm BSA-gold particles (procedure B). As found by
procedure A, 5-nm BSA-gold particles apically internalized for 1 h
localized to AP-EE as well as more complex compartments corresponding
to LE-1h. Using procedure B, the late endosomes could be separated into
at least two different subpopulations: multivesicular bodies containing
5-nm BSA-gold particles only (mvb-1h) located apically to the nucleus,
and compartments containing both 5-nm and 10-nm BSA-gold particles
(LE-ON). The morphology of the LE-ON compartments varied from
multivesicular to more multilamellar, and the LE-ON contained more
endocytosed BSA-gold particles (both 5 and 10 nm) compared with mvb-1h.
Compartments containing 10-nm BSA-gold particles only resembled
typically electron-dense lysosomes. Sections were labeled with Abs
against class II or Ii, and the endosomal distribution of class II and
Ii was estimated as described above. In cells expressing wild-type
(
ß) (Fig. 4
B),
point-mutated (
ßFL-AA), or truncated
(

ß) class II molecules alone,
70% of the class
II-positive compartments were mvb-1h and
30% were LE-ON, with
little or no class II detected in lysosomes (Table II
). In cells
coexpressing Ii and wild-type (Fig. 4
C), point-mutated, or
truncated (Fig. 4
D) class II molecules,
90% of the class
II-positive vesicles were mvb-1h and <10% LE-ON (Table II
). In these
cells, 8090% of the Ii-positive compartments were defined as mvb-1h.
A corresponding shift in class II localization from LE-ON to mvb-1h
upon Ii coexpression was also observed when the number of class
II-positive LE-ON/mvb-1h was compared with the total number of
LE-ON/mvb-1h in these cells (data not shown). Thus, the major fraction
of both wild-type and mutant class II molecules was located in
multivesicular compartments, which were reached by an apical endocytic
marker within 1 h. The efficiency of class II localization to
these compartments was increased upon Ii coexpression, which could be
caused by Ii-induced retention of the gold particles in early endocytic
compartments (44), although we cannot exclude that Ii may
regulate endosomal distribution of class II molecules in a more
specific way.
|
To further characterize the class II-positive compartments, MDCK cells coexpressing wild-type class II molecules and Ii were transfected stably with the fusion protein CD8-DMß, containing the cytoplasmic tail of the HLA-DMß-chain fused to the transmembrane and luminal domains of the plasma membrane resident protein CD8. CD8-DMß has been shown to colocalize with HLA-DM molecules in HeLa cells (45). Polarized cells expressing class II, CD8-DMß, and Ii were processed for immuno-EM analyses by uptake of endocytic markers as described in procedure B. The extent of colocalization of class II, Ii, and CD8-DMß was determined by indirect sequential double labeling.
We found extensive colocalization of class II and Ii (Fig. 4
E), of class II and CD8-DMß, and of Ii and CD8-DMß
molecules (micrographs not shown) in mvb-1h. Some colocalization was
also detected in LE-ON, but not in AP-EE or BL-EE. A different
intracellular distribution of intact HLA-DM cannot be excluded due to
the protease sensitivity of CD8-DMß as opposed to HLA-DM. A
tyrosine-based signal in the DMß cytoplasmic tail has been identified
as essential for endosomal targeting of HLA-DM molecules
(45, 46, 47). CD8-DMßY-A molecules,
having the tyrosine mutated to alanine, were not detected in mvb-1h,
although some labeling of LE-ON has been
observed,6 indicating
that this is a signal for sorting to the mvb-1h compartments. The
specific localization of class II, Ii, and HLA-DM molecules, all
components required for class II peptide loading, in the mvb-1h suggest
that these structures may be a class II peptide loading
compartment.
Ag presentation by HLA-DR1-transfected MDCK cells
As the majority of class II molecules were found in late multivesicular endocytic compartments both when expressed alone and together with Ii, we were interested in whether MDCK cells expressing HLA-DR1 molecules alone were able to process and present the hemaglutinin H3 and the matrix protein M1 epitope of influenza virus to HLA-DR1-restricted T cells. Previous studies have shown that the H3 epitope can bind to recycling class II molecules in early endocytic compartments, whereas binding of the M1 epitope occurs in later endocytic compartments and requires newly synthesized class II molecules and Ii (11, 12).
Nonpolarized MDCK cells expressing
ß were labeled with
51Cr, pulsed with H3 or M1 synthetic peptides or
UV-inactivated influenza virus (UV-flu) particles, and tested for lysis
by the H3- or M1-specific cytotoxic T cell clones, C1.6 and C3.5,
respectively. As shown in Fig. 5
a, the synthetic H3 peptide
and the H3 epitope of processed influenza virus were presented
efficiently to C1.6. The cells were also able to present synthetic M1
peptide and the M1 epitope of processed UV-flu particles to C3.5 T
cells (Fig. 5
b). Together, these data show that MDCK cells
expressing DR1-molecules were able to process and present influenza
virus Ags and that both the H3 and the M1 epitope are presented
independently of Ii coexpression. These results are consistent with our
EM data, showing that class II molecules are present in typical peptide
loading compartments independent of Ii or intact cytoplasmic tails.
|
| Discussion |
|---|
|
|
|---|
We found that class II molecules were located on the basolateral plasma
membrane, in BL-EE, and in multivesicular compartments that were
accessible to both apical and basolateral endocytic markers, but not at
the apical plasma membrane or in AP-EE. Hershberg et al.
(48) found that class II-restricted peptides were only
presented when fed basolaterally. Our data are in agreement with this
observation if HLA-DR were loaded at the plasma membrane or in early
(recycling?) basolateral endosomes. The FL residues in the cytoplasmic
tail of the HLA-DR1 ß-chain were required for efficient basolateral
sorting and internalization at the basolateral plasma membrane and thus
for localization in BL-EE. A dihydrophobic motif (LL or FL) is found in
the ß-chain cytoplasmic tail of several species (13),
indicating an important role of this evolutionary conserved motif in
the function of class II molecules. The
-chain cytoplasmic tail was
also required for efficient basolateral transport of DR1 molecules, as
truncation of either tail lead to nonpolarized sorting of the class II
molecules. The
-chain tail may contribute to correct folding of the
FL motif to be recognized by a sorting machinery or the
-chain may
contain sorting information. These results are consistent with a
previous study (11) where we found that both the class II
and ß cytoplasmic tails were required for internalization of
HLA-DR1 from the plasma membrane. Moreover, Zhong et al.
(13) identified a dileucine motif in the cytoplasmic tail
of the mouse Ak class II ß-chain, corresponding
to the FL residues in DRß, as essential for Ag presentation by
recycling class II molecules. Several studies have shown that tailless
and full-length class II molecules differed in their Ag presenting
capacity (10, 11, 49, 50, 51), and truncation of the ß-chain
were found to reduce the level of class II-associated invariant chain
peptide complexes (52), possibly caused by different
intracellular routing of wild-type and truncated class II
molecules.
The cytoplasmic tail of Ii contains two leucine-based basolateral sorting signals, which also function as internalization signals (30). Ii coexpression only marginally increased the basolateral distribution of wild-type and truncated class II molecules, which is not surprising, considering that only a few percent of the surface class II molecules are in complex with Ii and that this fraction is efficiently internalized (43). Thus, the main role of Ii may be to target newly synthesized class II molecules to endosomes. Basolateral sorting signals within the class II cytoplasmic tails may then be important for transport of these molecules to the basolateral surface after dissociation from Ii in endosomes. Together, our data suggest that intracellular transport of class II-Ii is directed by the function of several different leucine-based signals. The amino acid context requirements for motifs involved in both basolateral sorting and internalization may be identical or differ (27, 28, 53), and it is tempting to speculate that the motifs have different affinities for components of the sorting machinery and may even work differently at different sorting stations, resulting in a fine tuning of the final routing and destination of the class II-Ii complex within the cell.
Class II-Ii complexes can be transported to the endocytic pathway
directly from the trans-Golgi network (54, 55) or via the
plasma membrane (6, 43). Class II molecules can also enter
endosomes in the absence of Ii (for reviews see Refs. 8, 56, 57). In polarized MDCK cells, class II molecules were detected
both in BL-EE and in late endosomes, containing internalized apical and
basolateral markers, both in the presence and the absence of Ii.
Localization to BL-EE, but not to late endosomes, required both the
- and the ß-chain cytoplasmic tails, including the FL motif in the
ß tail. Interestingly, less Ii-positive BL-EE were observed in cells
expressing truncated
ß, suggesting that information in the class
II cytoplasmic tails maybe influenced the initial sorting of Ii before
degradation. Ii-independent localization of class II molecules to late
endosomes correlates with the data obtained in the cytotoxic T cell
assay showing that both the H3 and the M1 epitope of inactivated
influenza virus can be presented by cells expressing DR1 molecules.
Previous studies in human fibroblasts demonstrated that presentation of
the M1 epitope requires Ii-dependent sorting of newly synthesized class
II molecules to late endocytic compartments, whereas the H3 epitope
could be presented by class II molecules recycling from the plasma
membrane into earlier endosomal loading compartments (11, 12). We have used the same HLA-DR1, Ag, and T cells as in Pinet
et al. (11, 12), so the discrepancy is most likely due to
cell type-specific differences in protein sorting and/or Ag
processing.
The class II-positive late endosome population was morphologically heterogeneous, and we distinguished between two different class II-positive LE-1h populations: multivesicular bodies located apically of the nucleus which contained 1-h apically endocytosed marker (mvb-1h), and more complex multivesicular and multilamellar vesicles containing both 1-h apical marker and basolateral marker internalized for 3 h with a subsequent 18-h chase (LE-ON). The mvb-1h contained less endocytosed marker and was probably earlier in the endocytic pathway than the LE-ON. The majority of class II-positive vesicles in cells expressing wild-type class Ii molecules were mvb-1h. Interestingly, also in cells expressing mutant class II molecules, most class II-positive late endosomes were defined as mvb-1h. As mutant class II molecules were not detected in early endosomes, this indicates that newly synthesized class II molecules may be sorted directly to the mvb-1h compartments independently of information within their tails. This may be due to the luminal 8082 aa segment of the class II ß-chain, reported to control late events in the intracellular sorting of class II molecules (58, 59). A different explanation may be that a slow delivery of mutant class II molecules to mvb-1h through the endocytic pathway by normal plasma membrane turnover leads to accumulation in these compartments as the class II molecules are relatively resistant to degradation. We could detect typical mvb-1h also in cells lacking class II molecules, indicating that these structures were not induced by the class II expression. The mvb-1h may be apical recycling compartments, previously described to contain other basolateral proteins as the transferrin receptor and transcytosed IgA (60). These data are in line with other studies showing that APC harbor a major pool of their intracellular class II molecules in special endocytic compartments, termed MHC class II compartments (MIICs) (61) or class II containing vesicles (62). Both multivesicular and multilamellar MIICs have been described (61), where the multivesicular MIICs, thought to represent the main entry site of newly synthesized class II molecules in the endocytic pathway, represented an earlier endocytic compartment than the multilamellar MIICs (63).
Several studies have shown that MIICs contain high levels of HLA-DM molecules (64, 65, 66), suggesting that these compartments are peptide loading compartments. In our study, most Ii-positive vesicles were mvb-1h, and the fraction of class II-positive mvb-1h were increased upon Ii coexpression, indicating that Ii facilitated transport to or retention in mvb-1h. Moreover, we found that the fusion protein CD8-DMß, containing the cytoplasmic tail of HLA-DMß fused to CD8, was sorted to the same mvb-1h as class II and Ii, whereas a fusion protein having the tyrosine-sorting motif of the DMß tail mutated to alanine was excluded from the mvb-1h. This is in line with previous data from several groups (45, 46, 47) who found this tyrosine signal to be required for sorting of HLA-DM molecules to MIIC. Localization of class II, Ii, and HLA-DM molecules in compartments that are accessible to both apical and basolateral endocytic markers is consistent with a specific function of such compartments in epithelial Ag presentation. However, the steady-state distribution of HLA-DM may be different from the fusion protein, as CD8-DMß is protease sensitive whereas intact HLA-DM is long lived. The localization of class II molecules expressed in the absence of Ii and HLA-DM to similar peptide loading compartments is in line with the studies of Hershberg et al. (48), who showed that an HLA-DR-positive intestinal epithelial cell line may present apically internalized Ag in the absence of detectable Ii and HLA-DM. Thus, the specific localization of class II molecules in compartments that are accessible for endocytic markers internalized both from the apical and the basolateral side of the epithelial cell monolayer makes APC able to survey both external environments for foreign intruders.
| Acknowledgments |
|---|
| Footnotes |
|---|
2 Current address: Department of Biochemistry, The Norwegian Radium Hospital, Montebello, N-0310 Oslo, Norway. ![]()
3 Current address: Institute of Pathology, The National Hospital, N-0027 Oslo, Norway. ![]()
4 Address correspondence and reprint requests to Dr. Oddmund Bakke, P.O.Box 1050, Blindern, 0316 Oslo, Norway. E-mail address: ![]()
5 Abbreviations used in this paper: Ii, invariant chain; G
M, goat anti-mouse; MDCK, Madin-Darby canine kidney; EM, electron microscopy; AP-EE, apical early endosomes; BL-EE, basolateral early endosomes; LE-1h, late endosomes containing both 1 h apical and basolateral endocytosed marker; mvb, multivesicular bodies; mvb-1h, multivesicular bodies with 1 h uptake of endocytosed marker; LE-ON, late endosomes with endocytosed markers after both 1 h and overnight incubation; MIIC, MHC class II compartments; UV-flu, UV-inactivated influenza virus particles. ![]()
6 Røe, M., K. Egdalen, A. Simonsen, A. Kelly, and O. Bakke. Polarized targeting of HLA-DM by a tyrosine motif in the ß-chain. Submitted for publication. ![]()
Received for publication October 16, 1998. Accepted for publication June 16, 1999.
| References |
|---|
|
|
|---|
interferon-stimulated MHC antigens and transferrin receptors in a clonal cell line isolated from Fisher rat thyroid (FRT cells). Cell Tissue Res. 272:23.[Medline]
interferon. J. Exp. Med. 162:1645.
ß cytoplasmic domains reveal signaling-independent defects in antigen presentation. Int. Immunol. 7:665.This article has been cited by other articles:
![]() |
E. Walseng, O. Bakke, and P. A. Roche Major Histocompatibility Complex Class II-Peptide Complexes Internalize Using a Clathrin- and Dynamin-independent Endocytosis Pathway J. Biol. Chem., May 23, 2008; 283(21): 14717 - 14727. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. Beers, A. Burich, M. J. Kleijmeer, J. M. Griffith, P. Wong, and A. Y. Rudensky Cathepsin S Controls MHC Class II-Mediated Antigen Presentation by Epithelial Cells In Vivo J. Immunol., February 1, 2005; 174(3): 1205 - 1212. [Abstract] [Full Text] [PDF] |
||||
![]() |
U. Sundberg, N. Beauchemin, and B. Obrink The cytoplasmic domain of CEACAM1-L controls its lateral localization and the organization of desmosomes in polarized epithelial cells J. Cell Sci., March 1, 2004; 117(7): 1091 - 1104. [Abstract] [Full Text] [PDF] |
||||
![]() |
T. F. Gregers, B. Fleckenstein, F. Vartdal, P. Roepstorff, O. Bakke, and I. Sandlie MHC class II loading of high or low affinity peptides directed by Ii/peptide fusion constructs: implications for T cell activation Int. Immunol., November 1, 2003; 15(11): 1291 - 1299. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |