 |
INTRODUCTION |
Major histocompatibility complex
(MHC)1 class II molecules are
heterodimeric cell surface glycoproteins that present antigens to
CD4+ T cells (1). The antigenic peptide-class II complexes
are expressed on specialized antigen-presenting cells such as
macrophages and B lymphocytes. Following their synthesis, the 
and
subunits of the class II molecule associate in the endoplasmic
reticulum (ER) together with the invariant chain (Ii) (2). The latter folds in part through the groove of the class II molecule, stabilizing the heterodimer and preventing the undesirable binding of ER polypeptides (3-6). Studies using mice with inactivated Ii genes suggested that Ii is necessary for efficient exit of newly synthesized class II molecules from the ER (7, 8). However, it was later demonstrated that high levels of class II molecules reach the surface
of Ii
dendritic cells (9). Moreover, transfected cells
express a substantial amount of class II molecules at their plasma
membrane even in the absence of Ii (10). Such a phenotype probably
results from the binding of endogenous peptides or polypeptides present in the ER (6, 11) and supports the notion that occupancy of the
peptide-binding groove is sufficient to allow ER egress (12).
Another function of Ii is to direct efficiently MHC class II molecules
to the endocytic antigen-loading compartments (13-15). Two short
leucine-based sequences located in the cytoplasmic tail of Ii are
responsible for trafficking through the endocytic pathway (16, 17).
Similar motifs in many proteins are specifically recognized at the cell
surface and trans-Golgi by elements of the sorting machinery
(reviewed in Ref. 18).
Once the class II-Ii complex reaches the endosomal compartments, the
invariant chain is progressively degraded by various proteinases
depending on the cell type (19). A residual class II-associated Ii
peptide (CLIP) must be removed from the peptide-binding groove to allow
the binding of an antigen and the subsequent export of the MHC molecule
to the cell surface (20, 21). Removal of the CLIP is catalyzed by the
non-classical HLA-DM heterodimer (22-24). This intracellular chaperone
plays three critical roles in antigen presentation as it facilitates
the withdrawal of the residual CLIP peptide and stabilizes the empty
class II prior to peptide loading (25-28). This loading occurs in late
endosomal vesicles and, in particular, in the MHC class II-rich
compartments (29). In addition, it was demonstrated that HLA-DM
functions as a peptide editor limiting the repertoire of peptides bound to the class II molecule (23, 30-32). In this matter, HLA-DM favors
the loading of high affinity peptides.
The activity of HLA-DM is regulated by another non-classical class II
molecule called HLA-DO. Experiments using purified complexes demonstrated the ability of DO to inhibit DM-mediated CLIP release (33,
34). Furthermore, it was shown in transfected class II+
cells that overexpression of HLA-DO alters the peptide repertoire bound
to class II molecules and caused the accumulation of CLIP. Subsequent
studies demonstrated that HLA-DO regulates the activity of HLA-DM by
limiting the pH range and thus probably the endosomal compartments,
where HLA-DM is active (35, 36). Based on these results and on
experiments using H-2O knock-out mice, it was proposed that H-2O and
HLA-DO favor presentation of antigens internalized by membrane
immunoglobulins. This would be achieved by the selective inhibition of
H-2M and HLA-DM in endocytic compartments rich in proteins internalized
by fluid-phase endocytosis (35). However, the group of Hämmerling
(37) suggested that HLA-DO is not involved in the release of CLIP but
rather edits the repertoire of class II-bound peptides by chaperoning
empty DR molecules through its interaction with HLA-DM. In addition to
the controversial DM-related functions of HLA-DO, its role in classical
class II-negative thymic epithelial cells remains elusive (38).
In HeLa cells and in murine B lymphocytes, under the experimental
conditions tested so far, HLA-DO was shown to be totally dependent of
its association with HLA-DM to egress the ER (39). The complex then
moves through the Golgi network and traffics to endocytic compartments.
Although the cytoplasmic tail of HLA-DM contains a functional
tyrosine-based motif needed for its own targeting to lysosomal
compartments (40-43), the sorting signals responsible for the
intracellular localization of the DO-DM complex have not yet been
characterized. Indeed, we have recently proposed that the HLA-DO
chain might also contain a sorting motif to the endocytic pathway (44).
In these experiments, a mixed isotypic pair between the classical DR
chain and a recombinant DR/DO chain specifically accumulated in
lysosomal compartments of transfected HeLa cells. This phenotype was
imparted by the DO chain since the wild-type DR-DR heterodimer
could only be detected at the cell surface. In the present report, we
have used chimeric reporter molecules and site-directed mutagenesis to
determine 1) the domain localization of the DO-encoded lysosomal
sorting motif, 2) the exact nature of this sorting signal, and 3) its
role in the trafficking of the DO-DM complex.
| |
MATERIALS AND METHODS |
Plasmids
RSV.5neo DRTM (45), coding for a HLA-DR 0101 chain lacking
its cytoplasmic domain, was a gift from Dr. R. P. Sékaly. RSV.3 DR008 (46), RSV.3 DO (47), and RSV.3 DR18/DO
(44, 48) have been previously described. RSV.5gptDN1 (DO.) (49), 45.1 DO, and 45.1 DR008 (47) cDNAs were obtained from Dr. E. Long.
pBS DR008 was generated by inserting the BamHI insert of
45.1 DR008 into pBluescript KS+ vector (Stratagene). pBS
DO.9 was generated by inserting the SalI-PstI
fragment of RSV.5gptDN1 into pBluescript KS+ vector. This
was then subcloned in RSV.5 neo by digesting with SalI/XbaI to generate RSV.5neoDO. To construct
pBS DO, a BamHI-fragment from 45.1 DO was
inserted into pBluescript KS+ vector. Before cloning
HLA-DO/ gene into pBudCE4 vector (Invitrogen, CA), we integrated
a -lactamase gene, which allows for selection by resistance to
ampicillin, by inserting a 1-kilobase pair fragment from pBluescript
KS+ digested with BspHI/blunt-ended into
NheI/blunt-ended site of pBudCE4 (pBudCE4-A). The HLA-DO
cDNA was then cloned into pBudCE4-A by digesting pBS DO with
SalI-XbaI. The HLA-DO cDNA was inserted in
the same vector by cutting pBS DO.9 with BamHI and
cloning the fragment into the BglII site of the vector
(pBudCE4-A DO/). The DO cDNA in 45.1 was excised with
BamHI, and the 1.3-kilobase pair fragment was cloned in
RSV.5neo plasmid to generate RSV.5neoDO.
In order to generate the p305 DOLL242-243AA, several subcloning
steps were done to insert a unique HindIII restriction site in a 5'-position of the gene. The first step consisted in subcloning a
5' portion of the RSV.5neoDO, which contained a HindIII
site, in the RSV.3DOLL242-243AA vector by a
MluI/EcoRV digestion. From this construct, a
HindIII fragment containing DOLL242-243AA was then
subcloned into p305 expression vector from which we had previously excised a XbaI stuffer to finally yield p305DOLL242-243AA plasmid.
DO Mutagenesis and Construction of the DR18/DO
Molecules
Introduction of mutations into the DO cytoplasmic tail
cDNA sequence was performed by PCR overlap extension (50). Briefly, a 5' PCR product was generated from RSV.3 DR18/DO using
the RSV LTR primer (5' to the coding region) and a mutagenic primer
which included the desired mutation as well as a unique restriction site used to identify clones harboring the mutation
(DOY227ABspE1b, 5'-CGT CCG GAC
AGC TCC TTT CTG-3'; DOT230LPstIb, 5'-C AGA CAT CTG
CAG CCT CAC ATA-3'; DOLL242-243AAXbaIb,
5'-TGA CTG AGG GGC CGC AAC AGC TCT AGA GAC-3').
The 3' PCR product was generated using the complementary mutagenic
primers (DOY227ABspEIc, 5'-CAG AAA GGA GCT
GTC CGG ACG-3'; DOT230LPstIc,
5'-TAT GTG AGG CTG CAG ATG TCT G-3'; and
DOLL242-243AAXbaIc, 5'-GTC TCT AGA GCT GTT
GCG GCC CCT CAG TCA-3') and DR008 3' (5'-ACT
CGA TCT TTG AGA AAC ATT-3') which hybridized to the non-coding 3' end
of the cDNA. The two overlapping PCR products were mixed, and a
final PCR was performed using the flanking primers. This product was subsequently subcloned into SacI and HindIII
sites of RSV.3 DO, thereby replacing the wild-type fragment with the
PCR product containing the various mutations. The nucleotide sequence
was confirmed by DNA sequencing using T7 polymerase (Amersham Pharmacia Biotech) before the mutations were introduced into RSV.3
DR18/DO (RSV.3 DR18/DOT230L, RSV.3
DR18/DOY227A, RSV.3 DR18/DOLL242-243AA) by replacing the EcoRV-HindIII fragment with the
equivalent fragment from mutated DO cDNA in RSV.3 described
above. Underlined nucleotides in primer sequences correspond to
mutations introduced in the PCR products.
DR/DOcyto and DR/DMcyto Reporter Molecules
The cytoplasmic tail of HLA-DR was replaced by that of either
HLA-DO or HLA-DM. To construct RSV.3 DR/DOcyto chimeric reporter molecule, we used the PCR overlap extension method. A first
fragment was amplified from RSV.3 DR008 using the RSV-LTR primer and
a mutagenic primer overlapping the end of the DR transmembrane and
the N-terminal region of the HLA-DO cytoplasmic tail
(oligonucleotide DR/DOCyto222BssHIIb, 5'-ATA TCC TTT
CTG CGC GCG GAA GTA GAT GAA C-3').
A second reaction was made on RSV.3 DR18/DO using a
complementary fusion primer (DR/DOCyto222BssHIIc, 5'-G
TTC ATC TAC TTC CGC GCG CAG AAA GGA
TAT-3') and the DR008 3' primer. Following the overlap reaction, the
PCR product was subcloned into StuI and HindIII
sites of RSV3 DR008. In order to introduce mutations in the
cytoplasmic tail of the chimeric molecule, we used a similar strategy
to the one described above but used RSV.3 DR18/DO T230L
and RSV.3 DR18/DOLL242-243AA plasmids as templates for
the amplification of the DO cytoplasmic regions. Briefly, the DR
portion was amplified from pBS DR008 with the reverse primer and
DR/DOCyto222BssHIIb primer which makes the junction between DR and DO cytoplasmic regions. The mutated DO
cytoplasmic region was amplified from the RSV.3 DR18/DO
mutant with complementary fusion primer
DR/DOCyto222BssHIIc and the DR008 3' primer. The
product of the overlapping reaction was digested with StyI to generate a fragment containing the mutated DR/DOcyto fusion region and subcloned into StyI-digested pBS DR008 to
replace the wild-type sequence. The resulting construction was
sequenced, excised with BamHI, and subcloned into the SR
vector (51) which allows for selection by resistance to puromycin
(puroSR was kindly provided by Dr. François Denis). To
generate SR puro DR/DOcytoY227A, a different strategy was used. A
first PCR product was made from RSV.3 DR18/DO using the
RSV-LTR primer and the mutagenic primer (DOY227ABspE1b,
5'-CGT CCG GAC AGC TCC TTT CTG-3'). The second PCR product was done using the complementary mutagenic primer (DOY227ABspE1c, 5'-CAG AAA GGA GCT GTC
CGG ACG-3') and DR008 3'. The overlapping PCR product
was subcloned into StyI of pBS DR008 to replace the
wild-type sequence. The resulting construct was sequenced, excised with
BamHI, and cloned into the SR vector as above. In order
to generate the triple mutant DR/DOcytoLL242-243AA/T230L, we
performed two successive overlap PCRs. The first overlap PCR was done
as described above to generate a DR/DOcytoLL242-243AA PCR fragment.
By using the latter PCR product as a template, we subsequently
introduced the third mutation at position T230L. The 5' overlap
fragment was amplified with the reverse primer and the
DOT230LPstIb primer. The 3' fragment was obtained using DOT230LPstIc primer and the universal primer. Following
the overlap reaction, the PCR product was subcloned into
StyI of pBS DR008 to replace the wild-type sequence. The
resulting construct was sequenced, excised with BamHI, and
cloned into the SR vector (SR puro DR/DOcytoAA/T230L).
In order to make a fusion between the DR and DM cytoplasmic
tails, we first cloned the HLA-DM cDNA from Raji cells by PCR. Total mRNA was extracted using TRIZOL Reagent (Life Technologies, Inc.), and 10 µg was used for cDNA synthesis in the presence of the DM5'SalI (5'-CTG GAA GAG CTG GTC GAC GGG ACT G-3')
and the DM3'EcoRI (5'-GAA GTT GTA GAA TTC TGC CTC TAG-3')
oligonucleotides. The first strand cDNA was made using
Moloney murine leukemia virus reverse transcriptase (Life
Technologies, Inc.) and Taq polymerase at 42 °C for 15 min in Taq polymerase buffer (BIO/CAN, Canada). Double-stranded cDNA was then amplified by PCR for 20 cycles, digested with SalI-EcoRI, and cloned in
pBluescript KS+ (pBS 1-DM.1). The complete DNA coding
sequence was verified by sequencing, and the predicted amino acids
sequence corresponds to the published sequence (52).
In order to mutate tyrosine 230 in the DM cytoplasmic region, a
first PCR was performed on pBS 1-DM.1 using the reverse primer and a
mutagenic primer (DMY230AEco47IIIb, 5'-AGG AAG AGG AGT
AGC GCT AGA GTG GCC AGC-3'). A second fragment
was amplified from pBS 1-DM.1 using a complementary mutagenic primer
(DMY230AEco47IIIc, 5'-GCT GGC CAC TCT AGC
GCT ACT CCT CTT CCT-3') and the universal primer. The two
overlapping PCR products were mixed, and a final PCR was completed
using the flanking primers. This PCR product was subsequently digested
with SacI and HindIII, cloned into pBS 1-DM.1,
and sequenced (pBSDMY230A).
Fusion of the DM cytoplasmic tail to the DR transmembrane domain
was performed by PCR. A first fragment was amplified from pBS DR008
using the reverse primer and a fusion primer (DR/DMcytob, 5'-GA
GTG GCC AGC TCT GAA GTA GAT GAA CAG-3'), which makes the junction
between the DR and DM cytoplasmic regions. The DM cytoplasmic
region was amplified from pBS 1-DM.1 with a complementary fusion
primer (DR/DMcytoc, 5'-CTG TTC ATC TAC TTC AGA GCT GGC CAC TC-3')
and the universal primer. The overlap product was digested with
StyI-XbaI to generate a fragment containing the
DR/DMcyto fusion region and subcloned into pBS DR008. The
resulting cDNA was sequenced, excised with BamHI, and
cloned into the BamHI site of SR puro vector (SR puro
DR/DMcyto). In order to generate a DM cytoplasmic tail mutated at
position Y230A and fused to the DR molecule, we used a similar
strategy to the one described above but using pBS DMY230A as
template (instead of wild-type pBS 1-DM.1) for the amplification of
the DM cytoplasmic region.
Antibodies
L243 monoclonal antibody (IgG2a) binds a specific
DR conformational determinant (53). Coupling to biotin using
biotin-7-NHS or to FITC using FLUOS was performed as suggested by the
manufacturer (Roche Molecular Biochemicals). The anti-Lamp-1 (CD107a)
monoclonal antibody H4A3 (IgG1,
) reacts with the heavy
glycosylated 110-kDa lysosomal-associated membrane protein
(Developmental Studies Hybridoma Bank, NICHD, University of Iowa, IA).
XD5.117 is an anti-DR (IgG1) monoclonal antibody (54).
DA6.147 monoclonal antibody (IgG1) is directed at the
cytoplasmic tail of the DR chain (55). The anti-DO serum was
produced in C3H mice (H-2d) by repeated
intraperitoneal injections of DAP fibroblasts transfected with DR
and DR18/DO cDNA (48). Rabbit antisera were raised against
keyhole limpet hemocyanin conjugated to peptides corresponding to the cytoplasmic tail of HLA-DO (CMGTYVSSVPR) or HLA-DM (CRAGHSSYTPLPGSNYSEGWHIS).
Cells Lines and Transfections
HeLa DR1 (DR + DR0101) and HeLa DRTM +DR cells were
kindly provided by Dr. R. P. Sékaly. HeLa
DR+DR18/DO and HeLa DRTM + DR18/DO
have been described previously (44). HeLa DM.5 is a clone from HeLa
DM+ cells (a gift from Dr. R.P. Sékaly) obtained
after transfection of the DM and DM cDNAs (22). Cells were
cultured in Dulbecco's modified Eagle's medium, 10% fetal bovine
serum (Wisent, St.-Bruno, Quebec, Canada), and appropriate selective
agents (see below). HeLa cells were co-transfected by the calcium
phosphate precipitation method as described (57) using 2 µg of
truncated DR cDNA (DRTM) along with 10 µg of the different
chain chimeras. Independently, duplicate transfections were done
using Fugene6 (Roche Molecular Biochemicals). Selective agents were
added to a final concentration of 500 µg/ml G-418 (Life Technologies,
Inc.), 400 µg/ml puromycin (Sigma), 50 units/ml hygromycin (Cederlane
Laboratories, Ontario, Canada) or 100 µg/ml ZeocinTM (Cayla,
Toulouse, France). Cells expressing the desired class II molecules were
sorted using magnetic beads (Dynal Inc.) coated with L243 monoclonal
antibody, except for HeLa DR/DMcyto which was cloned by limiting
dilutions, and for the HeLa DRTM+DR, HeLa DR/DOcyto T230L,
HeLa DR/DOcyto AA/T230L, HeLa DM.5/DO+DO, HeLa
DM.5/DO+DOAA cells which represent unsorted populations. The
nature of the different reporter molecules expressed in transfected
cells was confirmed by Western blot using antibodies specific for the
different cytoplasmic tails (data not shown).
Flow Cytometry Analysis
Cell Surface Staining--
Cells were harvested using trypsin,
washed and incubated with FITC-conjugated L243 antibody or L243-bio in
complete medium. After 45 min at 4 °C, cells were washed twice in
PBS, and L243-bio-stained cells were incubated for another 45 min at
4 °C with phycoerythrin-conjugated streptavidin (SA-PE) in PBS.
Cells were washed twice in PBS and analyzed by flow cytometry on a
FACS® caliber (Becton Dickinson). As negative control, cells were
stained using only FITC-conjugated goat anti-mouse antibody
(Cerderlane, Canada) or SA-PE (Coulter, Ontario, Canada).
Intracellular Staining--
Cells were harvested using trypsin,
rinsed in PBS, and fixed in 4% formaldehyde for 20 min at room
temperature. After two washings in PBS, cells were treated with 50 mM NH4Cl for 15 min in 0.05% saponin (Sigma)
in PBS containing 1% BSA (Bioshop, Canada). Intracellular staining was
performed using L243-FITC. After 30 min at room temperature, cells were
washed twice, fixed in 1% formaldehyde, and analyzed by flow
cytometry. As negative control, untransfected HeLa cells were stained
under the same conditions. In order to analyze by flow cytometry the
total (surface and intracellular) versus surface expression
of the various class II molecules, the profiles obtained for DR1 were
first precisely superimposed, and the other samples of transfected
cells were acquired under the same exact settings. A shift between the
two curves for a given reporter molecule indicates a cellular
redistribution. When unsorted populations were analyzed, the peaks of
negative cells were used to align the staining profiles and define the settings.
Fluorescence Microscopy
Cells were plated on coverslips in 24-well plates and cultivated
for 3 days before intracellular staining. The coverslips were rinsed in
PBS and cells were cold-fixed in pre-cooled (
80 °C)
methanol/acetone (80:20%, v/v) for 20 min at 20 °C. Subsequent manipulations were done at room temperature. After four washings of 10 min in PBS, the coverslips were immersed in blocking solution (PBS
containing 0.2% BSA and 0.2% gelatin) for 10 min. Intracellular staining was performed by using L243-bio- and Lamp-1-specific antibodies. After 1 h, cells were washed twice in PBS, 1% BSA buffer and incubated for 1 h with Texas Red-coupled streptavidin (Amersham Pharmacia Biotech) and anti-mouse IgG1 coupled to
fluorescein (PharMingen, San Diego, CA). Cells were washed twice, and
the coverslips were mounted using Gelvatol (polyvinyl alcohol, a gift from Dr. M. Desjardins).
When stained with mouse DO- and/or rabbit DM-specific antibodies,
cells were washed twice in PBS, 1% BSA buffer and incubated at room
temperature for another 20 min with a biotinylated goat anti-mouse
antibody and/or FITC-conjugated goat anti-rabbit antibody (Bio/Can
Scientific, Ontario, Canada). When necessary, cells were washed twice
and incubated with Texas Red-coupled streptavidin (Amersham Pharmacia
Biotech) for 20 min.
Cells were then analyzed by fluorescence microscopy on a Zeiss
axioscope microscope (Carl Zeiss, Germany). Photographs were taken with
a Zeiss microscope camera MC 100 on Kodak elite chrome 400 film.
Confocal laser microscopy was performed on a Zeiss LSM 410 system
equipped with a PLAN-APOCHROMAT 63× oil immersion lens and a
Ar/Kr laser.
Immunoprecipitations and Western Blotting
Cells (1.2 × 107) were trypsinized, washed in
PBS and lysed into Triton X-100 1% as described previously (58). After
centrifugation, supernatants were harvested and incubated with 20 µl
(50% beads) of CL-6B Sepharose (Amersham Pharmacia Biotech) for 1 h at 4 °C with agitation. Cell lysates were transferred into a new
microtube containing 20 µl (50% beads) of protein A-Sepharose
(Amersham Pharmacia Biotech) coupled to HLA-DO antibody. Samples
were agitated overnight at 4 °C. Beads were centrifuged, washed four
times with lysate buffer, and resuspended in Laemmli buffer. Samples
were boiled and loaded on 10% SDS-polyacrylamide gels. Proteins were transferred to Hybond ECL membranes (Amersham Pharmacia Biotech) and
blotted with the rabbit anti-DO serum overnight at 4 °C with agitation. Secondary antibody (peroxidase-coupled goat anti-rabbit; BIO/CAN Scientific, Ontario, Canada) was used at 1:1000 dilution for
2 h at room temperature. After washings, the membrane was developed by chemiluminescence (Roche Molecular Biochemicals) on Kodak
XAR-5 films.
| |
RESULTS |
The Cytoplasmic Tail of HLA-DO Encodes a Lysosomal Sorting
Signal--
We recently described the lysosomal and cell surface
localization of a mixed isotypic pair between DR and a chimeric
recombinant DR18/DO chain (44, 48). As wild-type
DR/DR (DR1) molecules did not accumulate in intracellular
compartments under the same experimental conditions, we concluded that
a lysosomal sorting motif was encoded in the HLA-DO chain. In an
effort to localize such a signal and to evaluate the relative
importance of the cytoplasmic tail in lysosomal transport as compared
with the other domains of HLA-DO, we have designed a new reporter
molecule where the cytoplasmic tail of the DR chain was replaced by
that of DO (DR/DOcyto) (Fig. 1). To
eliminate any possible contribution from the cytoplasmic tail of the
DR chain partner, HeLa cells were stably transfected with the
DR/DOcyto construction together with a HLA-DR chain cDNA
encoding a stop codon immediately after the transmembrane domain
(DRTM). Since this chimeric heterodimer can be detected at the
plasma membrane, cells were incubated with the DR chain-specific
L243 antibody and sorted using magnetic beads. Surface expression was
compared with the wild-type DR or DR18/DO-expressing
HeLa cells (Fig. 2) (44). The structural integrity of the DR/DOcyto reporter molecule was confirmed by its
ability to bind superantigens and various DR-specific antibodies (data
not shown).
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 1.
Chimeric molecules and mutants.
A, schematic representation of the different molecules used
in this study. These reporter molecules consist of heterodimers between
different and chains. Cytoplasmic domain truncations and
recombinant molecules are described under "Materials and Methods."
B, amino acid sequence of the chimeric molecules and derived
mutants. The asterisk indicates the end of the protein. The
first amino acid in DO corresponds to the first amino acid of the
cytoplasmic tail. Dashed lines represent identity in the
amino acid sequence. Squares represent a potential or
documented sorting motif. Bold amino acids identify
introduced mutations.
|
|
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 2.
Cell surface expression of the class II
chimeric molecules in transfected HeLa cells. Flow cytometry
analysis was performed using FITC-labeled DR chain-specific L243
antibody (solid line). Dotted lines represent the
control fluorescence of cells incubated with a FITC-conjugated goat
anti-mouse serum.
|
|
The intracellular expression pattern of expression of the DR/DOcyto
molecule was analyzed by confocal microscopy using L243. This
monoclonal antibody recognizes a conformational epitope on the DR
chain apparent only upon association with the chain and maturation
through the Golgi apparatus (3). Fig.
3A shows that the class II
staining was characterized by the presence of large, well defined
perinuclear vesicles reminiscent of endosomal/lysosomal compartments.
The endocytic nature of these vesicles was confirmed using a
Lamp-1-specific monoclonal antibody (H4A3) as described (Fig.
3B) (44). Fig. 3C shows a perfect co-localization
between the Lamp+ and class II+ vesicles, as
was reported previously for cells expressing the mixed isotypic pair
between DRTM and DR18/DO (Fig. 3, D-F
(44)). The concomitant cell surface expression and intracellular
sorting of these molecules is reminiscent of the Ii expression pattern in B lymphocytes (see "Discussion"). On the other hand,
DR1-transfected HeLa cells displayed the highest level of surface
expression as determined by flow cytometry (Fig. 2) but failed to show
strong intracellular accumulation (Fig. 3G). As described
previously by other groups (14, 44, 59, 60), the intracellular class II
staining in Ii HeLa cells was rather diffuse (Fig.
3G) with, occasionally, a few cells showing small faint
vesicles. No co-localization of DR1 and Lamp-1 was detected by confocal
microscopy under the exact same settings used above for the analysis of
reporter molecules (Fig. 3, H and I). Moreover,
it is interesting to note the absence of enlarged, swelled
Lamp-1+ lysosomal compartments in cells expressing the
wild-type DR1 molecule as compared with those containing the DO
cytoplasmic tail (Fig. 3, H and E). This
important observation was consistent through all the independent
transfections generated and supports the conclusions that 1) the
wild-type DR1 does not accumulate in large amounts in lysosomes and 2)
the DO tail is responsible for the redistribution and accumulation
of the reporter molecules to the endocytic pathway, affecting the
morphology of the compartments.
View larger version (42K):
[in this window]
[in a new window]
|
Fig. 3.
The cytoplasmic tail of
HLA-DO is responsible for the lysosomal
localization of reporter molecules. Permeabilized cells expressing
the DR/DOBcyto (A-C) or the DR18/DO
(D-F) molecules paired with the truncated DR chain
(DRTM) devoid of its cytoplasmic tail were stained intracellularly
to analyze lysosomal sorting. A, D, and G show
staining with the class II-specific L243 antibody (red), and
B, E, and H highlight the lysosomal
staining using the H4A3 antibody specific for Lamp-1
(green). Co-localization of the two antibodies is shown in
C, F, and I (yellow). Control cells
express wild-type DR+DR (G-I). Immunofluorescence was
monitored by confocal microscopy.
|
|
Taken together these results show that the cytoplasmic tail of
HLA-DO is responsible and sufficient for the efficient targeting of
these reporter molecules to Lamp+ lysosomal compartments.
This was confirmed by the diffuse staining and the lack of vesicular
accumulation of a mixed isotypic heterodimer between an intact DR
chain and a truncated chimeric DR18/DO chain devoid of
its DO cytoplasmic tail (data not shown).
Tyrosine 227 in HLA-DO Is Not Involved in Lysosomal
Sorting--
The cytoplasmic tail of the DO chain contains two
putative sorting signals that have been shown in other molecules to
mediate the intracellular targeting to various cellular compartments
(Fig. 1B). The first signal is an imperfect tyrosine-based
GYVRT motif reminiscent of the GYXXL (where X is
any amino acid) motif found, for example, in LgpA protein (reviewed in
Ref. 61). It has been suggested that the glycine preceding the tyrosine
residue might mediate the direct sorting of proteins from the
trans-Golgi network to the endocytic pathway (62). Although in
HLA-DO a polar threonine is found instead of the typical hydrophobic
or aromatic residues at the end of the putative sequence, some reports
have suggested that the GY alone is sufficient to mediate intracellular
sorting events (61). The second signal is a leucine-leucine motif
analogous to those found in Ii and that mediate the transport of
associated class II molecules to the endocytic pathway (14, 16, 63). It
is not known which of these two signals in the mixed pairs is
responsible for their lysosomal distribution.
To dissect further the intracellular signals of the DO cytoplasmic
tail, we next introduced point mutations in the cDNA sequence coding for these putative motifs. First, the potentially critical tyrosine residue in the GYVRT sequence was changed to an alanine in
both the DR/DOcyto and the DR18/DO chains (Fig.
1B). HeLa cells stably transfected with these
DR/DOcytoY227A or DR18/DOY227A cDNAs together
with DRTM were sorted on magnetic beads, and flow cytometry analysis
confirmed high levels of surface expression (Fig.
4, A and C). The
confocal microscopy analysis using class II- and Lamp-specific
antibodies showed a perfect co-localization between the two markers for
both Y227A reporter molecules, containing either the cytoplasmic tail
(Fig. 5, A-C) or the almost
entire DO chain (Fig. 5, D-F). This phenotype was
similar to that of DR/DOcyto-expressing cells suggesting that
tyrosine 227 is not involved in lysosomal targeting or the generation
of enlarged vesicles.
View larger version (30K):
[in this window]
[in a new window]
|
Fig. 4.
Cell surface expression of the mutated
reporter molecules in transfected HeLa cells. HeLa cells were
transfected with DRTM and the indicated cDNAs. Positive cells
were sorted and analyzed by flow cytometry using FITC-labeled L243
antibody (filled line). Dotted lines represent
the control fluorescence of cells incubated with a FITC-conjugated goat
anti-mouse serum.
|
|
View larger version (36K):
[in this window]
[in a new window]
|
Fig. 5.
The di-leucine motif in
DO mediates lysosomal targeting.
Permeabilized cells expressing the DR/DOcytoY227A
(A-C), DR18/DOY227A (D-F),
DR/DOcytoLL242-243AA (G-I), and
DR18/DOLL242-243AA (J-L) molecules paired
with the truncated-DR chain (DRTM) were stained to analyze the
intracellular localization. A, D, G, and J show
staining using the L243 antibody (red). B, E, H,
and K represent the lysosomal staining obtained using H4A3
(green), and C, F, I, and L illustrate
the co-localization of the two proteins (yellow).
Fluorescence was monitored by confocal microscopy.
|
|
The Cytoplasmic Tail of HLA-DO Encodes a Functional LL
Signal--
We then examined the importance of the di-leucine motif in
the lysosomal sorting. The DR/DOcytoLL242-243AA cDNA (Fig.
1B) was co-transfected with DRTM, and the distribution of
this class II molecule was monitored by flow cytometry (Fig.
4B) and immunofluorescence (Fig. 5, G-I). This
molecule was expressed at high levels at the cell surface, but in
contrast to the wild-type DR/DOcyto or DR/DOcytoY227A phenotypes
observed above (Fig. 3A and 5A), confocal
microscopy revealed a diffuse intracellular labeling and the absence of
vesicular accumulation (Fig. 5G). The pattern was similar to
the one observed using the DR1-transfected HeLa cells (Fig.
3G). There was a clear loss of the enlarged perinuclear
class II+ vesicles such as those found in DR/DOcyto
cells. In addition, the Lamp-1 pattern is made principally of small
discrete vesicles scattered all over the cytoplasm. This pattern is
reminiscent of the signal obtained using DR1-expressing cells (Fig.
3H (44, 60)) and further demonstrates that the di-leucine
motif is functional and necessary for lysosomal sorting of the reporter molecule.
The critical role of the di-leucine motif was confirmed in independent
experiments using the mutated mixed pair
DRTM+DR18/DO. The di-leucine mutation
(DR18/DOLL242-243AA) (Fig. 4D) inhibited lysosomal accumulation of this reporter molecule as judged by the
diffuse intracellular staining and the lack of co-localization with the
Lamp-1 molecule (Fig. 5, J-L). Altogether, these results clearly establish that the cytoplasmic tail of HLA-DO contains a
functional di-leucine lysosomal sorting motif.
Mutation-induced Cell Surface Redistribution--
Lack of
lysosomal sorting caused by mutation of the leucine-leucine motif must
result in cell surface accumulation, and flow cytometry was used to
evaluate this redistribution within a cell population. Similar analysis
were previously reported by others (64) to highlight the plasma
membrane redistribution of class II molecules upon dendritic cell
maturation. To analyze the distribution of the reporter molecule, the
staining obtained at the cell surface was compared with the total
amount of class II molecules as determined by staining simultaneously
the cell surface and the intracellular content following membrane
permeabilization with saponin. To validate this assay in our system, we
first used DR molecules fused with the well characterized DM and
DMY230A cytoplasmic tails (Fig. 1, A and B).
Indeed, it has been shown previously that the tyrosine motif encoded in
the cytoplasmic tail of HLA-DM mediates trafficking to lysosomal
compartments and that disruption of this tyrosine motif results in
enhanced cell surface expression (40-42). In accordance with these
previously published data, Fig. 6 shows
that the mutation of the tyrosine 230 to an alanine in DR/DMcyto
(DR/DMcytoY230A) causes a dramatic redistribution of the molecules
to the cell surface. Such stably transfected populations contain
negative cells that serve as an internal control. These cells probably express only one or the other chain of the heterodimer and did not
stain with L243. Under these conditions, given the high background obtained with the intracellular staining method used for flow cytometry, there is an apparent underestimation of the total number of
class II molecules.
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 6.
Redistribution of the
DR/DMcytoY230A molecule to the cell
surface. Flow cytometry analysis using the L243 illustrates the
surface (empty region) and total expression (filled
region) of the DR/DMcyto molecule and mutant DR/DMcytoY230A
in unsorted HeLa-transfected cells. Cells were recovered using trypsin,
split, and stained separately either at the cell surface or fixed,
permeabilized, and stained with the FITC-conjugated L243. Marker
indicates the boundary between negative or positive cells for surface
expression of the chimeric molecules.
|
|
We then analyzed the DO cytoplasmic tail mutants in this assay. When
compared with the superimposed total and surface profiles of the
DR1-expressing cells (Fig.
7A), the curves obtained for the cells expressing the DR/DOcyto reporter molecules (Fig.
7B) showed a distinct pattern; the signal for the surface
amount of class II molecules is much weaker than the one for total
class II. This result is consistent with an intracellular accumulation of the reporter molecules and the role of the cytoplasmic tail of DO
in the sorting to the endocytic pathway. These results have been
reproduced consistently using either unsorted and sorted populations
expressing low or high levels of reporter molecules, as well as with
the mixed DRTM+DR18/DO pair (data not shown). The
data corroborate the confocal microscopy analysis presented above
showing the existence of a lysosomal sorting signal in reporter molecules containing the DO cytoplasmic tail. We next determined the
distribution of class II molecules in cells expressing
DR/DOcytoY227A. The intracellular and surface staining profiles were
similar to those of the wild-type DR/DOcyto chimera inasmuch as the
intracellular and cell surface curves did not overlap under these
conditions (Fig. 7C). The experiment confirmed that the
Y227A mutation does not alter lysosomal sorting.
View larger version (35K):
[in this window]
[in a new window]
|
Fig. 7.
Cell surface redistribution of the
DR/DOcytoLL242-243AA. Surface
(bold line) and total (dark region) class II
expression was monitored in stably transfected cells using
FITC-conjugated L243: DR (A), DR/DO cyto
(B), DR/DO cytoY227A (C), and
DR/DOcytoLL242-243AA (D) transfectants. Cells were
treated as in Fig. 6. All these cells were transfected with DRTM.
Data acquisition was performed under the same settings for all
comparable samples after matching the surface and total profiles
obtained for the DR cells.
|
|
Finally, we examined the effect of the LL 
AA mutations in this
assay. Fig. 7D shows that the DR/DOcytoLL242-243AA
molecules were clearly redistributed to the cell surface as the
patterns for total versus surface expression are comparable
between cells expressing DR or DR/DOcytoLL242-243AA molecules.
This result confirmed the critical role of the di-leucine motif in
sorting the reporter molecules to the endocytic pathway.
A Consensus Tyrosine-based Motif Rescued the Lysosomal Expression
of the LL AA Reporter Molecule--
By having generated a mutated
reporter molecule that is defective in lysosomal sorting
(DR/DOcytoLL242-243AA), it became feasible to characterize further
the potential role of tyrosine 227. Indeed, the possibility remained
that our cellular or experimental settings did not allow the efficient
recognition of the GYVRT sequence. Thus, to test if a known functional
tyrosine-based motif would have been recognized in our chimeras, we
designed a "gain" experiment by reconstructing a consensus
YXXL signal around the pre-existing Tyr-227 of the
DR/DOcytoLL242-243AA molecule. This was performed by replacing the
threonine 230 by the critical hydrophobic leucine residue to create the
YVRL sequence (DR/DOcytoAA/T230L) (Fig. 1B). The mutated
cDNA was transfected in HeLa cells together with the DRTM
cDNA, and the pattern of expression of this heterodimer was
verified in the bulk population. Wild-type DR and DR/DOcyto molecules were detected mostly at the cell surface by fluorescence microscopy (Fig. 8, A and B). However, analysis
of the HeLa DR/DOcytoAA/T230L cells revealed a defined intracellular
staining for the class II molecule with the presence of many enlarged
vesicles (Fig. 8C). Little if
any surface expression was detected on these cells (and see below), and
this pattern is similar to the one observed for the DR/DMcyto
molecule (Fig. 8D). These results confirmed the efficient
intracellular sorting of molecules containing a functional tyrosine
motif. The reduced surface expression of these molecules was confirmed
by flow cytometry using unsorted cells from a series of independent
transfections (Fig. 9). The weak but
detectable surface expression of the DR/DOcytoAA/T230L molecules could be due to the location of the tyrosine with respect to the transmembrane region. When compared with the cytoplasmic tail of
HLA-DM, the tyrosine residue in DR/DO is closer to the transmembrane (Fig. 1B) probably resulting in less stringent sorting
toward lysosomes (61, 65). Taken as a whole, these results confirm that
the tyrosine 227 in wild-type DO is not part of a classical if any
tyrosine-based sorting signal.
View larger version (35K):
[in this window]
[in a new window]
|
Fig. 8.
Characterization of the putative tyrosine
motif in HLA-DO. Total class II
expression was analyzed simultaneously on permeabilized cells by
fluorescence microscopy using biotinylated-L243 followed by Texas
Red-coupled streptavidin: DR (A), DR/DO cyto
(B), DR/DO cytoAA/T230L (C),
DR/DMcyto(D), and DR/DO cytoT230L (E).
These chains were transfected together with DRTM.
|
|
View larger version (31K):
[in this window]
[in a new window]
|
Fig. 9.
Reduced surface expression of reporter
molecules containing a consensus tyrosine motif. Surface
(bold line) and total (dark region) class II
expression was monitored in independent, stably transfected, unsorted
cell populations using FITC-conjugated L243. All these cells were
transfected with DRTM. Cells located left of the marker
in the DR panel are negative for class II expression and served as
internal controls to adjust acquisition settings.
|
|
We then tested the phenotype of cells expressing a reporter class II
molecule bearing both the di-leucine and tyrosine motifs. The presence
of two distinct signals is reminiscent of the situation in the DM-DO
complex that displays two motifs on separate cytoplasmic tails. The
YXXL motif was introduced in the "wild-type"
DR/DOcyto molecule that contains the original LL sequence (Fig.
1B), and stable transfectants were generated
(DR/DOcytoT230L). The class II staining patterns obtained by flow
cytometry or fluorescence microscopy were comparable to the ones
observed for HeLa DR/DMcyto or DR/DOcytoAA/T230L (Figs. 8 and 9).
Again, the presence of a tyrosine motif precluded high levels of
surface expression of a DR/DO reporter molecule demonstrating that
the two families of motifs, LL and YXXL, are not equivalent
and suggesting that the tyrosine-based sorting signal is dominant over
the di-leucine motif.
The Di-leucine Motif of HLA-DO Is Not Absolutely Required for
Lysosomal Sorting of the DO-DM Complex--
A previous study showed
that HLA-DO must first associate with HLA-DM to egress the ER and gain
access to the lysosomal compartments (39). Since both DO and DM contain
a sorting motif susceptible to target the complex to the lysosomes, we
sought to determine if the di-leucine motif of HLA-DO was necessary in
this process. We reasoned that surface expression of the DM-DO complex
and loss of intracellular vesicles would result from mutation of the
DOLL sequence if this motif was indeed critical for the sorting. To this end, we compared in HeLa cells the intracellular localization of
the HLA-DO-DM complex with that obtained for the same complex but
formed with an HLA-DO mutated at its di-leucine motif. We transfected
wild-type HLA-DO or HLA-DOAA into HeLa DM+ cells and
performed co-immunoprecipitations to confirm by Western blotting the
presence of associated DO and - chains in these cells (Fig.
10A). We then analyzed the
localization of the DO-DM complexes by fluorescence microscopy using
antibodies specific for DO or DM. Immunostaining revealed well
defined vesicles for both the DM-DO and DM-DOAA complexes (Fig.
10A). These patterns were similar to the one described
previously in HeLa cells for the wild-type DM-DO complex and clearly
different from the ER staining obtained for HeLa cells expressing DO
alone (Fig. 10B (39)). Contrary to the drastic effect of the
di-leucine mutation introduced in reporter molecules, this amino acid
change did not affect the vesicular accumulation of DO in the presence
of HLA-DM. Interestingly, both the HeLa DM.5 cells expressing HLA-DO
and HLA-DOAA were negative when assayed by flow cytometry for the class
II surface expression using the polyclonal mouse serum against DO
(data not shown). Together, these results agree with the observation described above that the tyrosine motif seems to dominate over the LL
signal and suggest that the di-leucine is not critical for the
lysosomal delivery of the DO-DM complex.
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 10.
Mutation of the LL motif in
DO does not prevent lysosomal accumulation of
the DO-DM complex. A, immunofluorescence microscopy
showing the pattern of expression of DO and DM. Permeabilized HeLa DM.5
cells expressing either the full-length DO + DO or the
full-length DO + DOAA molecules were stained intracellularly to
analyze localization. Cells were stained with the DM-specific antibody
(green) and with the mouse serum specific for HLA-DO
(red). The presence of both the DO and DO chains in
transfected DM.5 HeLa cells was confirmed by Western blot analysis of
samples immunoprecipitated with the mouse DO-specific serum and
revealed with a DO-specific rabbit serum. B,
intracellular staining of control HeLa cells expressing HLA-DO
molecules alone revealed a weak localized ER pattern but no scattered
endosomal vesicles. Western blot analysis as above revealed the
presence of both the DO and DO chains in transfected HeLa DO
cells.
|
|
| |
DISCUSSION |
The first indication that DO contains targeting information
came from experiments on the expression of a mixed isotypic pair between DR and DR18/DO (44). Immunofluorescence
microscopy revealed that this heterodimer was localized in
Lamp+ vesicular structures. However, we were concerned that
these molecules might accumulate in intracellular compartments
nonspecifically as a result of increased stability conferred by the
lumenal portion of DO. Indeed, HLA-DO normally resides in the lysosomal
compartments and must therefore be highly stable in this harsh
environment (39). To circumvent this problem and as a first step toward the disclosure of potential DO-encoded sorting signals, we addressed the importance of the cytoplasmic domain by constructing a chimeric DR1
molecule where the cytoplasmic tail of the chain was replaced by
the one of HLA-DO (Fig. 1). Our results clearly demonstrate that
this portion of DO is sufficient to bring the reporter molecule to
Lamp+ compartments (Fig. 3A). However, we cannot
rule out that amino acids motifs in DO and DO chains might
cooperate to increase sorting efficiency of the native HLA-DO molecule.
The two most common endosomal sorting signals are tyrosine- or
leucine-based sequences (18). The former is found in the cytoplasmic
tail of such protein as HLA-DM and are usually responsible for
specific, rapid internalization from the cell surface. The di-leucine
motif is thought to direct proteins to the endosomal/lysosomal compartments principally from the trans-Golgi network (reviewed in Ref.
66). Interestingly, the cytoplasmic tail of the HLA-DO contains two
such sequences, one imperfect tyrosine motif centered around position
227 and one di-leucine motif at position 242-243 (Fig. 1). Our
site-directed mutagenesis analysis revealed that the di-leucine motif
is responsible for the localization of reporter molecules in lysosomal
compartments. Furthermore, our results demonstrate that the putative
tyrosine signal is not recognized as such by the cellular machinery.
Since functional tyrosine motifs impede plasma membrane accumulation,
the conclusion that the tyrosine 227 of DO plays no role in sorting
probably holds true for B lymphocytes as well. Indeed, expression of a
mixed DR+DO heterodimer in a class II negative mutant B cell line
also resulted in significant surface expression (not shown).
The exact route taken by molecules of the antigen presentation
machinery is still a matter of debate (reviewed in Ref. 66). Early
experiments in HeLa cells showed that the AP1 adaptor was responsible
for the transport of the Ii-class II complex to the endocytic
compartment suggesting that this process is clathrin-dependent (67). AP-1 would mediate sorting at the trans-Golgi network, and it
would be recruited by the leucine-based motifs in Ii. It remains to be
determined if our reporter molecules are also sorted by AP-1 adaptor proteins.
Although the bulk of the Ii-class II complexes reaches the endocytic
pathway directly from the trans-Golgi, a significant proportion of the
proteins would first access the cell surface (68-70). Then, AP-2 is
probably recruited at the plasma membrane for integration of Ii-class
II complexes in clathrin-coated vesicles that mediate transport to the
endocytic pathway (66). Plasmon resonance experiments confirmed the
interactions between Ii and both AP1 and AP2 chains of the
clathrin-coated vesicles (71). We cannot rule out that reporter
molecules containing DO reach the lysosomes only after transiting at
the cell surface. However, these molecules are internalized and recycle
back to the cell surface in HeLa cells with similar kinetics as DR1
which does not accumulate in intracellular compartments (44) (data not shown). This suggests that sorting, at least in HeLa cells, does not
operate from the surface but rather directly at the trans-Golgi. Also,
it is not clear if plasma membrane expression conferred by the
cytoplasmic tail of HLA-DO is the result of trafficking through the
default pathway or exocytosis from lysosomes (72). Whatever the
mechanism, cell surface expression of our reporter molecule or of Ii in
B lymphocytes suggests that sorting directed by the di-leucine motif is
not very stringent. It is unlikely that the surface accumulation of our
chimeric molecules results from the overexpression of the transfected
proteins. Indeed, saturation of the sorting machinery in cells
expressing the DR/DMcyto would have resulted in the surface
expression of molecules such as Lamp-1 which also contain a
tyrosine-based motif (73, 74). No such expression could be detected by
flow cytometry (data not shown).
It is intriguing that mouse and rabbit O chains do not include an LL
motif (75, 76). This suggests a more specialized function for this
molecule in humans. However, it might explain why some H-2O was
detected at the surface of splenocytes by radioiodination studies (77).
Upon dissociation from DM, few H-2O molecules might escape
endosomal retention (see below) and end up at the cell surface.
Although HLA-DO contains a functional sorting signal, association with
HLA-DM is an obligatory step in its maturation. Karlsson and co-workers
(39) showed that the DO and DO chains associate in the ER, but
that egress and acquisition of complex sugars require an interaction
with HLA-DM. DM most probably assists in the folding of DO allowing its
release from ubiquitous ER chaperones. From there, however, the
relative importance of DO- and DM-encoded motifs in directing the
complex to lysosomes was not known. Our results show that substitution
of the two leucines for alanines in HLA-DO has no significant effect
on vesicular accumulation and suggest that the tyrosine-based signal of
DM is indeed the functional motif in this complex. Again, the fact that
the DO-DM complex does not accumulate at the plasma membrane of B
lymphocytes or transfected cells (on the contrary to the reporter
molecules relying on the di-leucine motif of DO) suggests that the
tyrosine motif of DM is recognized by the sorting machinery.
The fact that no DM-free HLA-DO has been found outside the ER (39) and
that the tyrosine motif of DM is sufficient to sort the complex to
lysosomes suggests that the role of the LL motif of HLA-DO needs to be
reconsidered. Maybe this di-leucine motif has a complementary role by
targeting some of the DO-DM complexes to earlier, less acidic
compartments than MHC class II-rich compartments where it might also
modulate class II loading. Presentation of antigens that are
independent of Ii expression and are dependent on the capacity of MHC
*
§,
,
TOP
ABSTRACT
INTRODUCTION
