International Immunology, Vol. 12, No. 10, pp. 1409–1416
© 2000 The Japanese Society for Immunology
Blocked transport of soluble Kb molecules
containing connecting peptide segment
involved in calnexin association
Shu-Bing Qian and Shi-Shu Chen
Department of Biochemistry & Molecular Biology, Shanghai Second Medical University, 280 South
Chongqing Road, Shanghai 200025, PRC
Keywords: calnexin, connecting peptide segment, kinetic association, MHC, transmembrane segment
Abstract
The molecular event governing the assembly of the MHC class I heavy chain–β2-microglobulinpeptide complex is still not fully understood. In order to characterize the transport properties of
MHC class I molecules, several truncated H-2Kb genes were constructed and expressed in COS7
cells. Surprisingly, the expressed soluble molecule containing connecting peptide (CP) segment
(sKbCP) did not secrete as efficiently as the one without CP (sKbCYT). When the sKbCP gene was
transfected into a calnexin-deficient cell line CEM.NKR, the amount of soluble Kb molecules in the
supernatant was comparable with sKbCYT-transfected CEM.NKR. To further demonstrate the different
transport of sKbCP and sKbCYT within living cells, we attached green fluorescent protein (GFP) to
the C-termini of both molecules and, as a comparison, to the full-length transmembrane
counterpart (mKb–GFP). While the mKb–GFP-transfected cells showed the green fluorescence in
the reticular network and the nuclear envelope, sKbCP–GFP showed obviously lump fluorescence of
high intensity within cells. However, the distribution of sKbCYT–GFP was fairly uniform.
Furthermore, GFP-tagged molecules allow us to analyze their interaction with other proteins in a
direct, simple and quantitative method, designated immunofluorescence precipitation. The results
showed that 60% of sKbCP–GFP molecules were associated with calnexin, while <10% with tapasin.
Taken together with the results from sKbCYT–GFP and mKb–GFP, it is reasonable to deduce that the
CP segment is involved in the association of class I molecules with calnexin and the
transmembrane region might play a dynamic role in the dissociation from calnexin. The suggested
kinetic association of class I molecules with calnexin is likely to contribute to the different
maturation rate between several class I alleles.
Introduction
The role of MHC class I molecules as indicators of intracellular
protein transport is reflected in their molecular design which
does not allow them to leave the endoplasmic reticulum (ER)
unless they have completely assembled to form class I heavy
chain–β2 microglobulin- (β2m) peptide complexes (1,2). While
the critical elements of the class I presentation pathway have
been defined, the molecular event governing the assembly
of this complex is still not fully understood. There are a number
of unresolved issues surrounding the interaction of assembling
class I molecules with calnexin, an ER type I transmembrane
chaperone. It was suggested that oligosaccharide residues
might serve as substrates for calnexin binding. However,
unglycosylated class I molecules can also bind calnexin (3).
As a chaperone, calnexin facilitates folding of the nascent
heavy chains and promotes assembly of heavy chains with
β2m. Surprisingly, the assembly and function of class I molecules appear to be normal in a calnexin-negative cell line
(4). While the lumenal domain, transmembrane region and
cytoplasmic tail (CYT) of calnexin have all been shown to
play a role in its interaction with various targets (5–7), the
identification of the class I domain which determines the
interaction with calnexin remains controversial (8). Previous
experiments showed that a glycophosphatidylinositol (GPI)linked Q7b class I molecule was associated with calnexin,
whereas a soluble Q7b isoform was not calnexin associated.
These results implicated that the 9 amino acid fragment,
connecting the α3 domain with the transmembrane region,
might be the site of interaction with calnexin (3). However,
Correspondence to: S.-B. Qian
Transmitting editor: T. Sasazuki
Received 31 January 2000, accepted 21 June 2000
1410 Calnexin association with Kb molecules
direct evidence is still lacking. On the other hand, studies
using soluble calnexin suggested that the transmembrane
domain might be important in the interaction of these integral
membrane proteins (6).
To investigate this problem, we generated two modified Kb
cDNAs from which both the transmembrane region and
the CYT were absent. One construct, sKbCP, contains the
connecting peptide (CP) segment following the α3 domain of
Kb and the other replaces the CP segment with terminal CYT
sequences of the same length (sKbCYT). Both engineered
soluble Kb molecules were expressed in COS7 cells and
the calnexin-deficient cell line CEM.NKR. Unlike sKbCYT, the
secretion of sKbCP was almost blocked in transfected COS7
cells, suggesting that the CP segment is directly involved
in the association with calnexin. However, the behavior of
membrane-anchored Kb molecules implies that the transmembrane domain of heavy chains may regulate the calnexin
dissociation. The unexpected role of the transmembrane
domain was further confirmed by the retarded transport of
integral membrane Kb tagged with green fluorescent protein
(GFP) at its C-terminus. The intracellular status of various
GFP-tagged versions was analyzed quantitatively by immunofluorescence precipitation (IFP). The suggested kinetic association of class I molecules with calnexin is likely to contribute
to the different maturation rate between several class I alleles.
Methods
Construction of truncated and fusion genes
Full-length H-2Kb cDNA was cloned from the splenocytes of
C57BL/6 mouse using RT-PCR. The forward oligonucleotide
primer (mKb5), 5⬘-TAGGATCCATGGTACCGTGCACGCTGCTCCTGCTGT-3⬘, includes a BamHI site (underlined) upstream
of the initiation ATG codon, and the reverse primer (mKb3), 5⬘CGGTCGACTCAAAGCTTCGCTAGAGAATGAGGGTC-3⬘,
introduces both a HindIII site and SalI site in the flank of the
termination TCA codon. The H-2Kb cDNA was cloned into
unique BamHI and SalI sites in pBluescript II SK⫹ plasmid
(Stratagene, La Jolla, CA), generating pBluescript/Kb followed
by sequencing to confirm its identity. For the fused Kb–GFP
construct, the GFP was first removed from phGFP-S65T (Clontech, Palo Alto, CA) with HindIII and XbaI, and then cloned into
pBluescript/Kb digested with the same restriction enzymes.
Thus the termination TCA codon of Kb was excised and the
GFP was fused in-frame to the 3⬘ end of Kb coding region. For
transient expression studies, pcDNA3/Kb–GFP was generated
by inserting the Kb–GFP fragment of 1.8 kb into BamHI–XhoI
digested pcDNA3 (Invitrogen, Carlsbad, CA). One soluble form
of Kb, sKbCP, was amplified by PCR using mKb5 and another 3⬘
primer (sKb3), CGGTCGACTCAAAGCTTGTTGGAGACAGTGGATGG, which includes the same enzyme sites as mKb3.
The PCR product, with a stop codon inserted at 1003, encodes
a truncated Kb molecule from which both the transmembrane
region and CYT were absent. The sKbCP–GFP fusion gene was
constructed using the similar method as mKb–GFP. For another
soluble form, sKbCYT–GFP, a pair of overhang primers was
designed. The 5⬘ primer (CG5), GTTCATGACCCTATTCTCTAGCGAAGCTTGCCGCCACC, encodes a portion of terminal
CYT residues and the 5⬘ end of GFP, and the 3⬘ primer (SC3),
CGCTAGAGAATGAGGGTCATGAACCATCCATCTCAGGGTGAGGG, contains the sequence for the 3⬘ end of α3 domain of
Kb and a portion of terminal CYT residues, an overlap to CG5.
The PCR products of mKb5 ⫹ SC3 and CG5 ⫹ GFP3 were
mixed and spliced by overlap extension. The GFP-untagged
version, sKbCYT, was then directly amplified with primers of
mKb5 and mKb3 using sKbCYT–GFP as template.
Cell culture and reagents
Transformed simian cell line COS7 and calnexin-deficient cell
line CEM.NKR (generously provided by Dr Peter Cresswell, Yale
University) were maintained in DMEM (Gibco, Gaithersburg,
MD), supplemented with 10% (v/v) heat-inactivated newborn
calf serum (Gibco). EL4, Jurkat, TAP-deficient cell line RMA-S
and hybridoma cell line secreting anti-Kb mAb AF6-88.5.3
(Ig2a) were obtained from ATCC (Rockville, MD), and cultured
in RPMI 1640 medium (Gibco) containing 10% (v/v) fetal bovine
serum (Hyclone, Logan, UT). Media for all cultures routinely
included 2 mmol/l glutamine, 1 mmol/l pyruvate, 100 U penicillin
and 100 µg/ml of streptomycin. All cultures were maintained at
37°C in a 5% CO2 humidified atmosphere.
DNA transfection
COS7 cells grown to 60–80% confluence on coverslip in a sixwell culture plate (Nunc, Rochester, NY) were transfected with 1
µg of recombinant pcDNA3 plasmid using 8 µl of Lipofectamine
reagent (Gibco) per well, according to the manufacturer’s
instructions. Briefly, COS7 cells were incubated with a DNA/
Lipofectamine mixture in serum-free media for 6 h, the media
were replaced with complete media and the cells were analyzed 48 h after transfection. For suspension cells, like EL4,
Jurkat, CEM.NKR and RMA-S cells, the transfection was performed by electroporation (250 V and 960 µF; GenePulser;
BioRad, Hercules, CA) because of its resistance to liposomemediated transfection.
ELISA
The concentration of recombinant proteins in supernatants was
analyzed by ELISA. Primary antibodies used to coat PRO-BIND
plates (Falcon, Franklin Lake, NJ) were mouse anti-human β2m
or rabbit anti-mouse β2m (PharMingen). Supernatants (100 µl)
from transfected cells were incubated for 1 h at room temperature. Plates were washed extensively and incubated with AF688.5 for 1 h. After extensive washing, plates were incubated
with horseradish peroxidase-conjugated goat anti-mouse IgG
(100 µl of a 1:1000 dilution; Sino-American, Shanghai, PRC) for
1 h, washed and developed with 3,3⬘,5,5⬘-tetramethylbenzidine
dihydrochloride substrate for 10 min. The reaction was stopped
with the addition of H2SO4 (1 M) and absorbance was read at
450 nm.
Fluorescence microscopy
Cell samples were plated onto 35 mm dishes containing a
centered glass coverslip. Microscopic evaluation of GFP
expression was carried out by direct observation of living cells
using a fluorescence microscope equipped with a FITC filter
and a xenon power supply (Nikon). For laser scanning confocal
microscopy, cell monolayers grown on glass coverslips were
analyzed in a laser scanning confocal microscope (Zeiss
LSM510) equipped with a 488 nm argon. To detect the expres-
Calnexin association with Kb molecules 1411
sion of Kb on the cell surface, indirect immunofluorescence was
performed with mAb AF6-88.5.3 (Ig2a). A phycoerythrin (PE)conjugated anti-mouse Ig (Sigma) was used as secondary
antibody. Antibody incubation was performed at 4°C for 30 min,
and cells were washed 3 times between applications using PBS
containing 0.02% sodium azide and 1% BSA. Dual-channel
fluorescences (GFP and PE) were observed simultaneously.
Digitized images of confocal optical sections (2 µm) were compiled into final figures using Adobe PhotoShop software (version 5.0, Adobe Systems).
IFP
Cells were washed twice with ice-cold PBS, and lysed with 1%
3-[(3-cholamidopropyl) dimethylammonio]-1-propanesulfate
(Pierce, Rockford, IL) in Tris-buffered saline, pH 7.4, containing
10 mM leupeptin, 1% aprotinin and 0.25 mM phenylmethylsulfonyl fluoride. After incubation for 30 min on ice, lysates were
centrifuged. The postnuclear supernatant was cleared overnight at 4°C with normal rabbit serum and PANSORBIN Staphylococcus aureus (Calbiochem, San Diego, CA). One aliquot
was directly analyzed the fluorescence intensity and the others
were performed by immunoprecipitation. The following antibodies were used: R.gp48N, a rabbit anti-peptide antibody to
the N-terminal region of tapasin (9); anti-calnexin mAb (kindly
provided by Dr Peter Cresswell in Yale University) (10); 28-86s, a mouse mAb that binds assembled form of Kb and Kd
(PharMingen) and AF6-88.5. Aliquots were then incubated for
1 h at 4°C with antibody, followed by 30 min at 4°C with Protein
A–Sepharose (Pharmacia). The post-precipitated supernatant
was analyzed by fluorometry or re-precipitated by different
antibody for dual-IFP experiments.
Western blot analysis
For protein immunoblotting, the first round immunoprecipitates
described above were subjected to SDS–PAGE and transferred
to a nitrocellulose membrane. The membrane was incubated
with antibody against GFP (Clonetech). After incubation with
horseradish peroxidase-conjugated secondary antibody, the
membrane was visualized by enhanced chemiluminescence
(ECL; Amersham, Piscataway, NJ).
Pulse–chase analysis
For pulse–chase analysis, transfected cells were starved in
methionine-free medium for 50 min, pulsed with [35S]methionine at 150 µCi/ml for 15 min and chased in medium containing
excess of methionine (0.5 mM) for 180 min. Cells (5⫻106) were
collected, washed and lysed as described above. GFP-tagged
Kb molecules were recovered with anti-GFP antibody, while
wild-type Kb with antiserum to the C-terminal peptide encoded
by exon 8 of the Kb gene (anti-X8, kindly provided by Dr Nathenson, Albert Einstein College of Medicine) (11). The immunoprecipitates were washed and treated with Endo H (Boehringer
Mannheim, Germany) or under control conditions before analysis by SDS–PAGE and autoradiography.
Results
Secretion of soluble Kb molecules in COS7 and CEM.NKR cells
We generated two modified Kb cDNAs from which both the
transmembrane region and the CYT were absent. One con-
Fig. 1. Secretion of CP-containing soluble Kb molecules was almost
blocked in calnexin-positive cell lines but normal in calnexin-deficient
CEM.NKR cells. Supernatants (100 µl) from transfected COS7,
CEM.NKR, EL4, Jurkat and RMA-S cells were tested for reactivity with
Kb-specific mAb (AF6-88.5).
struct, sKbCP, contains the CP segment following the α3 domain
of Kb and the other replaces the CP segment with terminal CYT
sequences of the same length (sKbCYT). Both constructs were
expressed in COS7 cells in a mammalian expression vector
pcDNA3. Supernatants from the transfected cells were tested
for the presence of the secreted form of Kb molecules by an
enzyme-linked immunosorbent assay (ELISA). As shown in Fig.
1, while a certain amount of soluble Kb was produced by sKbCYTtransfected cells, only trace was detected in sKbCP-transfected
cells. This result suggested that CP containing soluble Kb molecules could not be efficiently secreted from the cells. However,
the fate of sKbCP molecules within cells is unclear, for Northern
blot analysis showed both constructs were expressed at almost
the same level (data not shown). In order to determine whether
calnexin plays a role in blocking the secretion of sKbCP from
COS7, a similar experiment was done in CEM.NKR, which does
not express calnexin at the mRNA and protein level (12). Surprisingly, both sKbCP and sKbCYT were secreted from transfected CEM.NKR cells. Thus the results obtained from both
COS7 and CEM.NKR indicate that the CP segment may participate in the association of class I molecules with calnexin.
We expanded the results in COS7 to other calnexin-positive
cell lines, including mouse lymphoma EL4 and human leukemia
Jurkat (Fig. 1). Both cell lines did not secrete sKbCP after transfection. Due to different transfection efficiency, the secreted
amount of sKbCYT was relatively lower as compared with transfected COS7. There was no marked difference in the expression
pattern of soluble Kb in both cells, although several studies have
suggested that murine class I molecules behave differently in
human and mouse cells (13). It was shown that human β2m
could much more readily associate with Kb to form a stable
conformation (14). However, it still remains to be determined
whether soluble HLA class I molecules have the same fate
as Kb.
The results obtained from a TAP-deficient cell line, RMA-S,
were interesting. There was no dramatic difference of expres-
1412 Calnexin association with Kb molecules
sion between sKbCP and sKbCYT, especially the GFP-tagged
versions. Thus, the secretion of soluble Kb molecules might be
in a peptide-dependent manner. It was known that RMA-S could
express some endogenous empty Kb molecules in its cell surface when cultured at 26°C (15). However, because our ELISA
system detected only the conformational epitope, we could not
exclude the possibility that some soluble Kb molecules were
secreted from RMA-S cells.
Distribution of GFP-tagged soluble Kb molecules
To further characterize the features of both truncated Kb molecules within living cells, we attached GFP to the C-termini of both
constructs, generating chimeric sKbCP–GFP and sKbCYT–GFP,
and, as a comparison, to the full-length transmembrane counterpart (mKb–GFP). COS7 cells were transfected with all the
three GFP-fused chimeric constructs respectively. The results
from laser scanning confocal microscope showed markedly
different fluorescence patterns (Fig. 2B). While the mKb–GFP
transfected cells showed the green fluorescence in the reticular
network and the nuclear envelope, which are characteristic of
the ER, sKbCP–GFP showed obviously lump fluorescence of
high density within cells, indicating strong accumulation within
some organelles. However, the distribution of sKbCYT–GFP was
fairly uniform. In addition, both ELISA and fluorometry were
used to examine the secretion of GFP-tagged soluble Kb from
transfected COS7 cells. Similar to GFP-untagged counterparts,
only small amounts of sKbCP–GFP were produced in the supernatant, while much larger amounts of sKbCYT–GFP can be
detected (Fig. 1). On the other hand, the secretion of sKbCP–
GFP was normal in CEM.NKR, resulting in a similar fluorescence
pattern to sKbCYT–GFP (Fig. 2). In contrast to CEM.NKR, GFPtagged Kb molecules were obviously accumulated within RMAS cells because of TAP deficiency, resulting in a similar fluorescence pattern as adherent COS7 cells transfected with sKbCP–
GFP. It is noteworthy that mKb–GFP can be expressed on the
cell surface as determined by Kb-specific antibody staining
(Fig. 2A). However, it is surprising to find the different cell
surface expression level between mKb–GFP and the GFPuntagged natural counterpart. The mRNA expression level of
both Kb–GFP and Kb was almost similar as determined by
Northern blot analysis (data not shown), thus excluding the
possibility of different transfection efficiency.
IFP analysis and Western blot
Not only can the intracellular localization of the GFP-tagged
protein be monitored within living cells non-invasively (16), its
interaction with other proteins can also be analyzed quantitatively in a simple and rapid way. This approach, designated IFP,
utilizes the fluorometry to detect the change of fluorescence
intensity in the whole cell lysates before and after immunoprecipitation. Unlike standard immunoprecipitation using metabolic radiolabeling, in which only newly synthesized proteins
can be labeled and analyzed, the method using biochemical
tagging could put all of the tagged proteins into consideration.
A similar quantitative method was the assay of β-galactosidase
activity in the immunoprecipitates by fusing the β-galactosidase gene at the end of the cytoplasmic tail of the interested
protein (17,18). More steps were needed in measuring β-galactosidase activity in the immunoprecipitates than IFP that avoids
extensive washing of the immunoprecipitates. The amount of
co-precipitated GFP-tagged proteins can be easily determined
by calculating the different fluorescence intensity between preand post-precipitated whole cell lysates, thus increasing the
sensitivity of detecting any weak interaction between proteins.
The association with calnexin and tapasin may represent two
stages of class I assembly within the ER. Calnexin interacts
with newly synthesized chain (19), while tapasin bridges the
class I heavy chain–β2m with TAP, awaiting the loading of the
appropriate peptide (20). Generally, the amount of calnexinassociated, tapasin-associated and completely folded mature
form represents almost the total class I molecules within cells.
Thus, evaluation of the status of GFP-tagged class I molecules
within cells could be performed by IFP using anti-calnexin, antitapasin and Kb conformation-specific antibody.
Among the total mKb–GFP molecules within transfected
COS7 cells, the Kb conformation-specific antibody-reactive
molecules were ⬍10%, while nearly half of the mKb–GFP molecules were co-precipitated with calnexin (Fig. 3A). This feature
implies that GFP tagging may impede the maturation of membrane-anchored Kb molecules. The retarded transport of mKb–
GFP is consistent with the low cell surface expression detected
by dual-cannel fluorescence analysis after PE-conjugated antibody staining as compared with wild-type Kb (Fig. 2A). To rule
out the possibility that the association occurs after solubilization, cells transfected with GFP-untagged constructs and cells
transfected with GFP-tagged counterparts were mixed before
solubilization. No change in the proportion of co-precipitated
fluorescence intensity was observed (data not shown). In contrast to COS7, the maturation of mKb–GFP was rapid in transfected CEM.NKR cells. IFP demonstrated that 65% of mKb–GFP
were conformation-specific antibody-reactive in CEM.NKR
cells.
An overlap can be seen from the results of IFP in the mKb–
GFP-transfected CEM.NKR cells, in which the proportion of
both tapasin-associated and mature form was ⬎100%. It
is still possible that the non-specific co-precipitation might
exist in IFP experiments. We next performed dual-IFP
analysis to examine possible non-specific co-precipitation.
A slight overlap was observed in mKb–GFP-transfected
COS7 cells between the association with calnexin and
tapasin (Fig. 3B). It is unlikely that GFP-tagged transmembrane Kb can form intermediates consisting of both calnexin
and tapasin. Thus non-specific co-precipitation indeed
existed, but at ⬍5%. It is interesting to find that some (~20%)
sKbCP–GFP in transfected COS7 was neither associated with
calnexin nor with tapasin. The fluorescence pattern of
sKbCP–GFP in COS7 cells strongly suggests that the
accumulation might occur in other organelles in addition to
the ER. A possible explanation is that these molecules
might under degradation within cytosol. Such possibility
remains to be investigated. To further confirm the results
obtained from IFP analysis, we next performed Western
blot of the immunoprecipitates using anti-GFP antibody.
Consistent results were obvious, although it is difficult for
quantitative analysis (Fig. 4).
Pulse–chase analysis
As an additional approach to the above question, we performed a pulse–chase analysis and compared the maturation
rate of GFP-tagged and wild-type Kb molecules in transfected
Calnexin association with Kb molecules 1413
Fig. 2. Immunofluorescence localization of various GFP-tagged Kb molecules in transfected cells. (A) GFP tagging retards the intracellular
transport of Kb in COS7. COS7 cells were transfected with GFP, mKb and mKb–GFP. At 48 h after transfection, cells were washed and
incubated with 28-8-6s (mouse anti-Kb and Kd), followed by staining with PE-conjugated anti-mouse IgG. Both green (GFP) and red (PE)
channels were analyzed with excitation at 488 nm. (B) CP-containing soluble Kb molecules were accumulated within COS7 cells. COS7 cells
were transfected with mKb–GFP, sKbCP–GFP and sKbCYT–GFP by Lipofectamine reagent, and transfection of CEM.NKR and RMA-S cells was
performed by the electroporation method. Immunofluorescence was directly analyzed on living cells with confocal microscope (LSM 510).
Bar, 5 µm.
COS7 and CEM.NKR cells. As shown in Fig. 5, processing of
chimeric mKb–GFP proteins into Endo H-resistant species
was reduced in COS7 as compared with wild-type. It is
noteworthy that the maturation of mKb–GFP was also retarded
in CEM.NKR, while the transport of sKbCYT–GFP was much
more rapid in CEM.NKR than in COS7 cells. The nature of the
slight slower maturation rate of sKbCYT–GFP molecules in
COS7 than in CEM.NKR is unclear. One possibility is that the
assembly machinery might not be identical between these
two different cell lines. An alternative explanation is that the
1414 Calnexin association with Kb molecules
Fig. 4. Western blot analysis of GFP-tagged Kb molecules in
immunoprecipitates of COS7 cells transfected with mKb–GFP, sKbCP–
GFP and sKbCYT–GFP. Immunoprecipitation was performed using
anti-calnexin, anti-tapasin (R.gp48N) and anti-Kb (28-8-6s).
Immunoisolated molecules were subjected to 10% SDS–PAGE and
transferred to a nitrocellulose membrane. The membrane was blotted
with antibody against GFP and visualized by enhanced
chemiluminescence.
Fig. 3. The intracellular status of mKb–GFP, sKbCP–GFP and sKbCYT–
GFP in transfected COS7 and CEM.NKR (A). Transfected cells were
lysed, and fluorescence intensity was determined in whole cell lysates
before immunoprecipitation with various antibodies. The proportion
of total fluorescence intensity co-precipitated with various molecules
is expressed as a percentage in every experiment. (B) Dual-IFP in
transfected COS7 cells. After the first round of IFP (white), the second
round of precipitation (black) was performed using another antibody
different from the one used in the first round.
Discussion
Fig. 5. Intracellular transport of various GFP-tagged Kb molecules in
COS7 (A) and in CEM.NKR (B). COS7 cells were transfected with
various constructs by Lipofectamine reagent (Gibco) and transfection
of CEM.NKR cells was performed electroporation. At 48 h after
transfection, cells were incubated with [35S]methionine for 15 min
and then in the presence of excess unlabeled methionine for 180
min. Wild-type Kb molecules were immunoisolated with anti-X8
antibody and GFP-tagged Kb molecules were recovered with antiGFP antibody. Immunoisolated molecules were digested (⫹) or mockdigested (–) with Endo H and subjected to 8% SDS–PAGE. The
mobilities of the Endo H-resistant (R) and Endo H-sensitive (S)
molecules are indicated. Molecular sizes are indicated on the left
(in kDa).
We firstly found that the soluble Kb molecule containing CP
segment (sKbCP) did not secrete efficiently as the one without
CP (sKbCYT). It is unlikely that the sKbCP molecules misfolded
within the ER of COS7, while sKbCYT folded properly and
transported out of the ER. Both constructs were similar, only
differing in their terminal nine residues. To further characterize
the features of both soluble Kb molecules within living cells,
we attached GFP to the C-termini of both truncated molecules.
The resulting intensified fluorescence from sKbCP–GFP indicates that most of such molecules were accumulated within
cells. Furthermore, we designed a novel method, named IFP,
in order to evaluate the association of various GFP-tagged
Kb molecules with other proteins. The results showed that
60% of sKbCP–GFP molecules were associated with calnexin,
while ⬍10% with tapasin. Taken together with the results from
sKbCYT–GFP and mKb–GFP, it is reasonable to deduce that
the CP segment is involved in the association of class I
molecules with calnexin.
However, it is critical to define whether the additional GFP
tag might interfere with the folding process of the target
protein (21). Different consideration should be given to GFPtagged transmembrane Kb and its soluble form. After the
synthesis of GFP-tagged transmembrane Kb heavy chain,
only proximal domains of Kb (α1, α2 and α3) entered into the
lumen of the ER, while both the cytoplasmic domain and the
following GFP-tag faced the cytosol. Thus, it is unlikely
that the GFP-tag affects the lumenal domains across the
membrane. As to the GFP-tagged soluble version, however,
all the full-length chimeric chain entered into the lumen of the
ER after its synthesis. The fact that the level of secreted
expression level of delivered genes in COS7 cells is much
stronger than in CEM.NKR, resulting in more newly synthesized
class I molecules per cell for COS7.
Calnexin association with Kb molecules 1415
GFP-tagged soluble molecules was comparable with that of
GFP-untagged counterparts suggests that the interference of
GFP tagging with the folding process of Kb molecules, if it
exists, is minimal.
The fact that the association of calnexin with the
CP-containing soluble Kb molecule almost blocks its transport
raises an interesting question about how the wild-type Kb,
which also contains the CP segment, dissociates from calnexin
and then transports to the cell surface. As yet, there have
been no direct demonstrations of the molecular events determining the release of class I molecules from calnexin. The
most plausible explanation is that the conformational change
during class I assembly may induce the dissociation of many
chaperones (22). In human cells, class I binding to β2m is
thought to cause dissociation of calnexin (23). However, it is
not the case for murine cells in which an increased association
between calnexin and Db–β2m complexes was observed in
TAP-deficient RMA-S cells (24).
We speculate that the association of integral class I molecules with calnexin may involve a kinetic model in which the
CP segment is necessary for the association with calnexin,
while the transmembrane or cytoplasmic domain may regulate
the release from calnexin. The fact that GPI-anchored class
I molecules can be expressed normally on the cell surface
suggests that the cytoplasmic tail may not play a major role
in this process. Our results argue that the transmembrane
domain or GPI might play an unexpected role in regulating
the transport of class I molecules, although the difference
between the transmembrane domain and GPI anchor cannot
be defined at present. Generally the driving force could be
generated from the potential movement of the transmembrane
region along the plane of the membrane and can be perturbed
by the addition of GFP at the end of the CYT (Fig. 2A).
However, deletion of the transmembrane region might lose
the driving force, resulting in the almost permanent association
with calnexin, as confirmed by the blocked transport of sKbCP–
GFP in COS7 cells.
During the class I assembly process, the conformational
change of the lumenal domains of class I molecules could
induce calnexin dissociation by binding to other chaperones,
such as calreticulin and tapasin. Compelling evidence suggested that the absence of either β2m or peptide could cause
the increased association with calnexin (25,26). Thus the
kinetic association of class I molecules with calnexin is a
result of a balance involving several forces. The different role
played by the CP and transmembrane domain is unexpected,
and may provide evidence to explain the behavior of different
MHC class I molecules. The analysis of the maturation rate
of MHC class I molecules has shown marked differences
between several class I alleles (27–29). Interestingly, the fast
and slow maturation alleles have different residues in the CP
segment. The former, such as Kb and Kk, contain EPPPSTVSN
in their CP segment, whereas the CP of the latter, such as
Dk, Db and Ld, consists of EPPPSTDSY. Whether the two
different residues (italic letter) determine the varied association forces with calnexin remains to be identified.
On the other hand, the forces provided by the transmembrane region might also play a role in preventing the degradation of class I molecules. Previous experiments suggested
that the misfolded heavy chain could be removed from the
ER to the cytosol and degraded by the proteasome (30).
However, the machinery responsible for this event is unclear.
Our results imply that the calnexin-associated sKbCP could
be more easily translocated from the ER lumen to the cytosol,
suggesting calnexin might play an important role in mediating
the ‘retrotranslocation’ by translocan (31). Up till now, it has
been much more difficult to analyze the exclusive forces than
the associate interaction between transmembrane domains.
The results obtained from detergent solution may not reflect
the situation in vivo where the components of the complex
remain membrane bound. It is known that the assembly
machinery of the ER consists of some integral membrane
proteins and several lines of evidence have also suggested
that the transmembrane regions play different roles in their
function (32,33). The subtle nature of the transmembrane
interaction raises the prospect that other integral proteins
may be similarly engaged in the putative kinetic model in
regulating their transport within cells.
Acknowledgements
We thank P. Cresswell (Yale University) for providing anti-Tapasin,
R.gp48N and cell line CEM.NKR, S. G. Nathenson (Albert Einstein
College of Medicine) for providing antiserum of ant-X8, the Department of Histology & Embryology for the help with confocal microscope
analysis, and Drs Yewdell and Bennink (National Institute of Health)
for critical review. This work was supported by the National Natural
Science Foundation in China (to S. B. Q.) and the Developing
Foundation of Shanghai Education Committee (to S. B. Q.).
Abbreviations
β2m
CP
CYT
ER
GFP
GPI
IFP
PE
β2-microglobulin
connecting peptide
cytoplasmic tail
endoplasmic reticulum
green fluorescent protein
glycophosphatidylinositol
immunofluorescence precipitation
phycoerythrin
References
1 Townsend, A., Ohlen, C., Bastin, J., Ljunggren, H. G., Foster, L.
and Karre, K. 1989. Association of class I major histocompatibility
heavy and light chains induced by viral peptides. Nature 340:443.
2 Yewdell, J. W. and Bennink J. R. 1992. Cell biology of antigen
processing and presentation to MHC class I molecule-restricted
T lymphocytes. Adv. Immunol. 52:1.
3 Carreno, B., Schreiber, K., McKean, D., Stroynowski, I. and
Hansen, T. 1995. Aglycosylated and phosphatidylinositolanchored MHC class I molecules are associated with calnexin.
J. Immunol. 154:5173.
4 Scott, J. and Dawson, J. 1995. MHC class I expression and
transport in a calnexin deficient cell line. J. Immunol. 155:143
5 Wada, I., Ou, W. J., Liu, M. C. and Scheele, G. 1994. Chaperone
function of calnexin for the folding intermediate of gp80, the major
secretory protein in MDCK cells. Regulation by redox state and
ATP. J. Biol. Chem. 269:7464.
6 Ho, S. C., Rajagopalan, S., Chaudhuri, S., Shieh, C. C., Brenner,
M. B. and Pillai, S. 1999. Membrane anchoring of calnexin
facilitates its interaction with its targets. Mol. Immunol. 36:1.
7 Capps, G. G. and Zuniga, M. C. 1994. Class I histocompatibility
molecule association with phosphorylated calnexin. Implications
for rates of intracellular transport. J. Biol. Chem. 269:11634.
8 Cerundolo, V. and Veronique, B. 1996. Cell biology of MHC class
1416 Calnexin association with Kb molecules
9
10
11
12
13
14
15
16
17
18
19
20
21
I molecules. In Browning, M. J. and McMichael, A. J., eds, HLA
and MHC: Genes, Molecules and Function, p. 193. BIOS, Oxford.
Sadasivias, B., Lehner, P. J., Ortmann, B., Spies, T. and Cresswell,
P. 1996. Roles for calreticalin and novel glycoprotein, tapasin, in
the interaction of MHC class I molecules with TAP. Immunity 5:103.
Hochstenbach, F., David, V., Watkins, S. and Brenner, M. B.
1992. Endoplasmic reticulum resident protein of 90 kilodaltons
associates with the T- and B-cell antigen receptors and major
histocompatibility complex antigens during their assembly. Proc.
Natl Acad. Sci. USA 89:4734.
Smith, M. H., Parker, J. M., Hodges, R. S. and Barber, B. H. 1986.
The preparation and characterization of anti-peptide
heteroantisera recognizing subregions of the intracytoplasmic
domain of class I H-2 antigens. Mol. Immunol. 23:1077.
Prasad, S. A., Yewdell, J. W., Porgador, A., Sadasivian, B.,
Cresswell, P. and Bennink, J. R. 1998. Calnexin expression does
not enhance the generation of MHC class I–peptide complexes.
Eur. J. Immunol. 28:907.
Alexander, J., Payne, J. A., Murray, R., Frelinger, J. and Cresswell,
P. 1989. Differential transport requirements of HLA and H-2 class
I glycoproteins. Immunogenetics 29:380.
Fahnestock, M., Tamir, I., Narhi, L. and Bjorkman, P. 1992. Thermal
stability comparison of purified empty and peptide-filled forms of
a class I MHC molecule. Science 258:1658.
Ortiz, N. V. and Hammerling, G. J. 1991. Surface appearance
and instability of empty H-2 class I molecules under physiological
conditions. Proc. Natl Acad. Sci. USA 88:3594.
Misteli, T. and Spector, D. L. 1997. Applications of the green
fluorescent protein in cell biology and biotechnology. Nat.
Biotechnol. 14:610.
Cosson, P. and Bonifacino, J. S. 1991. Membrane protein
association by potential intramembrane charge pairs. Nature
351:414
Cosson, P. and Bonifacino, J. S. 1992. Role of transmembrane
domain interactions in the assembly of class II MHC molecules.
Science 258:659.
Jackson, M. R., Cohen-Doyle, M. F., Peterson, P. A. and Williams,
D. B. 1994. Regulation of MHC class I transport by the molecular
chaperone, calnexin (p88, IP90). Science 263:384.
Ortmann, B., Androlewicz, M. J. and Cresswell, P. 1994. MHC
class I/β2-microglobulin complexes associate with TAP
transporters before peptide binding. Nature 368:864.
Cubitt, A B., Heim, R., Adams, S. R., Boyd, A. E., Gross, L. A.
22
23
24
25
26
27
28
29
30
31
32
33
and Tsien, R. Y. 1995. Understanding, improving and using green
fluorescent proteins. Trends Biol. Sci. 20:448.
Koopmann, J.-O., Hammerling, G. J. and Momtury, F. 1997.
Generation, intracellular transport and loading of peptides
associated with MHC class I molecules. Curr. Opin. Immunol. 9:80.
Sugita, M. and Brenner, M. B. 1994. An unstable beta-2microglobulin: major histocompatibility class I heavy chain
dissociates from calnexin and then is stabilized by binding
peptide. J. Exp. Med. 180:2163.
Degen, E., Cohen-Doyle, M. F. and Williams, D. B. 1992. Efficient
dissociation of the p88 chaperone from major histocompatibility
class I molecules requires both β2 microglobulin and peptide.
J. Exp. Med. 175:1653.
Ou, W.-J., Cameron, P. H., Thomas, D. Y. and Bergeron, J. J. M.
1993. Association of folding intermediates of glycoproteins with
calnexin during protein maturation. Nature 364:771.
Rajagopalan, S. and Brenner, M. B. 1994. Retention of
unassembled components of integral membrane proteins by
calnexin. Science 263:387.
Williams, D., Sweiedler, S. and Hart, G. 1985. Intracellular transport
of membrane glycoproteins: two closely related histocompatibility
antigens differ in their rates of transmit to the cell surface. J. Cell.
Biol. 101:725.
Weis, J. and Murre, C. 1985. Differential expression of H-2Dd and
H-2Ld histocompatibility antigens. J. Exp. Med. 161:356.
Beck, J., Hansen, T., Cullen, S. and Lee, D. 1986. Slower
processing, weaker β2-m association, and lower surface
expression of H-2Ld are influenced by its amino terminus.
J. Immunol. 13:916.
Hughes, E. A., Hammond, C. and Cresswell, P. 1997. Misfolded
major histocompatibility complex class I heavy chains are
translocated into the cytoplasm and degraded by the proteasome.
Proc. Natl Acad. Sci. USA 94:1896.
Walter, P. and Johnson, A. E. 1994. Signal sequence recognition
and protein targeting to the endoplasmic reticulum membrane.
Annu. Rev. Cell Biol. 10:87.
Lee, N., Malacko, A. R., Ishitani, A., Chen, M. C., Bajorath, J.,
Marquardt, H. and Geraghty, D. E. 1995. The membrane-bound
and soluble forms of HLA-G bind identical sets of endogenous
peptides but differ with respect to TAP association. Immunity
3:591.
Lehner, P. J., Surman, M. J. and Cresswell, P. 1998. Soluble
tapasin restores MHC class I expression and function in the
tapasin-negative cell line 220. Immunity 8:221.