| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ß Relative to MHC Class I1
*
Department of Laboratory Medicine and Section of Immunobiology, and
Department of Genetics, Yale University School of Medicine, New Haven, CT 06520
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-MHC complex showed that one CD8 Ig domain provided the
majority of the contact with MHC class I and that residue R4 of that
domain contacted the 2 domain of MHC class I. We previously showed
by mutational analysis that this residue was critical for binding to
MHC class I. To determine which of the Ig domains for the
CD8ß heterodimer would make the most contact with class I
MHC, we expressed single-chain or dimeric forms of CD8 on COS-7 cells
and measured the adhesion of MHC class I positive cells. We found that
when one of the R4 residues was mutated in a CD8 homodimer
binding comparable to that of wild type was observed, whereas a
double R4 mutant severely impaired binding. However, when mutant CD8
(R4K) was coexpressed with wild-type CD8ß, binding was not observed.
These results support the model in which it is CD8, not CD8ß, that
is making the most of the contact with MHC class I, including the 2
domain. In addition, they demonstrate that a single-chain form of
CD8 can bind to MHC class I. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
ß heterodimer is
the major form present on thymocytes and mature T cells, while the
homodimer is exclusively present on subsets of NK cells and
intraepithelial lymphocyte (IEL) cells of the intestine (1, 2).
The interaction of CD8 with MHC class I has been studied
extensively. Salter et al. (3) identified a binding site for CD8 on a
negatively charged loop of the 3 domain of MHC class I, and we
identified critical residues located on the Ig domain of CD8 that were
important for this interaction (4, 5). As with other Ig molecules,
residues located on the surface containing the complementary
determining region
(CDR)3-like loops of CD8
were involved in recognition of MHC class I (5). However, unlike Ig
molecules, residues located on the A and B ß strands on the side of
the dimer were also found to be critical for this interaction (4).
These results allowed us to propose that while the residues located on
the CDR-like loops made contact with the negatively charged loop on the
3 domain of class I, amino acids on the side of the molecule would
interact with the MHC class I 2 domain. To support this, mutational
studies of the 2 domain were conducted and residues found that were
critical for CD8-MHC class I interaction (6). This model was further
supported by functional studies with murine/human chimeric MHC class I
(7).
The cocrystal structure of CD8/MHC class I confirmed that all
but one of the residues that we had identified to be critical
for the interaction with MHC class I by mutational analysis did, in
fact, make contact (8). It also demonstrated that CD8 interacted
with both the 2 and 3 domains of class I as well as with
ß2-microglobulin and, as we predicted, that CD8
made contact using residues from the A and B strands. Gao et al.
(8) also found that the contribution of the CD8 subunits to
the binding was asymmetric, with one domain contributing 70% of the
solvent-accessible area. While both subunits made contact with the 3
domain of class I (through their CDR-like loops), the subunit
designated -1 (Fig. 1
) made additional
contacts with the 2 domain of MHC class I and
ß2-microglobulin. Within the CD8-1 subunit, two
residues, R4 and L25, located on the A and B strands, respectively,
were shown to make contact with the 2 domain of MHC class I and
ß2-microglobulin, respectively. We had previously found
that both of these residues were critical for the interaction with MHC
class I as mutation of either severely reduced binding (4).
|
homodimer. However, since
the form predominantly expressed on T cells is the ß heterodimer,
we were interested in studying this interaction with MHC class I. Due
to the asymmetry in the binding of CD8 homodimers to MHC class I
described in the cocrystal (8), the contribution CD8ß makes to the
interaction with MHC class I depends on whether it adopts the -1 or
-2 subunit role. Based on modeling, Gao et al. (8) predicted
that CD8ß would function as the -2 subunit and, therefore, would
play a lesser role in the interaction.
To determine the orientation of CD8ß to MHC class I, we used a
cell-cell adhesion assay to study how mutated forms of the heterodimers
would interact with MHC class I. If CD8ß functioned as the -2
subunit, then coexpression with CD8 containing a mutation in a
residue involved in contacting the 2 domain of class I or
ß2-microglobulin (R4 or L25, respectively) would have a
dramatic effect on binding. However, if CD8ß functioned as the -1
subunit, then the R4 or L25 residues of CD8 would not contact MHC
class I, and coexpression with R4 or L25 mutant CD8 should have no
effect on binding. Results from this study demonstrate that there was
little or no binding of MHC class I positive cells to COS-7 cells
expressing mutant CD8 (R4 or L25) and wild-type ß, which was
similar to that observed with binding to COS-7 cells expressing the R4K
or L25A CD8 homodimers alone. However, cells expressing a dimer
of wild-type CD8 and mutant CD8 could bind MHC class I positive
cells at a level comparable to wild-type CD8. Therefore, one CD8
monomer that can make contact with 2 and
ß2-microglobulin is sufficient for binding. Thus, the
inability of wild-type CD8ß to compensate for the R4 or L25 mutation
in CD8 supports the model that in the CD8ß heterodimer CD8ß
is analogous to the -2 subunit.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Two CD8 Ig domains were linked via the carboxyl terminal of
one to the amino terminal of the other by means of a peptide spacer.
The second CD8 domain was linked to domains 3 and 4, together with
the transmembrane and cytoplasmic domains of CD4 that allowed the
chimeric protein to be expressed as a monomer. This approach has been
used previously in the expression of a chimeric CD48 (13). A peptide
spacer of 20 amino acids of 4 repeating units of GGGGS was determined
to be sufficient for the 2 Ig-like domains to adopt the correct
conformation using a computer graphic model of the CD8 homodimer
crystal structure. Fig. 2A
outlines the steps involved in synthesis of the single-chain construct.
Briefly, the spacer was synthesized in two stages, and the PCR products
of the individual -Ig domains were subcloned into the TA
cloning vector (Invitrogen, Carlsbad, CA). One StuI site was
deleted from the CD4 cDNA (position 934) using Kunkel mutagenesis,
which allowed us to use the other site that occurred at the
junction of domains 2 and 3 to subclone one CD8 domain. In addition,
this mutagenesis introduced XhoI and EcoRI sites
after the stop codon to facilitate future subcloning.
|
subunit was linked to the 5' end of the linker via the
introduced PstI site (a second PstI site is
located at the 5' end of CD8). All of the subcloning was conducted
in pBluescript II, then the entire construct was subcloned into the
expression vector pCDL-SR296 (14).
The various mutant constructs (Fig. 2
B) were synthesized by
Kunkel mutagenesis (15) of the wild-type single-chain construct in
pBluescript II. Different primers were used due to the 5' end of both
subunits being different (subunit 1 contains the leader sequence,
whereas subunit 2 has the linker peptide sequence 5'), which allowed us
to introduce mutations in either subunit independently. The mutation
introduced a DraI site that allowed us to confirm the number
of mutant sites within the entire construct. All constructs were
sequenced.
Expression of different forms of CD8 on COS-7 cells
COS-7 fibroblasts were transfected using a modification of a
previously described method (4). Briefly, an expression vector
containing a total of 2 µg of CD8 cDNA (single-chain CD8 WT
or mutants either alone or with CD8ß WT) was mixed with 8 µl of
lipofectamine per 100 µl of the serum-free medium Optimem (Life
Technologies, Grand Island, NY) for 30 min at room temperature. The
lipofectamine mixture was then added to several 35-mm dishes of nearly
confluent COS-7 cells. Transfection was stopped after 18 h by
replacing the lipofection mix with 2 ml of fresh medium containing 10%
FCS. Cells were fed again after 24 h, and, after an additional
24 h, the dishes were either analyzed for cell surface expression
of CD8 or used in the adhesion assay.
FACS analysis of cell surface molecules
For the cell surface expression of CD8 (including
single-chain constructs) and CD8ß, COS-7 transfectants were
stained with primary mAbs OKT8 (Coulter, Westbrook, ME), which
recognizes an epitope on CD8, and 2ST85H7 (Immunotech, Westbrook,
ME), which recognizes a conformational epitope formed by CD8ß.
Adhesion assay of class I positive cells to CD8 transfected COS-7
The assay used was a modification of the method previously
described in detail (4). CD8 transfectants were tested for their
ability to bind to an MHC class I positive B cell line, UC (16). The
HLA expressed by the UC cells were A1, A2, B5, B57, Cw4, DR7, DQ2, and
DQ3. These cells constitutively expressed the firefly luciferase gene
under the control of the CD8 promoter (17), a property that was
utilized to measure binding. Transfected COS-7 cells were washed once
with PBS, and 107 UC cells were added to each 35-mm dish.
The cells were incubated for 1 h at 37°C, and the UC cells were
aspirated off. After several washes, the number of bound cells was
determined by measuring the amount of luciferase activity in the cell
extract.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
ß
heterodimer, whether the CD8 or CD8ß Ig domain was making the
majority of contact with MHC class I. We first tested whether a
CD8 homodimer with only one functional R4 residue could bind to
MHC class I, as predicted, from the cocrystal structure of
CD8/MHC class I. In this structure, only one CD8 subunit made
contact with the 2 domain and ß2-microglobulin of MHC
class I (e.g., through residues R4 or L25, respectively). To test this,
we made a single-chain CD8 in which we could precisely control
the composition of the Ig domains in the dimer. If we had simply
cotransfected wild-type and mutant forms of CD8, a mixture of dimers
on the cell surface (WT/WT,
MT/WT,
WT/MT,
MT/MT) would have resulted. Four
different single-chain constructs were made (Fig. 2B),
consisting of either two wild-type CD8 Ig domains, two mixed CD8
homodimers consisting of one wild-type Ig domain and one with
the R4K mutation (in either domain), or a double mutant in which both
CD8 Ig domains contained the R4K mutation. The spacer used to link
the two Ig domains of CD8 consisted of four repeating units of GGGGS,
as these residues have previously been shown to provide a linker with
good flexibility (18, 19, 20). The constructs were expressed as a chimeric
protein linked to CD4 without domains 1 and 2.
Single-chain CD8 can support binding of MHC class I positive
cells
Expression of a single-chain CD8 was tested by
transfection into COS-7 cells. The cells were stained with a panel of
four Abs against CD8: OKT8, G10.1, Leu-2a, and 66.2. All Abs tested
were found to bind in a manner similar to the wild-type homodimer
(results not shown), indicating that no major structural changes had
occurred in the Ig domains. Mutant forms of the single chain were used
in a cell-cell binding assay in which we compared the binding of class
I positive cells to COS-7 cells expressing either wild-type single
chain, single chain with one wild-type domain and one mutant domain in
which R4 was mutated to a lysine (R4K), or a double mutant as outlined
in Fig. 2B. Expression of all forms of the single chain were
similar (Fig. 3A); therefore,
any differences in binding could not be attributed to differences in
expression levels. Results shown in Fig. 3B demonstrate that
the only construct that did not bind to class I positive cells
was the double mutant form of single-chain CD8. On the other hand, a
single mutant CD8 in either position in the single chain had no
major effect on binding. Therefore, one functional R4 residue is
sufficient for the binding of class I positive cells in this assay.
|
, we cotransfected cells with equal amounts of wild-type
CD8 and mutant (R4K) CD8 and compared the binding to cells
expressing either wild-type CD8 or double mutant CD8 alone.
The expression of the different CD8 dimers was similar, as shown
in Fig. 4A. Cotransfection of
wild-type CD8 and mutant CD8 results in a number of different
dimers on the cell surface: WT/WT, R4K/WT, WT/R4K and R4K/R4K. From the
single-chain results we would predict that only one of these forms
(R4K/R4K) would not be able to bind to class I. Therefore, we
expected the binding of the mutant/wild-type mix in this case to be
75% of wild-type homodimer (Fig. 4B). The binding of
class I positive cells to COS-7 cells expressing R4K/WT CD8 was
found to be 69 ± 12.9% of that of wild-type CD8
homodimers.
|
,
and not the CD8ß, domain
To determine the contribution that CD8ß makes to the
interaction, we tested the possibility that CD8ß would make similar
contacts as the CD8-2 subunit. To do this, we cotransfected
wild-type CD8ß with either R4K or L25A CD8. Cell surface
expression of CD8 was tested with two Abs, one against CD8 and the
other against an epitope formed by CD8ß. While CD8ß cannot come
to the surface on it own (21), it is possible that CD8 homodimers
and CD8ß heterodimers could be expressed on the cell surface.
Therefore, staining of both CD8 and CD8ß was compared to
ensure that the predominant form of CD8 expressed on the cell surface
was that of the CD8ß heterodimer, a finding that has previously
been quantified by Scatchard analysis (11). Fig. 5A demonstrates that this is
the case, as the level of CD8 expression is similar to that of
CD8ß. In addition, the level of expression for wild-type CD8ß
is similar to that of R4K/CD8ß and L25A/CD8ß, indicating that
differences in binding are not due to differences in expression level.
In the cell-cell adhesion assay, our prediction was that the binding
would only be reduced if CD8ß was not contacting the 2 domain of
class I. As shown in Fig. 5B, binding of class I positive
cells to COS-7 cells expressing R4K CD8/WT CD8ß or L25A CD8/WT
CD8ß was the same as to cells expressing only double mutant CD8
homodimers (i.e., little or no binding). Therefore, binding was not
rescued in the presence of wild-type CD8ß.
|
homodimers had a dominant effect on binding, wild-type CD8 was added
to the mix of wild-type CD8ß and mutant CD8 DNA that was
transfected into COS-7 cells. In this case, we were able to detect
binding, even though some mutant CD8 homodimers would still be
present. Fig. 4A shows the expression levels of the various
transfectants, which indicate that the levels of expression were
similar and, therefore, any differences in binding could not be
attributed to variations in expression level. Binding of CD8
WT/R4K/CD8ß was 46 ± 5% of wild-type CD8ß, as shown in
Fig. 4B. This is what we would predict since only half of
CD8ß expressed on the cell surface would be capable of binding
to MHC class I if CD8ß could not contact the 2 domain of
MHC class I. |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
form of CD8. However, there is a great
deal of interest in studying the interactions of CD8ß with other
proteins since this is the major form expressed on class I-restricted,
mature T cells. The interaction of CD8 with MHC class I occurs when CD8
functions either as a coreceptor with the TCR or independent of the TCR
as in the veto effect (22, 23). The cocrystal structure of
CD8-MHC class I showed that the interaction with CD8 was
asymmetric with one Ig domain contributing 70% of the surface area
buried by CD8 on complex formation. This domain makes contact
with residues located in the 2 and 3 domains of MHC class I, as
well as with ß2-microglobulin. Assuming that CD8ß
would bind in a similar manner, we used a mutational approach to
determine whether CD8 or CD8ß made most of the contact with MHC
class I. Based on our previous mutational studies and the cocrystal
structure, we mutated residues on the CD8 Ig domain (R4 and L25)
that could interact with the 2 domain of MHC class I and
ß2-microglobulin, respectively, and we determined their
effect when coexpressed with wild-type CD8ß. The consequences of
these mutations would be a severe reduction of CD8ß/MHC class I
binding or no effect, depending on the orientation of the CD8
Ig domain relative to the CD8ß Ig domain. We found the former to be
the case, and, therefore, our work supports the model in which the
CD8 Ig domain is contacting both the 2 and 3 domains of MHC
class I, whereas the CD8ß Ig domain is only contacting the 3
domain.
This conclusion is consistent with respect to the previously published
studies on CD8-MHC class I interaction. Biacore studies by Garcia et
al. (12) found no major difference in affinity of CD8 or
CD8ß for MHC class I, and Sun et al. (11) demonstrated that CD8ß
did not enhance the association of CD8 for MHC class I as had been
speculated. One might expect the affinities of CD8 vs CD8ß
to be similar if it were the CD8 Ig domain for both homodimers and
heterodimers making the most contact. It is known that CD8ß
functions as a better coreceptor (9, 10), a finding that in part was
attributed to stronger affinity for MHC class I; however, this is
unlikely to be the case.
We found that the binding of R4K/CD8 single chain was slightly
higher than that of CD8/R4K. This may be due to the different
structural constraints of the two Ig domains of the single chain. The
Ig domain at the N terminus of the protein that is connected to the
rest of the protein by a peptide linker may be more flexible and could
more readily adopt the -2 role and bind at a position further from
the membrane. The other Ig domain linked to domain 3 of CD4 appears to
be the one more likely to make contact with both the 2 and 3
domains of MHC class I because it is linked to CD4, and, to adopt the
-2 subunit role, it would have to "reach" further from the cell
membrane. An alternative, although not mutually exclusive,
hypothesis may be the effect of the different position of the
linker peptide relative to MHC class I when the different subunits make
contact with the 2 domain of MHC class I (i.e., the linker peptide
may be more likely to interfere in one position vs another).
There are multiple approaches for studying the orientation of two molecules relative to each other. For instance, to study TCR-MHC interaction, cocrystals were generated (24, 25), and mutations of the TCR (26) and peptide (27) have helped to elucidate the orientation of the TCR-MHC/peptide interaction. In addition, complementary mutations in CD2 and CD48 were made that restored binding or function in comparison to the affect of each mutation alone (28). These different approaches are complementary. Even a cocrystal represents the most efficient way that two molecules pack to form a crystal. This may not completely reflect what is happening on the cell surface, particularly if there are other molecules involved in the interaction that are not part of the cocrystal.
The orientation of CD8ß to MHC class I supported by the data we
have generated has implications for potential binding between CD8 and
the TCR. Evidence for association between coreceptor and the TCR was
initially described for CD4 based on cocapping studies (29, 30). More
recent cocapping studies with CD8 and the TCR showed that
anti-CD8ß Abs were significantly more efficient than
anti-CD8 Abs at inducing cocapping of the TCR (31). This
suggested to Lim et al. (31) that Abs to the CD8ß polypeptide
may preferentially promote a conformation of CD8 that stabilizes an
association with the TCR, independent of their binding to MHC (31). In
designing experiments to test this hypothesis, knowing the orientation
of CD8ß to the TCR is helpful. Our results provide strong support
for the orientation of CD8ß relative to MHC class I, with the
CD8 Ig domain providing the majority of the contact. The CD8ß
stalk is shorter and yet it must reach further from the membrane to
contact MHC class I. As a result, this may constrain the orientation of
CD8 and the TCR facilitating interaction between the CD8 associated
tyrosine kinase p56lck and the TCR 
-chain.
|
Acknowledgments |
|---|
, and Victoria Aitken
for help with all other figures. |
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Paula B. Kavathas, Department of Laboratory Medicine, Yale University School of Medicine, 333 Cedar Street, P.O. Box 208035, New Haven, CT 06520-8035. E-mail address:
3 Abbreviations used in this paper: CDR, complementary determining region; WT, wild type; MT, mutant; IEL, intraepithelial lymphocyte.
Received for publication July 6, 1998. Accepted for publication September 22, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and CD8ß chains. Tissue Antigens 35:82.[Medline]
3 domain of HLA-A2. Nature 345:41.[Medline]
/ homodimer in binding to major histo-compatibility complex class I molecules: support for a bivalent model. Proc. Natl. Acad. Sci. USA 91:1716.2 and 3 domains of MHC class I. J. Exp. Med. 182:1275. and HLA-A2. Nature 387:630.[Medline]
and CD8ß in interaction with MHC class I. J. Immunol. 159:6077.[Abstract]
promoter: an efficient and versatile mammalian cDNA expression system composed of the simian virus 40 early promoter and the R-U5 segment of human T-cell leukemia virus type 1 long terminal repeat. Mol. Cell. Biol. 8:466. promoter. J. Immunol. 150:4376.[Abstract]
chain on stimulator cells. Int. Immunol. 2:879.ß T cell receptor structure at 2.5 A and its orientation in the TCR-MHC complex. Science 274:209.This article has been cited by other articles:
|
|
 |
|
  J. Huang, L. J. Edwards, B. D. Evavold, and C. Zhu Kinetics of MHC-CD8 Interaction at the T Cell Membrane J. Immunol., December 1, 2007; 179(11): 7653 - 7662. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. A. Benoit and R. Tan Xenogeneic beta2-Microglobulin Substitution Affects Functional Binding of MHC Class I Molecules by CD8+ T Cells J. Immunol., September 15, 2007; 179(6): 3588 - 3595. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. E. Lyons, T. Moore, N. Brasic, M. Li, J. J. Roszkowski, and M. I. Nishimura Influence of Human CD8 on Antigen Recognition by T-Cell Receptor-Transduced Cells Cancer Res., December 1, 2006; 66(23): 11455 - 11461. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H.-C. Chang, K. Tan, and Y.-M. Hsu CD8{alpha}beta Has Two Distinct Binding Modes of Interaction with Peptide-Major Histocompatibility Complex Class I J. Biol. Chem., September 22, 2006; 281(38): 28090 - 28096. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. S. Wong, X. Wang, T. Witte, L. Nie, N. Carvou, P. Kern, and H.-C. Chang Stalk Region of {beta}-Chain Enhances the Coreceptor Function of CD8 J. Immunol., July 15, 2003; 171(2): 867 - 874. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. L. Hutchinson, L. Wooldridge, S. Tafuro, B. Laugel, M. Glick, J. M. Boulter, B. K. Jakobsen, D. A. Price, and A. K. Sewell The CD8 T Cell Coreceptor Exhibits Disproportionate Biological Activity at Extremely Low Binding Affinities J. Biol. Chem., June 27, 2003; 278(27): 24285 - 24293. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Devine, L. Rogozinski, O. V. Naidenko, H. Cheroutre, and P. B. Kavathas The Complementarity-Determining Region-Like Loops of CD8{alpha} Interact Differently with {beta}2-Microglobulin of the Class I Molecules H-2Kb and Thymic Leukemia Antigen, While Similarly with Their {alpha}3 Domains J. Immunol., April 15, 2002; 168(8): 3881 - 3886. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Gaspar Jr., P. Bagossi, L. Bene, J. Matko, J. Szollosi, J. Tozser, L. Fesus, T. A. Waldmann, and S. Damjanovich Clustering of Class I HLA Oligomers with CD8 and TCR: Three-Dimensional Models Based on Fluorescence Resonance Energy Transfer and Crystallographic Data J. Immunol., April 15, 2001; 166(8): 5078 - 5086. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Devine, L. J. Kieffer, V. Aitken, and P. B. Kavathas Human CD8{beta}, But Not Mouse CD8{beta}, Can Be Expressed in the Absence of CD8{alpha} as a {beta}{beta} Homodimer J. Immunol., January 15, 2000; 164(2): 833 - 838. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Kern, R. E. Hussey, R. Spoerl, E. L. Reinherz, and H.-C. Chang Expression, Purification, and Functional Analysis of Murine Ectodomain Fragments of CD8alpha alpha and CD8alpha beta Dimers J. Biol. Chem., September 17, 1999; 274(38): 27237 - 27243. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
