| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
-Helical Residues1
*
Department of Genetics, Washington University School of Medicine, St. Louis, MO 63110; and
Department of Biochemistry and Biophysics, Texas A & M University, College Station, TX 77843
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-helices of the MHC molecule (1). Interaction
of the TCR with MHC residues has been shown to be important for
allele-specific recognition (2); however, CTL clones
specific for the same MHC molecule can differ in certain MHC residues
that they contact (3, 4, 5, 6). Because each MHC allele can bind
a unique repertoire of peptides, the precise set of peptides bound to
each MHC molecule also contributes to allele-specific recognition. Yet,
it remains unclear whether this functional recognition by TCRs varies
in terms of the relative dependence upon interaction with peptide vs
MHC residues. Recent evidence suggests that alloreactive T cells, like self-restricted T cells, recognize a complex of both MHC and endogenous peptide ligand, and that alloreactive T cells are capable of the same degree of specificity as self MHC-restricted T cells (2, 7, 8, 9, 10, 11). However, there has been speculation that alloreactive TCRs are more dependent on interactions with MHC, rather than peptide residues, as compared with self MHC-restricted TCRs (12, 13, 14). Whereas T cells with high affinity for self MHC molecules will be deleted during development, a T cell with a high inherent affinity for an allo-MHC molecule does not undergo selection on this allo-MHC molecule, and thus can mature. As a consequence, alloreactive TCRs might be more biased toward interactions with MHC residues than syngeneic TCRs. Recent studies have established that TCRs do have an inherent reactivity toward MHC molecules (15, 16), and studies examining the effects of changes in peptide residues on recognition by a given T cell have reported a greater tolerance for changes in peptide residues with allo-, rather than self MHC ligands (13, 14). This has been used to argue that the recognition of an allogeneic ligand must be more dependent upon interaction with MHC residues than peptide residues. However, there have been no studies that have directly addressed the relative MHC dependence of allogeneic vs syngeneic TCRs.
Examination of how TCRs interact with their MHC/peptide ligands has led to the proposal that there is a common pattern for TCR recognition of a given MHC molecule and that TCRs may interact with class I and class II MHC proteins in this same orientation (5, 17). The solution of the structures of four individual TCR-MHC/peptide complexes, in addition to a functional study of TCR interaction with Kb, have led to the suggestion that this common orientation is parallel to the ß-pleated sheets of the MHC molecule (5, 18, 19, 20, 21). However, subtle differences in the angle of this interaction could occur if the residues contacted by TCRs differ (20, 21), and patterns of TCR-MHC engagement distinct from this proposed common orientation have been identified (22, 23). Thus, it is not clear whether all TCRs engage all MHC molecules in this same common orientation, or whether allele-specific recognition of MHC molecules is in part a result of distinct orientations of TCR engagement.
In this study, using an extensive panel of site-directed mutants of
Ld together with a large panel of CTL clones, we
determine which MHC class I residues are important for TCR recognition
by Ld-restricted and
Ld-alloreactive CTL clones that are specific for
the same and different peptides. Our approach is different from earlier
studies that compared self-restricted vs allorecognition of different
MHC alleles by a single T cell clone. In contrast to these earlier
studies, we find that alloreactive CTL clones are no more
cross-reactive with the Ld mutants than
Ld-restricted clones, indicating that allogeneic
and syngeneic TCRs are equally dependent on interaction with MHC
residues. Importantly, the same primary MHC contact residues are used
by self-restricted and alloreactive T cells. A high affinity clone, 2C,
interacts with the same primary contact residues, but is less affected
by mutation of Ld, suggesting that dependence on
interaction with MHC residues can be overcome by high affinity. This
analysis reveals an allele-specific recognition pattern of
Ld with greater dependence upon interaction with
residues in the center of both -helices of Ld
than what has been observed with other class I molecules, yet is
consistent with a common diagonal orientation of TCR-MHC/peptide
engagement.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
BALB/c (H-2d), BALB/c-H-2dm2 (dm2, Ld loss mutant), and DBA/2 (H-2d) mice were obtained from The Jackson Laboratory (Bar Harbor, ME) or Charles River Laboratories (Wilmington, MA) and were housed and bred in the barrier animal facility at Washington University School of Medicine (St. Louis, MO).
Peptides
The amino acid sequence of the murine CMV
(MCMV)4 peptide
corresponds to residues 168–176 (YPHFMPTNL) of the MCMV immediate
early protein pp89 (24). The amino acid sequence of the
tum- peptide corresponds to residues 14–22
(TQNHRALDL) of the mutant protein P91A- (exon 4)
from the tum- P815 variant (25).
The p2Ca and QL9 peptides are both derived from the endogenous
mitochondrial protein -ketoglutarate dehydrogenase, and the
sequences are LSPFPFDL and QLSPFPFDL, respectively (8, 12, 26). Peptides were synthesized using Merrifield’s solid-phase
method (27) on a peptide synthesizer (model 431A; Applied
Biosystems, Foster City, CA). Peptides were purified (>95%) by
reverse-phase HPLC, and purity was assessed, as described
(28).
Monoclonal Abs
28-14-8S is a mAb of the mouse IgG2a isotype specific for the
3 domain of Ld and was used for detection of
Ld and the Ld mutants
(29). The following mAbs were used in flow-cytometric
studies to establish the TCR Vß usage of our CTL clones. F23.1 is a
mouse IgG2a mAb that is specific for the mouse Vß8.1, Vß8.2, and
Vß8.3 regions (30). MR9-4, a mouse mAb specific for the
Vß5.1 and Vß5.2 regions; RR4-7, a rat mAb specific for the Vß6
region; and MR12-3, a mouse mAb specific for the Vß13 region were
generous gifts from Osami Kanagawa (Washington University). 1B2 is a
clonotypic mouse mAb specific for the 2C TCR (31).
Cell lines
DAP-3 is the murine Ltk- fibroblast cell line (H-2k), and L-Ld was generated by introducing the Ld gene into DAP-3 cells (32). LM1.8, a kind gift from Phillipe Kourilsky (INSERM, Institut Pasteur, Paris, France), is the L cell (H-2k) fibroblast cell line transfected with ICAM-1 under HAT selection. LM1.8-Ld is the LM1.8 cell line transfected with the Ld cDNA under G418 resistance. Cell lines were maintained at 37°C, 5% CO2 in RPMI 1640 medium (Life Technologies, Grand Island, NY) supplemented with 2 mM L-glutamine, 10% (v/v) bovine calf serum (HyClone Laboratories, Logan, UT), 0.1 mM nonessential amino acids, 1.25 mM HEPES, 1 mM sodium pyruvate, and 100 U/ml penicillin/streptomycin. Transfected cell lines were grown in medium containing 0.6 mg/ml G418 (Geneticin; Life Technologies, Grand Island, NY).
CTL clones
The Ld-alloreactive CTL clone 2C, generated from a BALB.B (H-2b) mouse that had been immunized with H-2d cells, and shown to specifically recognize Ld (33), was a generous gift from Herman Eisen (MIT, Cambridge, MA). The Ld-alloreactive clone, L3 (34), was provided by Matt Thomas (Washington University). The clone 42F3 was derived in vitro from a dm2 anti-BALB/c response and was shown to be specific for Ld+ p2Ca, and Vß8+, but 1B2- (35). The Ld-alloreactive, MCMV-specific CTL clone 2.11.2 and the Ld-restricted, tum--specific CTL clones P15 and P24 were previously described (2, 36, 37). All clones were maintained in 24-well plates in sensitization medium (RPMI 1640 medium supplemented with L-glutamine, sodium pyruvate, nonessential amino acids, 100 U/ml penicillin/streptomycin, 50 µM 2-ME, and 10% (v/v) FCS (HyClone Laboratories)) and stimulated weekly with 5 x 106 irradiated (2000 rad) BALB/c splenocytes/ml and 10 U/ml murine rIL-2 (Biosource, Camarillo, CA). Peptide-specific clones were supplemented with 10-5 M peptide.
The MCMV-specific, Ld-alloreactive clone 2.3.3 was generated in vitro in the presence of the MCMV peptide, as described (2). Clones 1C2, 1A1F7, and 1A1G9 were generated using dm2 mice that had been primed with BALB/c splenocytes in vivo. Two weeks after priming, spleens from the mice were removed and the splenocytes were isolated. A total of 7.5 x 106 responding dm2 splenocytes/well was cocultured with 3.5 x 106 BALB/c splenocytes/well (irradiated at 2000 rad) in 24-well Linbro plates (Flow Laboratories, ICN, Horsham, PA) containing 2 ml sensitization medium. After incubation for 5 days at 37°C and 5% CO2, effector cells were analyzed for Ld specificity. For the generation of CTL clones, effector cells were resuspended in fresh sensitization medium and cloned by limiting dilution into 96-well, round-bottom plates in the presence of 2.5 x 106 irradiated BALB/c splenocytes/ml and 10 U/ml rIL-2. The clones were restimulated weekly by replacing 100 µl of the medium with fresh medium containing 5 x 106 irradiated BALB/c splenocytes/ml and 10 U/ml rIL-2. Clones were selected for recognition of Ld-expressing targets and maintained by weekly restimulation in 24-well plates with 0.5–1 x 106 T cells/well, 5 x 106 irradiated splenocytes/well, and 10 U/ml rIL-2.
The generation of the tum-- and MCMV-specific CTL clones was performed as previously described (38). The MCMV-specific clones 8, C5, D7, and A8 were generated from BALB/c splenocytes, while the clones IF5, 1C6, and 2C4 were generated from DBA/2 splenocytes. The tum--specific clones ID3, IC10, and IG10 were generated from DBA/2 splenocytes.
Generation of the Ld mutants
The Ld mutants were made by PCR from an Ld cDNA template, and sequenced to confirm the presence of the mutation and the fidelity of the polymerase reaction. Each Ld mutant was cloned into the expression vector RSV.5.neo (39), and the constructs were transfected, following the instructions provided by the manufacturer, into LM1.8 cells using Lipofectin (Life Technologies, Gaithersburg, MD).
Flow cytometry and peptide induction
Cells were washed and incubated on ice in PBS containing 0.2% BSA in the presence of a saturating concentration of mAb, or PBS + 0.2% BSA alone, for 30 min, washed twice, and incubated with a saturating concentration of fluorescein-conjugated, Fc-specific, affinity-purified F(ab')2 fragment of goat anti-mouse, or goat anti-rat, IgG (Organon-Teknika Cappel, West Chester, PA) for 30 min on ice. Cells were washed twice and resuspended, and the viable cells, gated by forward and side light scatter, were analyzed on a FACScan flow cytometer (Becton Dickinson, Mountain View, CA). Mean fluorescence values were converted from logarithmic amplification by linear regression analysis using the CellQuest 30 software (Becton Dickinson).
For peptide inductions, cells were cultured overnight at 37°C in the presence or absence of peptide in RPMI medium (Life Technologies, Grand Island, NY) supplemented with 10% BCS (HyClone, Logan, UT), 2 mM glutamine, 1 mM sodium pyruvate, 0.1 mM nonessential amino acids, 1.25 mM HEPES, and 100 U/ml penicillin/streptomycin. After incubation, the cells were harvested, labeled, and analyzed, as described above.
51Cr release assay
A total of 1 x 106 target cells was labeled for 1 h with 150–200 µCi of 51Cr (Na51CrO4, NEN, Boston, MA; 1 Ci = 37 GBq) in 200 µl of RPMI 1640 medium + 10% bovine calf serum at 37°C in 5% CO2. Effector cells were plated at various concentrations into 96-well microtiter plates, and 2.5 x 103 or 5 x 103 washed target cells per well were added. In some experiments, the targets were resuspended in medium containing peptide before plating. The plates were centrifuged at 50 x g for 1 min and incubated for 4 h at 37°C in 5% CO2. Radioactivity in 100 µl of supernatant was measured in an Isomedic gamma counter (ICN Biomedicals, Huntsville, AL). The mean of triplicate samples was calculated, and percentage 51Cr release was determined according to the following equation: percentage 51Cr release = 100 x ((experimental 51Cr release - control 51Cr release)/(maximum 51Cr release - control 51Cr release)), where experimental 51Cr release represents counts from target cells mixed with effector cells; control 51Cr release represents counts from target cells incubated in medium alone (spontaneous release); and maximum 51Cr release represents counts from target cells lysed with 5% Triton X-100. Spontaneous release ranged from 8–20% of maximum release.
CTL assays were performed on day 4 after clone restimulation to
determine the lowest E:T ratio that gave maximal lysis on LM1.8 cells
expressing wild-type Ld. This was done to ensure
strong lysis by all of the clones and to avoid the effect of a mutation
being masked by too high of an E:T ratio. On day 5 after clone
restimulation, this E:T ratio was used for CTL assays with the entire
panel of target cell lines. All of the clones gave strong lysis of
wild-type LM1.8-Ld (mean specific lysis 48
± 2% (SEM)). For the data in Fig. 2
, for a given clone, the lysis of
each Ld mutant was calculated relative to the
lysis of wild-type Ld. The lysis of mutant
Ld by any one clone was divided by the lysis of
wild-type Ld by the same clone (x 100) to obtain
the relative lysis of that clone on the mutant.
|
A model of the Ld/MCMV complex was built
using the refined coordinates of the Ld/P29
complex (40), where residues P3(N), P4(V), P5(N), P6(I),
P7(H), and P9(F) of the P29 peptide (YPNVNIHNF) were replaced with
those of the MCMV peptide (YPHFMPTNL). The residues of the
Ld heavy and light chains were not tampered with
during the modeling. The atomic interactions between the heavy chain
atoms and the MCMV peptide atoms were manually checked for close
contacts. Molecular dynamics calculations and a 200-step conjugate
gradient energy minimization were performed using DISCOVER module in
INSIGHT (Molecular Simulation, San Diego, CA). Fig. 5 was generated
using the energy-minimized model of Ld/MCMV and
the program SETOR (41). The model was stable during the
molecular dynamics simulations and reached equilibrium on the basis of
the leveling with time of the root-mean-square deviation (1.11
angstrom) of the C atom positions. The proline at P6 of the MCMV
peptide in the energy-minimized conformation adopts the same
orientation as the proline at P6 of QL9 in the energy-minimized
Ld/QL9 model of Speir et al.
(42).
|
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To determine which residues of the MHC class I molecule
H-2Ld are important for functional interaction
with TCRs, amino acid residues of both the 1 and 2 domains of
Ld, predicted to interact with TCRs, were
mutated. Of the 18 Ld residues indicated in Fig. 1, 14 were mutated independently, and
residues 144, 145 and 155, 157 were mutated together. These residues,
with the exception of 107 on a loop on the 2 domain, span much of
the -helices, and point up and, thus, are predicted to interact
directly with TCRs (41, 43). Where possible, to minimize
gross alterations in the class I structure, the amino acid residue
found in Ld was changed to an amino acid residue
found in other class I molecules at that position. The nonconservative
amino acid substitutions made at each of these Ld
residues are shown in Table I.
Flow-cytometric analysis demonstrates that the cell lines transfected
with mutant Ld molecules express normal levels of
Ld, within a 4-fold range of the transfectant
expressing wild-type Ld,
LM1.8-Ld (Table II).
|
|
|
). This increase in expression of
Ld has been shown to result from peptide binding
to and stabilization of surface Ld molecules, and
thus a high proportion (
80%) of surface Ld
molecules is occupied with the exogenously added peptide ligand
(44, 45). This indicates that the mutant
Ld molecules clearly retain their ability to bind
these three different peptide ligands and that much of the surface
Ld is occupied with the added peptide. The
ability of all the mutant Ld molecules to bind
peptides together with their substantial surface expression validates
their use in functional analyses using peptide-specific CTL clones
reactive with Ld. Reactivity of self-Ld-restricted and alloreactive clones for the Ld mutants reveals a common recognition pattern
An extensive panel of CTL clones was used to determine whether
similar residues of Ld are involved in
recognition by self-Ld-restricted clones specific
for different peptides, as well as in recognition by alloreactive and
self-restricted clones specific for the same or different peptides. The
panel of independently derived CTL clones tested for recognition of the
Ld mutants includes eight
Ld-alloreactive clones, four of which have known
peptide specificity, and twelve Ld-restricted
clones (Table III). In addition, nine of
the clones are Vß8+, including alloreactive and
self-restricted clones with different peptide specificities. This
permits a comparison of the Ld residues
recognized by Vß8+ vs
Vß8- TCRs.
|
. The black boxes indicate a
dramatic reduction in T cell recognition, whereas the grey boxes
indicate an intermediate effect and white boxes indicate little or no
effect.
We find that mutations of the amino acid residues at positions 69, 72,
76, or 155/157 of Ld result in the most dramatic
effect on CTL recognition. Each of these mutations reduces CTL
recognition to less than 50% of wild-type Ld for
75% or more of the CTL clones. In fact, for many of the clones,
recognition is reduced to less than 25% of wild-type
Ld. Even though peptide is present at a high
concentration, and thus peptide occupancy is high, mutation of these
residues has a noticeable effect on recognition, suggesting that these
are major TCR contact points. In addition, for Ld
molecules mutated at residues 58, 82, 158, 162, or 166, recognition is
reduced to less than 50% of wild-type Ld for
50% or more of the clones. The importance of TCR interaction with each
of these amino acid residues, although clearly significant, appears to
vary depending upon the particular CTL clone. It is important to note
that a unique aspect of this analysis is that it includes both
Ld-restricted and
Ld-alloreactive CTL specific for the same
peptide, MCMV. Thus, a direct comparison can be made between allogeneic
and syngeneic CTL that recognize the same MHC/peptide complex. These
data reveal a recognition pattern of Ld by both
alloreactive and self-restricted T cells, which spans much of the 1
-helix and the middle of the 2 -helix toward the C-terminal
end, with the Ld residues most important for
recognition clustered in the middle of both -helices (Fig. 1). Thus,
the majority of clones analyzed show a common recognition pattern of
Ld that involves both relatively conserved (72,
76) and polymorphic (69, 155) amino acid residues.
Common MHC residues are recognized independent of the peptide ligand bound and independent of TCR Vß usage
Next we wanted to determine whether, in addition to common residues, there are differences in other residues that are involved in recognition of a specific MHC/peptide ligand. Given that the nature of the peptide bound to Ld can influence the conformation of the MHC/peptide complex (46), it might be expected that different MHC residues would be involved in TCR interaction depending on the sequence of the Ld-bound peptide. However, comparison of recognition of the Ld mutants by CTL clones specific for the MCMV peptide vs those specific for the tum- peptide shows a remarkable similarity. The most dramatic effect on CTL recognition is seen for mutation of the same key positions noted earlier. MHC residues 69, 72, 76, and 155/157 are important for recognition by both tum--specific and MCMV-specific clones, with 60% or more of each type affected by these mutations. There are some differences, namely the amino acids at positions 82 and 169 are more important for MCMV-specific clones than tum--specific clones. This suggests, although a larger sample size would need to be examined to verify this, that in addition to common MHC residues important for recognition of Ld (69, 72, 76, 155/157), the importance of other MHC residues for recognition may vary depending on the sequence of the peptide bound.
The preponderance of Vß8+ TCRs among Ld-alloreactive and Ld-restricted clones suggests that these TCRs may contact similar residues of Ld (35, 36, 47). However, our data show that the Ld residues important for interaction with TCRs are similar regardless of whether the CTL is Vß8+ or Vß8-. Differences between TCR-MHC contacts might be expected for TCRs with different Vß se- quences. Instead, comparing the nine clones that are known to be Vß8+ (2C, 42F3, D7, P15, P24, IG10, IF5, 1C6, and 2C4) with the six that are known to be Vß8- (1A1G9, 8, C5, ID3, IC10, and A8), there are only a few differences. Actually, our data demonstrate that clones that share usage of Vß8 show some differences in the Ld residues contacted, indicating similar TCRs can interact with distinct MHC residues. Because the TCR ß-chain contacts both peptide and MHC residues, as shown in crystal structure analyses (18, 20, 48), the precise MHC residues contacted by a given TCR most likely depend upon the specificity of the peptide bound to the MHC molecule.
Ld-alloreactive and Ld-restricted CTL clones show equivalent levels of cross-reactivity with the Ld mutants
To address whether allogeneic clones are biased toward interaction
with MHC residues compared with syngeneic clones, we assessed the
effect of the mutations on recognition by each type of clone. If
alloreactive clones are more MHC dependent, then they should recognize
fewer Ld mutants than the syngeneic clones.
However, both the Ld-alloreactive clones as well
as the Ld-restricted clones, on average,
recognize a similar number of the MHC mutants. The
Ld-alloreactive clones recognized 53% of the
mutants and the Ld-restricted clones recognized
48% of the mutants (Fig. 3). Thus, there
does not appear to be an increased requirement for interaction with MHC
amino acid residues by the Ld-alloreactive CTL
clones.
|
-helices; the amino acid
residue at position 82 is more important for Ld
recognition by alloreactive clones, while the residues at positions
144/145 and 162 are more important for the
Ld-restricted clones. For the remainder of the
Ld residues, a similar number of allogeneic and
syngeneic clones is affected. Together these data suggest that
Ld-restricted clones are as dependent on
interaction with MHC residues as Ld-alloreactive
clones. Ld residues important for recognition by 2C T cells are unmasked at lower peptide concentrations
The alloreactive clone, 2C, is not affected by any of the
Ld mutations at the high peptide concentration
used for analysis, whereas all of the other clones were affected by at
least some of the Ld mutations (Fig. 2). This
suggests that, unlike what has been reported (13), 2C
shows a decreased dependence upon interaction with MHC residues.
Another p2Ca-specific, Ld-alloreactive clone
(42F3) and two other peptide-specific,
Ld-alloreactive clones (2.3.3 and 2.11.2) are
affected by several of the Ld mutations. Thus,
the insensitivity of 2C to the Ld mutations does
not appear to be a common feature of peptide-specific,
Ld-alloreactive clones.
2C’s unique recognition pattern of Ld could
reflect a reduced dependence upon interaction with MHC residues due to
the high affinity interaction of the 2C TCR with ligand
(12), or 2C’s recognition pattern of
Ld may not include the mutated residues. To
distinguish between these two possibilities, we tested the ability of
2C to recognize those Ld mutants shown to have
the most dramatic effect on recognition by the other T cell clones,
over a wide range of peptide concentrations. Decreasing the peptide
concentration effectively lowers the ligand density and reduces the
avidity of the interaction. We observed that 2C recognizes wild-type
Ld and Q149A, which does not affect recognition
by most of the clones, in the absence of exogenous peptide (Fig. 4). However, 2C recognizes Q72A and
Y155H/R157K at high, but not lower peptide concentrations (Fig. 4).
This indicates that the 2C TCR recognizes similar MHC determinants as
other Ld-restricted and alloreactive clones, and
thus the common recognition pattern of Ld applies
also to 2C.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The most dramatic effects on T cell recognition of Ld include the mutation of residues at 155/157. Other studies, both structural and functional, have also implicated the importance of the class I MHC residue 155 for TCR recognition. These include recognition of Kb/dEV8 by 2C, and recognition of HLA-A2, HLA-B7, H-2Kd, and H-2Ld by different TCRs (2, 4, 6, 18, 20, 48, 49, 50, 51). Thus, position 155 serves as a principal anchor on most class I molecules, and may be required for docking the TCR on its MHC/peptide ligand. In addition to 155/157, mutation of residues at positions 69, 72, and 76 had the most dramatic effect on recognition of Ld, and residues at positions 58, 82, 158, 162, and 166 were also influential. It is important to note that the dramatic reduction in recognition of the Ld molecules mutated at these positions occurred at high peptide concentration, and thus could not be rescued at high ligand density. This underscores the importance of these residues for Ld recognition. Thus, these are the primary contact residues for Ld-reactive T cells.
The residues implicated in Ld interaction with
TCRs in our study are generally consistent with the diagonal
orientation of TCRs with MHC, first proposed in a study examining TCR
recognition of Kb (5). Based on the
effect of mutations at positions 80, 82, 158, 166, 167, and 174, it was
proposed that the TCR engages Kb/peptide parallel
to the MHC ß-pleated sheets and diagonal to the MHC -helices. This
model has more recently been supported by class II mutagenesis and
TCR-class I MHC crystallographic studies (17, 18, 19, 20, 21). Such a
common orientation for TCR-MHC engagement implies that TCR specificity
is achieved by TCR engagement of a unique combination of amino acid
residues, including polymorphic MHC residues and distinct peptide
residues.
Although our results are consistent with the diagonal model, there are
important differences that most likely dictate the allele specificity.
For example, the primary TCR contact residues at positions 69, 72, 76,
and 155/157 of Ld were not among the mutants that
inhibited recognition of Kb (5). As
shown in Fig. 1, the location of these residues on
Ld is distinct from the location of residues on
Kb implicated in TCR interaction, such that the
important Kb residues are skewed more to the
C-terminal regions of both -helices. Also, mutation of
Kb residues at positions 167 and 174 affected
recognition, while we did not see a dramatic effect on TCR recognition
of Ld by mutations at positions in this range
(positions 169 and 173). A difference between TCR interaction with
Ld vs Kb is also supported
by the crystal structure of the 2C TCR-Kb/dEV8
complex (48). In that study, the Kb
residue at position 69 is not contacted by the TCR, whereas in our
study, position 69 is a primary TCR contact residue for recognition of
Ld. Collectively, these findings suggest that
TCRs may engage Ld in a distinct orientation
compared with Kb, such that with
Ld the TCR is slightly more perpendicular to the
-helices (Fig. 1). However, independent of predicting orientation,
the reciprocal differences critical for TCR interaction with
Kb vs Ld most likely define
the structural basis of allotypic T cell recognition of these two class
I molecules.
Comparison of our study with the aforementioned
Kb mutagenesis study reveals significant
differences in the location of residues on Ld vs
Kb that are important for functional interaction
with their respective TCRs. However, it is difficult to adapt our data
to the recently published model predicting how the 2C TCR interacts
with Ld/QL9 (42). This model was
predicted from their structural resolution of Ld
loaded with heterogeneous peptides, and was based on resolution of the
structure of the 2C TCR with Kb/dEV8. It was also
assumed that most of the TCR interactions with MHC are through
conserved -helical residues shared by Kb and
Ld. However, our data indicate that both
conserved and polymorphic MHC residues are important for recognition of
Ld. In addition, we demonstrate that the primary
functional MHC interaction sites are the same for all TCRs reactive
with Ld, but differ between
Ld and Kb. Solution of a
crystal structure of Ld/peptide with TCR will
help to resolve this discrepancy.
To visualize the topology of the critical residues on the surface of
Ld/peptide complexes as seen by TCRs, we modeled
the Ld/MCMV structure based on our crystal
structure of Ld/P29 (41). We chose
to model the MCMV peptide into the Ld groove
because Ld-restricted and
Ld-alloreactive, MCMV-specific CTL were
characterized in this study and earlier experiments using
alanine-substituted peptides identified the P6 proline as the most
critical for CTL recognition (data not shown). Fig. 5 shows the orientation of the side
chains (the main chain carbonyl oxygen in the case of Gly) of the
residues we determined to be most critical for interaction of
Ld with TCRs, namely Gly69,
Gln72, Val76,
Tyr155, and Arg157. The
close proximity of the critical peptide side chain of the P6 proline
between the critical Ld side chains highlights
the importance of the central region of the MHC/peptide complex for TCR
interaction. The side chain of the P6 proline of MCMV points out of the
cleft, in part due to the hydrophobic ridge that transverses the center
of the Ld Ag-binding groove (41, 42). Importantly, the p2Ca and QL9 peptides have a proline in
the same position (P5 of the p2Ca octamer, P6 of the QL9 nonamer) that
is critical for CTL recognition (52, 53, 54). The side chains
of the P6 proline of QL9 and MCMV have a similar orientation when
modeled bound to Ld (42 and Fig. 5).
The fact that proline is the critical residue for TCR interaction with
both MCMV and QL9 is a unique feature of
Ld/peptide interactions with TCRs. Most
structure/function studies have identified a central charged residue in
the peptide that forms a salt bridge with the complementary charged
residue in complementarity-determining region 3 of the TCR (as first
shown by Jorgensen et al., in Ref. 22). Because proline
cannot form a salt bridge, its function as the most critical peptide
residue may be more a matter of shape complementarity.
As a corollary to the proposed difference in MHC dependency, it has also been speculated that self MHC-restricted TCRs are more peptide dependent than alloreactive TCRs (12, 13, 14). Recent studies have used peptide libraries to compare the peptide dependencies of individual T cell clones that display both self-restricted and alloreactive responses. For example, Brock et al. (13) reported that 2C recognition of self-Kb/SIYR-8 was highly peptide specific, whereas the allo-Ld/p2Ca response was very peptide degenerate. However, one recent study has shown that the recognition of self-Kb molecules by 2C T cells is peptide degenerate (55), and other studies have shown that 2C recognition of its allo-ligand is highly peptide specific (52, 53, 56). These approaches are limited by the use of a single T cell clone that detects different MHC molecules bound by different peptides with different anchor motifs. Thus, it is difficult to generalize from these studies in regard to the structural basis of allorecognition, and therefore the relative peptide dependency of 2C, and other alloreactive T cells, remains controversial. In this study, we have taken a reciprocal approach by comparing the MHC degeneracy of alloreactive vs self-restricted T cells. Another unique feature of our study is that all T cells compared were reactive with the same MHC molecule and several were specific for the same peptide.
The affinity of a TCR for MHC/peptide ligand can also influence whether
mutations in peptide and/or MHC residues affect T cell recognition.
Based on the higher affinity of the 2C TCR for its allo-ligand,
Ld/QL9, compared with the clone 4G3 for its self ligand,
Kb/SIINFEKL, it was speculated that alloreactive
TCRs may be of a higher affinity than self MHC-restricted TCRs
(12). A recent model of how Ld/QL9
interacts with the 2C TCR predicts that peptide, relative to class I,
contributes 33% of the binding energy and 46% of the contacts with
TCR (42, 57). This contribution of peptide to the binding
energy could explain why in our study 2C T cells were strikingly
tolerant to changes in Ld -helical residues at
high peptide concentrations in comparison with other
Ld-reactive T cells. However, when peptide was
limiting, 2C T cells displayed a similar pattern of recognition on the
panel of Ld mutants. This result is consistent
with 2C having a higher affinity for Ld than
other T cells, but sharing a common orientation when engaging
Ld. Given that 2C is atypical, at least among the
T cells we compared, generalizations regarding affinity of alloreactive
vs self-restricted T cells need to be confirmed using a wider panel of
CTL clones. However, regardless of whether or not there is an affinity
difference between alloreactive and self MHC-restricted TCRs, our study
definitively shows that alloreactive T cells are not more dependent on
MHC residues, but rather both types of T cells are comparably dependent
upon interaction with the same MHC -helical residues.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Current address: Eppley Institute for Research in Cancer and Allied Disease, University of Nebraska Medical Center, Omaha, NE 68198.
3 Address correspondence and reprint requests to Dr. Janet M. Connolly, Washington University School of Medicine, Department of Genetics, Box 8232, 660 South Euclid Avenue, St. Louis, MO 63110. E-mail address:
4 Abbreviation used in this paper: MCMV, murine CMV.
Received for publication April 8, 1999. Accepted for publication June 30, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-helices of the class I MHC molecule are simultaneously recognized by the T cell receptor. Cell 54:47.[Medline]
helices. J. Immunol. 158:2805.[Abstract]
ß T cell receptor structure at 2.5 Å and its orientation in the TCR-MHC complex. Science 274:209.3.2 selection in MHC class I mutant mice: evidence for an alternate orientation of TCR-MHC class I interaction. J. Immunol. 159:3322.[Abstract]
2 helix abolishes peptide recognition by T cells. Immunogenetics 34:141.[Medline]
and Vß T cell receptor chains. Mol. Immunol. 9:747.
-helices residues of the MHC class I molecule. J. Immunol. 161:148.ß T cell receptor: mapping the energy of antigen recognition. Immunity 8:413.[Medline]
This article has been cited by other articles:
|
|
 |
|
  L. Varani, A. J. Bankovich, C. W. Liu, L. A. Colf, L. L. Jones, D. M. Kranz, J. D. Puglisi, and K. C. Garcia Solution mapping of T cell receptor docking footprints on peptide-MHC PNAS, August 7, 2007; 104(32): 13080 - 13085. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Huang, S. Gilfillan, M. Cella, M. J. Miley, O. Lantz, L. Lybarger, D. H. Fremont, and T. H. Hansen Evidence for MR1 Antigen Presentation to Mucosal-associated Invariant T Cells J. Biol. Chem., June 3, 2005; 280(22): 21183 - 21193. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. K. Baxter, S. J. Gagnon, R. L. Davis-Harrison, J. C. Beck, A.-K. Binz, R. V. Turner, W. E. Biddison, and B. M. Baker Strategic Mutations in the Class I Major Histocompatibility Complex HLA-A2 Independently Affect Both Peptide Binding and T Cell Receptor Recognition J. Biol. Chem., July 9, 2004; 279(28): 29175 - 29184. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. J. Gagnon, Z. Wang, R. Turner, M. Damirjian, and W. E. Biddison MHC Recognition by Hapten-Specific HLA-A2-Restricted CD8+ CTL J. Immunol., September 1, 2003; 171(5): 2233 - 2241. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. M. C. Hornell, N. Myers, T. H. Hansen, and J. M. Connolly Homology Between an Alloantigen and a Self MHC Allele Calibrates the Avidity of the Alloreactive T Cell Repertoire Independent of TCR Affinity J. Immunol., May 1, 2003; 170(9): 4506 - 4514. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. A. Brehm, T. G. Markees, K. A. Daniels, D. L. Greiner, A. A. Rossini, and R. M. Welsh Direct Visualization of Cross-Reactive Effector and Memory Allo-Specific CD8 T Cells Generated in Response to Viral Infections J. Immunol., April 15, 2003; 170(8): 4077 - 4086. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. Wang, R. Turner, B. M. Baker, and W. E. Biddison MHC Allele-Specific Molecular Features Determine Peptide/HLA-A2 Conformations That Are Recognized by HLA-A2-Restricted T Cell Receptors J. Immunol., September 15, 2002; 169(6): 3146 - 3154. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. M. C. Hornell, S. M. Martin, N. B. Myers, and J. M. Connolly Peptide Length Variants p2Ca and QL9 Present Distinct Conformations to Ld-Specific T Cells J. Immunol., October 15, 2001; 167(8): 4207 - 4214. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. M. Baker, R. V. Turner, S. J. Gagnon, D. C. Wiley, and W. E. Biddison Identification of a Crucial Energetic Footprint on the {{alpha}}1 Helix of Human Histocompatibility Leukocyte Antigen (HLA)-A2 That Provides Functional Interactions for Recognition by Tax Peptide/HLA-A2-specific T Cell Receptors J. Exp. Med., February 26, 2001; 193(5): 551 - 562. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Garcia-Peydro, A. Paradela, J. P. Albar, and J. A. L. d. Castro Antagonism of Direct Alloreactivity of an HLA-B27-Specific CTL Clone by Altered Peptide Ligands of Its Natural Epitope J. Immunol., November 15, 2000; 165(10): 5680 - 5685. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.-P. Kovalik, N. Singh, S. K. Mendiratta, W. D. Martin, L. Ignatowicz, and L. Van Kaer The Alloreactive and Self-Restricted CD4+ T Cell Response Directed Against a Single MHC Class II/Peptide Combination J. Immunol., August 1, 2000; 165(3): 1285 - 1293. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. U.Y. Lee, H. R.O. Churchill, M. Daniels, S. C. Jameson, and D. M. Kranz Role of 2C T Cell Receptor Residues in the Binding of Self- and Allo-Major Histocompatibility Complexes J. Exp. Med., April 18, 2000; 191(8): 1355 - 1364. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
30o away from the 
