T cells are known to cross-react with diverse peptide MHC Ags through their αβ TCR. To explore the basis of such cross-reactivity, we examined the 2C TCR that recognizes two structurally distinct ligands, SIY-Kb and alloantigen QL9-Ld. In this study we characterized the cross-reactivity of several high-affinity 2C TCR variants that contained mutations only in the CDR3α loop. Two of the TCR lost their ability to cross-react with the reciprocal ligand (SIY-Kb), whereas another TCR (m67) maintained reactivity with both ligands. Crystal structures of four of the TCRs in complex with QL9-Ld showed that CDR1, CDR2, and CDR3β conformations and docking orientations were remarkably similar. Although the CDR3α loop of TCR m67 conferred a 2000-fold higher affinity for SIY-Kb, the TCR maintained the same docking angle on QL9-Ld as the 2C TCR. Thus, CDR3α dictated the affinity and level of cross-reactivity, yet it did so without affecting the conserved docking orientation.

The recognition of infected or transformed cells by CTLs requires the interaction of the αβ TCR with an antigenic peptide fragment bound to a product of the class I MHC. The specificity of this interaction is determined by the V domains of the TCR, which are structurally similar to Ab V domains (1). Each TCR V domain (α and β) contains three CDR loops that combine to form the TCR binding interface. Some general features of TCR-peptide MHC (pMHC)3 recognition have been defined by crystallographic studies, and all complexes examined to date are oriented such that the TCR Vα domain interacts with the C terminus of the peptide and the Vβ domain interacts with the N-terminal peptide residues (2, 3).

The ability of a single TCR to cross-react with different pMHC (sometimes called TCR degeneracy) is a common feature of T cell biology and is thought to be functionally important in mounting a rapid and effective response to diverse foreign Ags (4, 5). TCR cross-reactivity is also important in T cell selection processes, as a TCR must interact with a self-peptide to survive and be exported to the periphery (6). Finally, TCR cross-reactivity is important in the process that leads to transplant rejection through the recognition of a foreign MHC known as the alloantigen (7). Studies have linked degeneracy in TCR recognition to molecular flexibility at the binding interface, particularly in the CDR3 (8, 9). Analysis of x-ray crystal structures of TCR and TCR/pMHC complexes reveal substantial conformational adjustments that occur within the TCR binding interface upon pMHC binding (3, 10, 11, 12). Additionally, the high level of conformational mobility of the CDR3s is evident in the only nuclear magnetic resonance structure of an unliganded TCR, the class II-restricted D10 TCR (13).

Binding and kinetic studies of TCR-pMHC recognition have found that these interactions occur with low binding affinities ranging from ∼1 to 50 μM and that they have slow on-rates relative to other receptor-ligand interactions (14, 15), implying that some degree of conformational adjustment is required for binding to occur. Thermodynamic studies showing that TCR-pMHC interactions are highly temperature dependent and incur an entropic penalty upon binding due to stabilization of the interface further support the notion that a flexible interface is stabilized upon binding (15, 16, 17). However, some exceptions have recently suggested that a well ordered CDR3 in the unliganded state or a high degree of surface complementarity and the subsequent expulsion of water from the interface can compensate for such flexibility (18, 19). Conformational flexibility of TCR could represent a mechanism by which the immune system sacrifices high-affinity interactions to maintain a high level of degeneracy needed to maximize the recognition of potential pathogenic peptides.

Several studies of TCR cross-reactivity have focused on degenerate recognition of pMHC ligands that are similar in either peptide sequence (20) or MHC allele (21, 22), but less is known about TCR systems that cross-react with ligands that vary significantly in both the peptide sequence and MHC allele. The 2C TCR represents an excellent model system for examining TCR cross-reactivity because it recognizes two well-characterized and distinct pMHC ligands as strong agonists. The 2C clone was originally isolated as an alloreactive T cell that recognizes Ld as a foreign MHC (23). 2C binds to and is activated by octamer and nonomer peptide fragments of the enzyme α-ketoglutarate dehydrogenase, called p2Ca and QL9, respectively, when bound to Ld (24, 25). Additionally, 2C is positively selected on Kb, and a synthetic peptide called SIY acts as a strong agonist when bound to Kb (26). The Ld-binding peptide QL9 (QLSPFPFDL) and the Kb-binding peptide SIY (SIYRYYGL) share no sequence identity other than a C-terminal leucine, and although Ld and Kb are both murine class I MHC molecules, they differ by 31 residues in the α1α2 domains, making these two ligands for the 2C TCR quite diverse. Our recent structural analysis of 2C TCR in complex with QL9-Ld (27), and previous structural studies with 2C/SIY-Kb (28), showed that the TCR docks onto the Ld and Kb ligands with a different docking angle and interacts with distinct MHC residues in each case.

A panel of high-affinity TCR mutants was previously generated from a library containing random mutations of five residues in the CDR3α loop of 2C by selecting for enhanced binding to QL9-Ld or SIY-Kb (29, 30). In some cases, these mutants exhibited 2000-fold increases in affinity for the selecting pMHC ligand, and this was due primarily to an increase in the on-rate of the interaction. Kinetic analysis along with structural models of several mutants suggested that high affinity was achieved in part by increasing stabilizing interactions among the flexible CDR loops of the TCR (30). These findings suggested that high-affinity TCR might be less conformationally flexible, and thus less cross-reactive than their low-affinity counterparts. Studies of high-affinity TCR mutants provide a unique system to further examine the relationship between TCR cross-reactivity, affinity, and CDR loop flexibility.

In this study we examined the cross-reactivity of three high-affinity mutants of the 2C TCR: m6, which binds to QL9-Ld with an affinity of 10 nM (31), m13, which binds to QL9-Ld with an affinity of 116 nM, and m67, which binds to SIY-Kb with an affinity of 16 nM (30). Unexpectedly, we found that these high-affinity TCR, which are identical except for five residues of the CDR3α loop, vary significantly in their level of cross-reactivity. In the case of m67, cross-reactivity was relatively unaffected by mutations that generated a 2000-fold increase in affinity for SIY-Kb. Structural analysis of the m67/QL9-Ld complex revealed that the Ld docking orientation of the 2C TCR was maintained, even though the CDR3α and CDR1α loop conformations may be better suited for high-affinity Kb binding. Conversely, activation by the reciprocal pMHC, SIY-Kb, was completely abolished in both the m6 and m13 TCR. Structural analysis of m6 and m13 in the QL9-Ld complex suggested that increased stabilizing interactions between CDR loops in m6 and a rigid CDR3α loop containing three proline residues in m13 prevented the TCR from reorganizing to accommodate the structurally different SIY-Kb ligand. Thus, the CDR3 itself can dictate the affinity, the conformational rigidity of neighboring loops, and the level of cross-reactivity, although it can do so without affecting the conserved docking orientation of a particular TCR/pMHC interaction, which appears to be dictated largely by CDR1 and CDR2 interactions.

Ld-binding peptides QL9 (QLSPFPFDL) and MCMV (YPHFMPTNL), and Kb-binding peptides SIY (SIYRGYYL) and OVA (SIINFEKL) were synthesized by standard F-moc chemistry at the Macromolecular Core Facility at Penn State University College of Medicine (Hershey, PA) and purified by reverse phase HPLC with a linear acetonitrile gradient. Hamster anti-mouse-CD3 Ab 145-2C11 and anti-Ld Ab 30-5-7 were purified from ascites fluid. Rat anti-mouse IL-2 Ab JES6-1A12 and biotinylated rat anti-mouse IL-2 Ab JES6-5H4 were obtained from BD Pharmingen.

A stabilized mutant Ld module called Ld-m31 that contains only the α1 and α2 domains was expressed in BL21(DE3) Codon Plus Escherichia coli (Stratagene) in both biotinylated and nonbiotinylated forms and refolded in the presence of Ld-binding peptides (QL9 or MCMV) as previously described (32). Three additional mutations (F9Y, V12Y, and I23T) were added to Ld-m31 for use in crystallization (27). The Kb H chain containing a C-terminal biotinylation sequence was coexpressed in BL21(DE3) E. coli with BirA ligase to generate in vivo biotinylated Kb. Human β2-microglobulin was expressed in BL21(DE3) E. coli. Biotinylated Kb H chain and human β2-microglobulin were refolded in the presence of Kb-binding peptides (SIY or OVA) using a standard protocol (33) and purified by gel filtration.

An sc version of the 2C TCR (2C-T7)-containing six stability mutations (H47Yβ, L81Sβ, G17Eβ, L43Pα, W82Rα, and I118Nα; Ref. 34) was cloned into pET28a as an NcoI-EcoRI fragment in the Vβ-Vα orientation, with an N-terminal 6-histidine tag and the two domains separated by a 25 amino acid linker (SSADDAKKDAAKKDDAKKDDAKKDA). This construct was transformed into BL21(DE3) E. coli and was grown at 37°C in Luria-Bertani medium until the OD600 reached 1.0 and expression was induced with 1 mM isopropyl β-d-thiogalactoside for 2 h. Cells were harvested and lysed, and inclusion bodies were separated and washed. Soluble scTCR were refolded from inclusion bodies as previously described (35) and used in surface plasmon resonance (SPR) experiments. scTCR in the Vα-linker-Vβ orientation and containing an additional stability mutation in the Vβ domain (I75Tβ) were obtained by E. coli periplasmic expression and purified by osmotic shock, nickel affinity, and gel filtration chromatography for crystallographic and SPR experiments as described (27, 32). High-affinity mutant TCR m6, m13, and m67 contained the T7 stability mutations in addition to the CDR3α affinity mutations (30). Both forms of scTCR were used in SPR experiments, and similar values were obtained for binding parameters.

Binding affinities and kinetics were determined by SPR using a BIAcore 3000 instrument. An uncharged mutant of streptavidin called neutra-avidin (Pierce) was immobilized on a CM5 sensor chip using standard amine chemistry. Biotinylated pMHC ligands (SIY-Kb or QL9-Ld-m31) were immobilized to the neutra-avidin at levels ranging from 200 to 600 response units. Purified scTCR were injected over the surface at a flow rate of 30 μl/min at varying concentrations, depending on the affinity of the interaction. Binding to a control surface with immobilized null pMHC ligand (OVA-Kb or MCMV-Ld) was subtracted from all experiments, with the exception of the m6/QL9-Ld measurement. Because of the short lifetime of the Ld-binding peptides and the long lifetime of the m6/QL9-Ld interaction, surface degeneration did not allow for the use of immobilized MCMV-Ld-m31 as a control surface. Instead, binding to a blank flow cell was subtracted from the m6/QL9-Ld-m31 data, and 1 μM excess QL9 peptide was added to the scTCR samples to extend the lifetime on the surface. All kinetic data were fit to 1:1 Langmuir binding using simultaneous ka/kd determination with BIAevaluation 3.0 software. Equilibrium affinities were obtained from Scatchard analysis of the binding curves (36).

In several cases, soluble scTCR were also tested for the ability to inhibit the binding of biotinlyated m67 scTCR to SIY-Kb on the surface of cells. T2-Kb cells were incubated with 100 μM SIY or OVA peptide for 2 h at 37°C in supplemented RPMI 1640 medium and washed twice with PBS-0.5% BSA. Purified biotinylated m67 at a final concentration of 5 nM was mixed with either T7, m6, unbiotinylated m67, or an irrelevant Ab against Ld, and the mixture was added to cells and incubated at 4°C for 45 min. Cells were washed and stained with streptavidin/PE for 20 min at 4°C. After washing, fluorescence was detected by flow cytometry.

To examine binding of SIY-Kb directly to TCR-transduced 58−/− cells, selected CD8-negative cell lines were incubated with various concentrations of streptavidin/PE SIYR/Kb tetramer on ice for 3 h. After washing, cells were suspended in ice-cold PBS containing 1% BSA and 0.02% azide, and analyzed for bound fluorescent tetramers by flow cytometry.

T cell lines were generated previously by transfection of the 58−/− T cell hybridoma with the 2C TCR or high-affinity mutant m6, m13, or m67 TCR. To generate hybridoma lines expressing the CD8 coreceptor, CD8α and CD8β genes were cotransfected with the TCR genes (37). T cell transfectants were stimulated with varying concentrations of SIY peptide and T2-Kb cells or QL9 peptide and T2-Ld cells for 24 h. Concurrently, transfectants were activated with immobilized anti-CD3 Ab 2C11 for maximal release of IL-2 and with null peptide and T2 cells as a negative control. Activation of transfectants was monitored by IL-2 release as detected by ELISA (37). Results were plotted as percentage of maximal IL-2 release = [(sample A450 − null peptide A450)/(anti-CD3 A450 − null peptide A450)] × 100.

Purified TCR and pMHC were combined in equimolar amounts and concentrated in a YM10 Microcon unit (Millipore) to ∼15 mg/ml for both m67/QL9-Ld and m13/QL9-Ld. Crystals of the m67 complex were obtained in 2 M ammonium tartrate (pH 7.0) where nucleation was obtained by streak seeding from crystals of m6/QL9-Ld using the sitting drop vapor diffusion method. Crystals of the m67 complex were in a different space group and had different unit cell dimensions than the seed source. m13/QL9-Ld crystals were grown in 0.2 M ammonium citrate, 20% polyethylene glycol, m.w. 3350, and 3% trimethyl amine N-oxide dihydrate. Crystals were cryoprotected using 22% glycerol before cryo-cooling to 100 K. Data were collected on beamline 8.2.2 at the Advanced Light Source, Berkeley, CA or beamline 11.1 at the Stanford Synchrotron Light Source, Stanford, CA. Crystals of m67/QL9-Ld diffracted to 3.8 Å, and crystals of m13/QL9-Ld diffracted to 2.95 Å with cell dimensions and space group as listed in Table II. The data were indexed, integrated, and scaled with HKL2000 (38). Rmerge values for the m67 complex are slightly higher than average due to anisotropy in the data and weak diffraction. We rely on measures of bond lengths and angles to insure reliable data quality, because geometry for the m67 complex structure is better than structures of similar resolution.

Table II.

Data collection and refinement statistics

m13/QL9-Ldm67/QL9-Ld
Data collection   
 Space group C222 P6422 
 Cell dimensions   
  a, b, c (Å) 158.5, 160.5, 357.2 112.6, 112.6, 272.5 
  α, β, γ (°) 90, 90, 90 90, 90, 120 
 Resolution (Å) 50-2.95 (3.06-2.95) 50-3.8 (3.94-3.80) 
Rmerge 0.130 (0.656) 0.194 (0.366) 
 Average II 10.3 (2.0) 7.9 (3.3) 
 Completeness (%) 97.4 (97.3) 95.4 (85.7) 
 Redundancy 4.1 (4.1) 6.2 (5.2) 
   
Refinement   
 Resolution (Å) 50-2.95 50-3.8 
 No. reflections 93,313 10,227 
Rwork/Rfree 0.2498/0.2747 0.2227/0.2765 
 No. protein atoms 25,808 3,227 
 Average protein B-factors 47.95 67.80 
 Ramachandran   
  Core, allowed, generous, disallowed 87.3, 11.8, 0.8, 0.0 79.6, 19.0, 1.2, 0.3 
 Root mean square deviations   
  Bond lengths (Å) 0.008 0.0091 
  Bond angles (°) 1.465 1.245 
m13/QL9-Ldm67/QL9-Ld
Data collection   
 Space group C222 P6422 
 Cell dimensions   
  a, b, c (Å) 158.5, 160.5, 357.2 112.6, 112.6, 272.5 
  α, β, γ (°) 90, 90, 90 90, 90, 120 
 Resolution (Å) 50-2.95 (3.06-2.95) 50-3.8 (3.94-3.80) 
Rmerge 0.130 (0.656) 0.194 (0.366) 
 Average II 10.3 (2.0) 7.9 (3.3) 
 Completeness (%) 97.4 (97.3) 95.4 (85.7) 
 Redundancy 4.1 (4.1) 6.2 (5.2) 
   
Refinement   
 Resolution (Å) 50-2.95 50-3.8 
 No. reflections 93,313 10,227 
Rwork/Rfree 0.2498/0.2747 0.2227/0.2765 
 No. protein atoms 25,808 3,227 
 Average protein B-factors 47.95 67.80 
 Ramachandran   
  Core, allowed, generous, disallowed 87.3, 11.8, 0.8, 0.0 79.6, 19.0, 1.2, 0.3 
 Root mean square deviations   
  Bond lengths (Å) 0.008 0.0091 
  Bond angles (°) 1.465 1.245 

Complex structures were determined by molecular replacement with the program phaser (39). Individual components (MHC, then TCR) of Brookhaven Protein Data Bank (PDB) file 2OI9 were used as search models. One round of rigid body refinement was performed, followed by rounds of model building using COOT (40), simulated annealing, and positional and individual B factor refinements (group B factor refinements were used for the m67 complex, with one value for main-chain atoms and one value for side-chain atoms) using the CNS software package (41), resulting in R and Rfree values of 0.250 and 0.275 for the m13 complex, and 0.223 and 0.277 for the m67 complex. A summary of refinement statistics is given in Table II. These values are better than those for structures of similar resolutions. Although the m67 complex consisted of one complex per asymmetric unit, the m13 complex had eight complexes per asymmetric unit. Noncrystallographic symmetry operators were used throughout data refinement and were not relaxed due to the lower resolution of the data. One residue of the m67 complex is in the disallowed region. This residue, Asp29 of the MHC, resides on a loop involved in crystal packing with Tyr8 of the MHC, forcing it into a poor geometry. Ramachandran values were determined with the program Procheck (42, 43).

To investigate the level of cross-reactivity in engineered high-affinity TCR, SPR was used to determine kinetic and equilibrium-binding affinities of 2C, m6, m13, and m67 TCR for both QL9-Ld and SIY-Kb (Figs. 1 and 2; Table I). Binding affinities of 31.9 μM and 3.9 μM have been reported for 2C with SIY-Kb and QL9-Ld, respectively, and these values were obtained using a 2C TCR construct that contained the variable and constant domains (44). In this study we have used a stabilized sc version of the 2C TCR (2C-T7 scTCR) containing only the variable domains (Vα and Vβ) connected by a flexible linker, as well as several mutations for increased protein stability (34). Additionally, we have previously described a stabilized mutant Ld protein, Ld-m31, that contains only the α1 and α2 domains and can be refolded from inclusion bodies without the addition of β2-microglobulin (32). The equilibrium-binding affinities assessed by Scatchard analysis (Fig. 2) of the 2C-T7 scTCR were 22.1 μM for SIY-Kb and 1.6 μM for QL9-Ld-m31 at 25°C, each within 2-fold of the previously reported values (44). Thus, the sc formats of the TCR or the Ld molecule did not appreciably alter the binding properties of the TCR, and for comparison of relative binding affinities of 2C and high-affinity mutants, we have used the equilibrium values obtained with the 2C-T7 scTCR.

FIGURE 1.

Surface plasmon resonance of wild-type and high-affinity TCR binding to QL9-Ld or SIY-Kb. SPR traces of immobilized QL9-Ld (left) or SIY-Kb (right) binding by soluble scTCR. Biotinylated QL9-Ld-m31 or SIY-Kb was immobilized on a neutra-avidin-coated CM5 sensor chip surface (Biacore) at 400–600 response units (RU). Soluble scTCR were flowed over the surface at a flow rate of 30 μl/min at varying concentrations: from top to bottom on immobilized QL9-Ld: [2C-T7]: 20, 10, 5, 2.5, 1.25, 0.625, 0.313, 0.156, and 0.078 μM; [m67]: 10, 5, 2.5, 1.25, 0.625, 0.313, 0.156, and 0.078 μM; [m6]: 0.050, 0.025, 0.013, 0.0063, 0.003, 0.0012, 0.0008, and 0.0004 μM; [m13]: 1.28, 0.64, 0.32, 0.16, 0.08, 0.02, and 0.01 μM; from top to bottom on immobilized SIY-Kb: [2C-T7]: 54, 27, 13, 6.8, 3.4, 1.7, 0.84, 0.42, and 0.21 μM; [m67]: 0.78, 0.39, 0.195, 0.098, 0.049, 0.024, 0.012, 0.006, and 0.003 μM; [m6]: 68, 34, 17, 8.5, 4.3, 2.1, 1.1, and 0.53 μM; [m13]: 100, 50, 25, 12.5, 6.3, 3.1, 1.6, and 0.8 μM. To account for nonspecific interactions, binding to a control surface containing immobilized null pMHC complexes OVA-Kb or MCMV-Ld was subtracted from the data. The bottom line in all traces represents an injection of buffer only.

FIGURE 1.

Surface plasmon resonance of wild-type and high-affinity TCR binding to QL9-Ld or SIY-Kb. SPR traces of immobilized QL9-Ld (left) or SIY-Kb (right) binding by soluble scTCR. Biotinylated QL9-Ld-m31 or SIY-Kb was immobilized on a neutra-avidin-coated CM5 sensor chip surface (Biacore) at 400–600 response units (RU). Soluble scTCR were flowed over the surface at a flow rate of 30 μl/min at varying concentrations: from top to bottom on immobilized QL9-Ld: [2C-T7]: 20, 10, 5, 2.5, 1.25, 0.625, 0.313, 0.156, and 0.078 μM; [m67]: 10, 5, 2.5, 1.25, 0.625, 0.313, 0.156, and 0.078 μM; [m6]: 0.050, 0.025, 0.013, 0.0063, 0.003, 0.0012, 0.0008, and 0.0004 μM; [m13]: 1.28, 0.64, 0.32, 0.16, 0.08, 0.02, and 0.01 μM; from top to bottom on immobilized SIY-Kb: [2C-T7]: 54, 27, 13, 6.8, 3.4, 1.7, 0.84, 0.42, and 0.21 μM; [m67]: 0.78, 0.39, 0.195, 0.098, 0.049, 0.024, 0.012, 0.006, and 0.003 μM; [m6]: 68, 34, 17, 8.5, 4.3, 2.1, 1.1, and 0.53 μM; [m13]: 100, 50, 25, 12.5, 6.3, 3.1, 1.6, and 0.8 μM. To account for nonspecific interactions, binding to a control surface containing immobilized null pMHC complexes OVA-Kb or MCMV-Ld was subtracted from the data. The bottom line in all traces represents an injection of buffer only.

Close modal
FIGURE 2.

Scatchard analysis of TCR:pMHC binding. Scatchard analysis was performed by plotting the maximum response (RU Max) vs the maximum response/protein concentration (RU Max/[protein]). All points were fitted using linear regression analysis with Microsoft Excel software. The R2 value for all data shown is 0.95 or better. The negative inverse of the slope represents the equilibrium affinity value.

FIGURE 2.

Scatchard analysis of TCR:pMHC binding. Scatchard analysis was performed by plotting the maximum response (RU Max) vs the maximum response/protein concentration (RU Max/[protein]). All points were fitted using linear regression analysis with Microsoft Excel software. The R2 value for all data shown is 0.95 or better. The negative inverse of the slope represents the equilibrium affinity value.

Close modal
Table I.

Binding properties of scTCR at 25°C measured by surface plasmon resonancea

scTCRCDR3α SequencepMHCKD eq (nM)KD kin (nM)Ka (103) (1/Ms)Kd (10−2) (1/s)t1/2 (s)
2C-T7 SGFASAL SIY-Kb 22,100 27,400 22 46.4 1.5 
  QL9-Ld 1,600 3,400 96 28.6 2.5 
        
m6 SHQGRYL SIY-Kb 39,000 25,800 0.6 1.6 44 
  QL9-Ld 33 14 370 0.3 213 
        
m67 SLERPYL SIY-Kb 1,300 0.5 140 
  QL9-Ld 4,600 7,500 45 28.5 2.7 
        
m13 SDPPPLL SIY-Kb ND     
  QL9-Ld 116 154 357 21 
scTCRCDR3α SequencepMHCKD eq (nM)KD kin (nM)Ka (103) (1/Ms)Kd (10−2) (1/s)t1/2 (s)
2C-T7 SGFASAL SIY-Kb 22,100 27,400 22 46.4 1.5 
  QL9-Ld 1,600 3,400 96 28.6 2.5 
        
m6 SHQGRYL SIY-Kb 39,000 25,800 0.6 1.6 44 
  QL9-Ld 33 14 370 0.3 213 
        
m67 SLERPYL SIY-Kb 1,300 0.5 140 
  QL9-Ld 4,600 7,500 45 28.5 2.7 
        
m13 SDPPPLL SIY-Kb ND     
  QL9-Ld 116 154 357 21 
a

Values shown are averages of two or more independent measurements. TCR half-lives were calculated using the formula ln 2/kd. ND, Not detected; eq, equilibrium; kin, kinetic.

The m67 mutant that was selected for high-affinity binding to the structurally related dEV8-Kb complex had an equilibrium-binding affinity of 7 nM for SIY-Kb, in agreement with previously determined values for the full-length TCR (30). TCR m67 retained the ability to bind to QL9-Ld, although its equilibrium-binding affinity was reduced by ∼3-fold, relative to 2C (Figs. 1 and 2 and Table I). The m6 mutant, selected for high-affinity binding to the QL9-Ld complex, had an equilibrium-binding affinity of 33 nM for QL9-Ld, similar to the affinity of the full-length TCR. The m13 mutant, also selected for high-affinity binding to QL9-Ld, had an equilibrium-binding affinity of 116 nM. This affinity is 10-fold lower than reported, although previous values were determined using cell-based methods rather than SPR (30).

Unlike m67, the m13 mutant no longer cross-reacted with SIY-Kb, as detectable binding could not be measured in SPR experiments at 25°C (Fig. 1). Low-affinity binding to SIY-Kb by the m6 mutant was detected in two independent SPR experiments; however, this interaction exhibited a very slow association rate and a half-life that was 7-fold longer than that of 2C (Fig. 1 and Table I). Due to the unusual kinetic parameters, we sought independent verification of m6/SIY-Kb binding. In one experiment, biotinylated m67 scTCR was used as the probe for binding to SIY-Kb on T2-Kb cells, as detected using streptavidin/PE and flow cytometry. Various concentrations of unlabeled scTCR were added to the biotinylated m67 scTCR in a competition format. Soluble m6 scTCR was unable to inhibit the binding of high-affinity m67 scTCR to SIY-Kb-bearing cells, whereas both unlabeled m67 and the 2C-T7 scTCR inhibited binding (supplemental Fig. 1)4. The binding affinities for m67 and 2C-T7 in this assay were calculated to be 3 nM and 7 μM, respectively. Given the difference in temperature of the two measurements (4°C for the inhibition method and 25°C for SPR), these values are in agreement with those measured by SPR. In addition, we performed yet a third independent binding analysis that involved the use of 2C and m6 TCR transduced 58−/− cells (CD8-negative). In this assay, SIY-Kb tetramers were added to the cell lines at various concentrations; only the 2C line showed binding above the parental cell line 58−/− (supplemental Fig. 2). These results suggest that the interaction between m6 and SIY-Kb is weak at best, and that the m6/SIY-Kb SPR binding results were likely spurious, a conclusion also supported by the lack of biological activity of SIY peptide for m6 tranductants (see below). In summary, in vitro engineered, high-affinity 2C TCR with different mutations in the CDR3α varied in their level of cross-reactivity as determined in binding assays.

T cells that express TCR with binding affinities lower than the detection limit of SPR can often still be activated in response to their cognate pMHC ligand on APC (45, 46). This level of sensitivity is achieved in large part by the presence of the coreceptor molecules CD4 or CD8 (47, 48). Thus, to further assess the level of cross-reactivity in high-affinity TCR, we analyzed the activation of T cell hybridoma lines transfected with the 2C TCR or one of the high-affinity TCR mutants (30). Transfectants that expressed wild-type or high-affinity TCR, and the CD8αβ coreceptor where indicated, were stimulated with various amounts of SIY or QL9 and APC with the corresponding MHC. Activation was monitored by release of IL-2 (Fig. 3). Activation of 2C, m6, and m13 transfectants in response to Ld- positive APC loaded with QL9 peptide was CD8 independent, as expected from their binding affinities and previous studies (31, 49). Despite having a 3-fold lower affinity relative to 2C, m67 transfectants maintained CD8-independent activation in response to the reciprocal ligand QL9-Ld, although the SD50 is enhanced by approximately two orders of magnitude when CD8 is present (Fig. 3, left panel).

FIGURE 3.

Functional activity of T cell lines transfected with 2C, m6, m13, and m67 TCR. Percentage of maximal IL-2 release of TCR transfectants stimulated with QL9-Ld (squares, left panel) or SIY-Kb (circles, right panel). Peptides were added to T2-Ld or T2-Kb cells and an equal number of T cell transfectant cells were added; closed circles or squares indicate CD8 T cell transfectants. Open circles or squares indicate CD8 T cell transfectants. Cells were incubated for 24 h at 37°C and supernatants were assayed for IL-2 using ELISA. Percentage of maximal IL-2 release = [(sample A450 − null peptide A450)/(CD3 A450 − null peptide A450)] × 100 where CD3 A450 is the absorbance measured for cells stimulated with an anti-CD3 Ab, and null peptide A450 is the absorbance value for cells stimulated with a null peptide (OVA for Kb and MCMV for Ld).

FIGURE 3.

Functional activity of T cell lines transfected with 2C, m6, m13, and m67 TCR. Percentage of maximal IL-2 release of TCR transfectants stimulated with QL9-Ld (squares, left panel) or SIY-Kb (circles, right panel). Peptides were added to T2-Ld or T2-Kb cells and an equal number of T cell transfectant cells were added; closed circles or squares indicate CD8 T cell transfectants. Open circles or squares indicate CD8 T cell transfectants. Cells were incubated for 24 h at 37°C and supernatants were assayed for IL-2 using ELISA. Percentage of maximal IL-2 release = [(sample A450 − null peptide A450)/(CD3 A450 − null peptide A450)] × 100 where CD3 A450 is the absorbance measured for cells stimulated with an anti-CD3 Ab, and null peptide A450 is the absorbance value for cells stimulated with a null peptide (OVA for Kb and MCMV for Ld).

Close modal

Similar experiments were performed with transfectants using Kb-positive APC loaded with SIY peptide (Fig. 3, right panel). In this case, the 2C TCR transfectants required CD8 for activity, as has been observed previously (49, 50), whereas the high-affinity m67 TCR was equally reactive in the presence or absence of CD8. Consistent with the lack of detectable binding of SIY-Kb by soluble TCR, m13 TCR transfectants were not active in either the presence or absence of CD8. The m6 TCR transfectant was also not active in the presence of CD8, in agreement with two of the three binding assays described above. Thus, the m67 TCR maintained a level of functional cross-reactivity similar to the wild-type 2C, whereas the m6 and m13 TCR have lost functional cross-reactivity compared with wild-type 2C.

To understand the structural mechanism by which high-affinity TCR maintain or lose cross-reactivity with reciprocal ligands, a 3.8 Å resolution crystal structure of the m67/QL9-Ld complex and a 2.9 Å resolution structure of the m13/QL9-Ld complex (supplemental Fig. 3 and Fig. 4), were solved for comparison to previously reported 2C/QL9-Ld and m6/QL9-Ld complex structures (Table II; Ref. 27). The resolution of the m67/QL9-Ld complex is sufficient for comparison of docking angles of the TCR on the pMHC and the CDR backbone conformations. The 2C, m6, m13, and m67/QL9-Ld and the 2C/SIY-Kb complexes were overlaid, and TCR docking orientations and CDR loop conformations in the five structures were compared (Fig. 4, A and B). As previously reported, the docking angle of 2C with QL9-Ld differs significantly from the docking angle of 2C with SIY-Kb (27). Interestingly, the m67 TCR, which contains mutations that confer a 2000-fold increase in affinity for SIY-Kb, maintained a docking solution on QL9-Ld that is nearly identical to that of the 2C/QL9-Ld complex and to that of the m6 and m13 TCR, which were engineered for high affinity against QL9-Ld (Fig. 4 A). Thus an increase in the affinity of a TCR for one pMHC via mutations in the CDR3α loop does not force the TCR into a docking strategy complimentary to the high-affinity, yet structurally diverse, ligand (SIY-Kb, in this case).

FIGURE 4.

Conserved QL9-Ld docking orientation and CDR conformations in 2C, m6, m13, and m67 TCR. A, Top view of 2C/QL9-Ld, m6/QL9-Ld, m13/QL9-Ld, m67/QL9-Ld, and 2C/SIY-Kb complex structures with MHC overlaid. SIY-Kb is shown in brown and QL9-Ld is shown in green. The CDR of the TCR are shown as loops with 2C/Kb in red, 2C/Ld in blue, m6/Ld in light blue, m13/Ld in purple, and m67/Ld in cyan. B, CDR loop conformations of 2C/Kb (red), 2C/Ld (blue), m6/Ld (light blue), m13/Ld (purple), and m67/Ld (cyan). CDR loop conformations of unliganded 2C are shown in gray for comparison. TCR and MHC alignments were done using PyMol. C, Composite omit map of m67/QL9-Ld. Composite omit map was calculated omitting 5% of the model at a time. The α-chain of m67 is shown colored raspberry red, with the QL9 peptide in yellow and Ld in green, where only main-chain atoms are shown because the location of side chains is not clearly defined by the electron density. For clarity, only those atoms near the CDR1α and CDR3α loops are shown. The structure was determined by molecular replacement using individual components of the 2C/QL9-Ld structure as search models (PDB ID 2OI9). CDR1α and CDR3α of 2C are shown in pink, where 2C/QL9-Ld superimposed globally on m67/QL9-Ld with a root mean square deviation of 0.52 Å. Molecular replacement originally yielded solutions where CDR1α and CDR3α were outside the electron density tube observed in composite omit maps, indicating that loop placement was not biased by molecular replacement models, as well as highlighting the subtle differences in m67 and 2C despite globally similar recognition of QL9-Ld.

FIGURE 4.

Conserved QL9-Ld docking orientation and CDR conformations in 2C, m6, m13, and m67 TCR. A, Top view of 2C/QL9-Ld, m6/QL9-Ld, m13/QL9-Ld, m67/QL9-Ld, and 2C/SIY-Kb complex structures with MHC overlaid. SIY-Kb is shown in brown and QL9-Ld is shown in green. The CDR of the TCR are shown as loops with 2C/Kb in red, 2C/Ld in blue, m6/Ld in light blue, m13/Ld in purple, and m67/Ld in cyan. B, CDR loop conformations of 2C/Kb (red), 2C/Ld (blue), m6/Ld (light blue), m13/Ld (purple), and m67/Ld (cyan). CDR loop conformations of unliganded 2C are shown in gray for comparison. TCR and MHC alignments were done using PyMol. C, Composite omit map of m67/QL9-Ld. Composite omit map was calculated omitting 5% of the model at a time. The α-chain of m67 is shown colored raspberry red, with the QL9 peptide in yellow and Ld in green, where only main-chain atoms are shown because the location of side chains is not clearly defined by the electron density. For clarity, only those atoms near the CDR1α and CDR3α loops are shown. The structure was determined by molecular replacement using individual components of the 2C/QL9-Ld structure as search models (PDB ID 2OI9). CDR1α and CDR3α of 2C are shown in pink, where 2C/QL9-Ld superimposed globally on m67/QL9-Ld with a root mean square deviation of 0.52 Å. Molecular replacement originally yielded solutions where CDR1α and CDR3α were outside the electron density tube observed in composite omit maps, indicating that loop placement was not biased by molecular replacement models, as well as highlighting the subtle differences in m67 and 2C despite globally similar recognition of QL9-Ld.

Close modal

We also compared the CDR loop conformations of the 2C, m6, m13, and m67 TCR bound to QL9-Ld, the 2C TCR bound to SIY-Kb, and the unliganded 2C TCR (Fig. 4,B). The CDR2α, CDR1β, CDR2β, and CDR3β loops retained almost identical conformations, even in the unliganded TCR and in the Kb complex where the 2C TCR has a different docking angle. The most significant differences among CDR are found in CDR1α and CDR3α. The CDR1α loop of m67/Ld is shifted by ∼3 Å relative to 2C/Ld, but interestingly the backbone of the m67 CDR1α is almost identical to that of 2C/Kb. Similarly, the position of the CDR3α backbone of m67 is most similar to the position of the CDR3α backbone from the 2C TCR bound to SIY/Kb. To show that even at this low resolution it is apparent that the backbones of the CDR1α and CDR3α loops differ between m67 and 2C, electron density from these loops in the m67 structure were superimposed with the backbone structures of 2C vs m67, from their respective QL9-Ld complexes (Fig. 4 C). This representation shows that it is not possible to fit the 2C CDR1/3α backbones into the density of the m67 complex, consistent with different conformations of these loops.

Although the remodeled CDR3α and CDR1α loops in the m67 TCR have a minor detrimental effect on binding to QL9-Ld (Table I), this does not prevent binding that is sufficient to maintain CD8-independent cross-reactivity with QL9-Ld. Comparison of the m67 CDR3α loops in the m67/Ld structure (Fig. 5,A) and in m67 manually docked onto SIY-Kb (Fig. 5 B) indeed show that in both cases CDR3α bends back away from the central peptide residues and thereby avoids steric hindrance problems associated with binding to either ligand (also see below).

FIGURE 5.

CDR3α loop and peptide conformations allow or disallow cross-reactivity in high-affinity TCR. A, The CDR3 loops of 2C, m6, m13, and m67 TCR on QL9-Ld. The QL9 peptide is shown in magenta and Ld in green. B, Models of CDR3 loops of 2C, m6, m13, and m67 TCR manually docked on SIY-Kb. The scTCR from the m6/QL9-Ld, m13/QL9-Ld, and m67/QL9-Ld structures were aligned with the 2C TCR from the 2C/SIY-Kb complex. The SIY peptide is shown in yellow and Kb in brown. The CDR3α and CDR3β of each TCR are shown: 2C in red for the Kb complex and blue for the Ld complex, m6 in light blue, m13 in purple, and m67 in cyan. All structural alignments were done using PyMol.

FIGURE 5.

CDR3α loop and peptide conformations allow or disallow cross-reactivity in high-affinity TCR. A, The CDR3 loops of 2C, m6, m13, and m67 TCR on QL9-Ld. The QL9 peptide is shown in magenta and Ld in green. B, Models of CDR3 loops of 2C, m6, m13, and m67 TCR manually docked on SIY-Kb. The scTCR from the m6/QL9-Ld, m13/QL9-Ld, and m67/QL9-Ld structures were aligned with the 2C TCR from the 2C/SIY-Kb complex. The SIY peptide is shown in yellow and Kb in brown. The CDR3α and CDR3β of each TCR are shown: 2C in red for the Kb complex and blue for the Ld complex, m6 in light blue, m13 in purple, and m67 in cyan. All structural alignments were done using PyMol.

Close modal

To better understand how the increased affinity of m6 and m13 for QL9-Ld was associated with the loss of SIY-Kb cross-reactivity, we compared the structures of 2C/QL9-Ld, m6/QL9-Ld, and m13/QL9-Ld, and we manually docked the m6, m13, and m67 TCR onto the SIY-Kb ligand. In the 2C/QL9-Ld complex, CDR3α is pointed away from the peptide, back over the α1 helix, where it contacts QL9 and Ld. In the high-affinity m6/QL9-Ld complex, the mutated CDR3α loop is moved over the peptide, increasing the contact of the CDR3α residues with QL9 (27 ; Fig. 5,A and Fig. 6). The CDR3α loop of the high-affinity m13 TCR is positioned similarly and also makes additional van der Waals contacts with QL9. In both high-affinity TCRs, these interactions are focused around the central phenylalanine residue of QL9. In m6, new contacts are mediated through the CDR3α backbone near the glycine residue at position 101, whereas in m13 they are primarily focused around the proline residues at the center of the loop (Fig. 6). During the original selection process, a glycine residue at position 101, or three proline residues from 100 to 102, were the only conserved motifs that were isolated in selection of 2C mutants with high affinity for QL9-Ld. It is likely that the overall effect of these residues on the position of the CDR3α loop, as well as their additional interactions with the QL9 peptide, contributed to the selective pressure associated with isolation of these motifs.

FIGURE 6.

Interactions between the CDR3 loops and QL9 peptide residues in 2C, m6, and m13 TCR. Interactions between residue side chains in the CDR3α and CDR3β loops and the QL9 peptide in 2C/QL9-Ld complex (A), m6/QL9-Ld complex (B), and m13/QL9-Ld complex (C). Interactions with distances between 2 and 3 Å are shown with green dashed lines. Distances between 3 and 4 Å are shown with blue dashed lines. Interactions between the CDR3β glycine residues and F7 and D8 of the QL9 peptide are conserved in all three structures and are not shown for clarity. All distances were determined using PyMol.

FIGURE 6.

Interactions between the CDR3 loops and QL9 peptide residues in 2C, m6, and m13 TCR. Interactions between residue side chains in the CDR3α and CDR3β loops and the QL9 peptide in 2C/QL9-Ld complex (A), m6/QL9-Ld complex (B), and m13/QL9-Ld complex (C). Interactions with distances between 2 and 3 Å are shown with green dashed lines. Distances between 3 and 4 Å are shown with blue dashed lines. Interactions between the CDR3β glycine residues and F7 and D8 of the QL9 peptide are conserved in all three structures and are not shown for clarity. All distances were determined using PyMol.

Close modal

In addition to increased contact with QL9, the m6 and m13 TCR exhibited differences in inter- and intra-CDR loop contacts compared with the 2C TCR. These included interactions between residues in the CDR3α and CDR2β loops. In the 2C complex, the phenylalanine (F100α) at the tip of the CDR3α loop makes van der Waals contacts with Y48β and Y50β in the CDR2β loop (Fig. 7,A). Although the high-affinity m6 mutant maintained contact with these CDR2β tyrosines by using an arginine residue at position 102α (Fig. 7,B), contact between the CDR3α and CDR2β residues was lost in the m13 TCR. The three tandem prolines of m13 may act to stabilize the CDR3α without the need for the CDR3α:CDR2β stabilizing interactions. The tyrosine residues Y48β and Y50β have been shown by alanine scanning mutagenesis to contribute significant binding energy to the formation of the 2C/QL9-Ld complex (51). In addition, it has been proposed that Y50β serves as an evolutionary conserved point of interaction with MHC (52 ; note that Y48β and Y50β correspond to Y46β and Y48β in their review). Thus, despite variability in CDR3α:CDR2β interloop interactions among the three TCRs, the two CDR2β tyrosines maintained similar positions (Fig. 7).

FIGURE 7.

CDR loop interactions in the 2C and m6 TCR. A, Interactions between CDR loop side chains in the 2C TCR from the 2C/QL9-Ld complex structure. B, Interactions between CDR loop side chains in the m6 TCR from the m6/QL9-Ld complex structure. C, Interactions between CDR loop side chains in the m13 TCR from the m13/QL9-Ld complex structure. Interloop interactions within 3.5 Å are shown as blue dashed lines, and intraloop interactions within 3.5 Å are shown as green dashed lines. All distances were determined using PyMol.

FIGURE 7.

CDR loop interactions in the 2C and m6 TCR. A, Interactions between CDR loop side chains in the 2C TCR from the 2C/QL9-Ld complex structure. B, Interactions between CDR loop side chains in the m6 TCR from the m6/QL9-Ld complex structure. C, Interactions between CDR loop side chains in the m13 TCR from the m13/QL9-Ld complex structure. Interloop interactions within 3.5 Å are shown as blue dashed lines, and intraloop interactions within 3.5 Å are shown as green dashed lines. All distances were determined using PyMol.

Close modal

The displacement of CDR3α in both high-affinity TCRs increased the extent of stabilizing interactions with the neighboring CDR1α loop. In the 2C TCR, there are no CDR3α:CDR1α interloop interactions (Fig. 7,A). In m6, the glutamine at position 100 contacts the tyrosine at position 26 in the CDR1α loop (Fig. 7,B). Additional contacts between the repositioned CDR3α and the CDR1α loop are also observed in the m13 TCR. In this case, the aspartic acid at position 99 and the proline at position 100 contact the same tyrosine at position 26 in the CDR1α loop (Fig. 7,C). Whereas residue Y26α showed a moderate effect in the alanine scan, neighboring CDR1α residues Y31α and T29α are pMHC contact residues and contributed significant energy to 2C/QL9-Ld complex formation (51). These results support the view that an increase in stability of the CDR1α loop, mediated through interactions with CDR3α residues, could contribute to the increase in binding affinity. Finally, whereas the CDR3α loop of 2C lacks intraloop contacts, four of the five affinity mutations in m6 CDR3α are involved in intraloop van der Waals interactions (Fig. 7,B). There are no additional intraloop contacts in the m13 TCR (Fig. 7 C); however, the conformational rigidity conferred by three sequential proline residues may stabilize the structure of the loop in a similar fashion.

Although limited in the number of structures available, there does appear to be a correlation between additional intraloop, interloop, and peptide contacts in the TCR and the affinity of the TCR for QL9-Ld. The lowest affinity 2C TCR has CDR3α-CDR2β-stabilizing interactions, but few CDR3α-peptide contacts and few inter-CDR3α loop interactions. The intermediate affinity m13 TCR has lost CDR3α-CR2β interactions, but gained CDR3α-CDR1α interloop contacts and CDR3α-peptide contacts, and m13 contains the more rigid, proline-rich CDR3α. Finally, the highest affinity m6 TCR has maintained the CDR3α-CDR2β interactions, gained CDR3α-peptide contacts and CDR3α-CDR1α interactions, and m6 contains a CDR3α loop with additional intraloop stabilizing interactions. These results support the view that high affinity is achieved not only by increasing contacts with the peptide, but by stabilizing the overall conformation of the CDR3α loop and its interactions with neighboring loops.

For the reasons described above, the m6, m13, and m67 high-affinity TCR are all likely to be restricted in terms of CDR mobility, but only m6 and m13 have lost their ability to cross-react with the alternative ligand (SIY-Kb). The loss of conformational mobility in CDR3α, and perhaps CDR1α, could lock the m6 and m13 TCR into the conformation found in the QL9-Ld complexes. Although we do not have unliganded structures of either the m6 or m13 TCR to confirm this, it is supported by the increase in association rate of both TCR with QL9-Ld (Table I), suggesting that fewer conformational adjustments are required for binding. By superimposing the three scTCR from the QL9-Ld complex structures onto the structure of 2C bound to SIY-Kb (root mean square deviation = 0.96 for m6, 0.98 for m67, 1.22 for m13), it is possible to examine the potential steric clashes that may occur in these three cases (Fig. 5,B). The CDR3α loop of 2C bound to SIY-Kb points outward toward the α1 helix of Kb, interacting with the peptide (arginine at position 4 of SIY) only through the CDR3α main chain (28). As described above, the m67 CDR3α conformation is only slightly altered from this configuration, and also points outward over the α1 helix. Accordingly, the m67 TCR could bind to SIY-Kb without steric interference and without substantial changes in the CDR3α conformation. In contrast, the CDR3α loop of both m6 and m13 is displaced inward, pointing toward the peptide, instead of away from it as in the 2C/SIY-Kb complex. Without CDR flexibility to allow for repositioning of the CDR3α loop, the arginine at position 4 of the SIY peptide would sterically interfere with binding of both the m6 and m13 TCR to SIY-Kb (Fig. 5,B). In m6, this clash would be with the CDR3α backbone, whereas in m13 the side chain of the central proline residue would interfere with binding. As there is not yet a structure of the unliganded SIY-Kb complex, it remains to be seen whether the side chain of arginine at position 4 might exhibit sufficient conformational mobility to mitigate these potential steric clashes. In the QL9-Ld structures, none of the QL9 peptide residues extend from the surface of the complex where they would clash with the m6 CDR3α (Fig. 5 A). Thus, we hypothesize that the loss of SIY-Kb cross-reactivity by the m6 TCR is most likely a result of decreased conformational mobility of the CDR3α loop and the resulting inability of the TCR to accommodate this structurally different pMHC surface.

Structural studies of wild-type affinity TCR have provided evidence for a link between conformational flexibility in the CDRs TCR degeneracy, and the low affinity of typical TCR/pMHC interactions (2, 3). There is also some evidence that the CDR3 loops could stabilize and alter the conformations of the CDR1 and CDR2 loops, suggesting that the CDR3 could influence interactions (and by inference, the docking) of other loops with MHC (53). In this study we looked more directly at the effects of CDR3 on TCR cross-reactivity and CDR flexibility by examining the binding and functional properties of three CDR3α mutants with high-affinity for different pMHC ligands. Our findings show that mutations in CDR3α that yielded high-affinity for one ligand can either maintain the wild-type cross-reactivity with the alternative ligand (e.g., with m67 TCR) or can lead to a loss of cross-reactivity with the alternative ligand (e.g., with the m6 and m13 TCRs). In all three cases, the exact docking orientation and even the configuration of the CDR loops (except for CDR3α and CDR1α) on QL9-Ld was maintained. This orientation was even present with the m67 TCR that has a 2000-fold higher affinity for the SIY-Kb complex, and thus one might have predicted it would adopt a 2C/SIY-Kb docking angle on Ld. The observation that MHC-specific docking orientation is maintained even in this case suggests that the binding energies associated with the germline-encoded CDRs (CDR1s and CDR2s) are sufficiently strong to compensate for the CDR3 differences, at least among this collection of TCR. In fact, the binding energies associated with CDR1 and CDR2 residues in the 2C TCR have been shown to dominate in the reactions with both QL9-Ld (51) and SIY-Kb (54).

It is interesting that in the QL9-Ld complex the CDR1α and CDR3α loops of m67 are positioned like the equivalent loops in 2C bound to Kb, despite the fact that this TCR docks onto Ld at a very different angle than 2C docks onto Kb. In the 2C TCR, both the CDR1α and CDR3α loops undergo the largest changes from the unliganded TCR upon Ld binding. These results are consistent with the notion that the increase in affinity of m67 for SIY-Kb is in part achieved by stabilizing the CDR1α and CDR3α loops into conformations that are best suited for interaction with Kb. This is supported by an increase in the on-rate of m67 for SIY-Kb (Table I). Results with the m6 and m13 TCR showed that mutations to CDR3α loop alone can confer large increases in affinity for one pMHC ligand, in part by increasing stabilizing interactions with nearby loops. These findings also suggest that the native conformations of the CDR loops may be more rigid than has been assumed, and that there is actually a very limited ensemble of possible conformations. This possibility, and the highly conserved docking orientation, are consistent with the idea that relatively rigid CDRs, and CDR1β, CDR2β, and CDR2α in particular, are largely responsible for the engagement of specific MHC molecules in an energetically favorable docking footprint (52, 55, 56). It remains to be seen whether the m67 TCR docks onto SIY-Kb with the same footprint as 2C, but our prediction is that it will maintain the same CDR loop conformations as in the m67/QL9-Ld complex, only docked at the angle of the 2C/SIY-Kb structure. Accordingly, the high-affinity of m67 for SIY-Kb would be largely achieved by enhancing the conformational stability of CDR3α and CDR1α for optimal complementarity with the SIY-Kb ligand. In fact, the similarities in all six CDR conformations between the m67 TCR bound to Ld, and the 2C TCR bound to Kb (Fig. 4 B), suggest that there may be energetically favorable conformations for binding regardless of the pMHC ligand.

TCR m6 and m13 both exhibited repositioning of the CDR3α loop that allowed for increased contacts with the QL9 peptide and neighboring CDRs. Thus the level of cross-reactivity depended on the positioning of the mutated CDR3α in the complexes, and whether the peptide side chains of the ligands allowed binding without steric constraints. Studies of the BM3.3 TCR have shown that movement of the CDR3, to accommodate the pMHC surface, were essential in cross-reactivity between peptides with little structural similarity (8, 22). A recent study highlighted how TCRs are influenced by interface-disrupting side chains, either on the MHC helices or the peptide (57). Given that it is virtually impossible to predict the conformation that CDR3 loops will take in a liganded structure, it is likely to be difficult or impossible to predict the potential for cross-reactivity with either alloantigens or self-peptides. Thus, analysis of autoreactivity of higher affinity TCR, engineered in vitro, will require empirical analyses as has now been done with several TCR (30, 36, 58). It may be possible to identify mutated TCR, like m6 and m13, which exhibit reduced cross-reactions because of the positioning of their CDR3 regions.

Although the process of MHC restriction makes interpretations of TCR cross-reactivity unique in comparison to Abs, there are useful analogies. For example, germline-encoded Abs from primary responses tend to exhibit low-affinity binding interactions, a high level of cross-reactivity (i.e., lower specificity), and flexible CDR loops. Abs that have undergone somatic hypermutation and clonal expansion have increases in affinity relative to their germline-encoded counterparts, very high specificity for Ag, and little or no conformational mobility in the CDR (59, 60, 61, 62). In Abs, a reduction in CDR mobility has been correlated with a decrease in cross-reactivity and an increase in specificity. In one sense, TCR cross-reactivity might be simplified compared Abs, if one assumes that the TCR CDR1 and CDR2 loops are largely fixed on the MHC. In this case, the central CDR (CDR3α and CDR3β) control cross-reactivity with different peptides. Cross-reactivity is then dependent only on the following: 1) the conformation adopted by the CDR3s, and 2) the nature of the peptide(s) bound in the groove. For example, despite multiple CDR3α conformations (Fig. 4 B), cross-reactivity was observed between QL9-Ld and all four TCR: 2C, m67, m13, and m6. This was in large part due to the planar surface formed by the QL9 peptide, in which there are few projecting side chains that would interfere sterically with the diverse CDR3α loops. Thus, the degeneracy of the 2C TCR is not only associated with conformational flexibility of the CDR3 in a process that has been assumed to account for such cross-reactions, but is in large part dictated by positions of the less flexible CDR3 and the projecting side chains from the bound peptide.

We thank the University of Illinois Immunological Resource Center for assistance and Phil Holler for early studies on the m6, m13, and m67 TCR. The PDB identification numbers of the m67/QL9-Ld and m13/QL9-Ld coordinates are 3E2H and 3E3Q, respectively.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported by National Institutes of Health Grants GM55767 (to D.M.K.) and AI48540 (to K.C.G.), a predoctoral grant from the National Science Foundation (to L.A.C.), and the Samuel and Ruth Engelberg/Irvington Institute Fellowship of the Cancer Research Institute (to J.D.S.).

3

Abbreviations used in this paper: pMHC, peptide MHC; sc, single chain; SPR, surface plasmon resonance; PDB, Brookhaven Protein Data Bank.

4

The online version of this article contains supplemental material.

1
Davis, M. M., P. J. Bjorkman.
1988
. T-cell antigen receptor genes and T-cell recognition.
Nature
334
:
395
-402.
2
Garcia, K. C., E. J. Adams.
2005
. How the T cell receptor sees antigen–a structural view.
Cell
122
:
333
-336.
3
Rudolph, M. G., R. L. Stanfield, I. A. Wilson.
2006
. How TCRs bind MHCs, peptides, and coreceptors.
Annu. Rev. Immunol.
24
:
419
-466.
4
Mason, D..
1998
. A very high level of crossreactivity is an essential feature of the T-cell receptor.
Immunol. Today
19
:
395
-404.
5
Wucherpfennig, K. W..
2004
. T cell receptor crossreactivity as a general property of T cell recognition.
Mol. Immunol.
40
:
1009
-1017.
6
Starr, T. K., S. C. Jameson, K. A. Hogquist.
2003
. Positive and negative selection of T cells.
Annu. Rev. Immunol.
21
:
139
-176.
7
Archbold, J. K., W. A. Macdonald, S. R. Burrows, J. Rossjohn, J. McCluskey.
2008
. T-cell allorecognition: a case of mistaken identity or deja vu?.
Trends Immunol.
29
:
220
-226.
8
Reiser, J. B., C. Darnault, C. Gregoire, T. Mosser, G. Mazza, A. Kearney, P. A. van der Merwe, J. C. Fontecilla-Camps, D. Housset, B. Malissen.
2003
. CDR3 loop flexibility contributes to the degeneracy of TCR recognition.
Nat. Immunol.
4
:
241
-247.
9
Wilson, D. B., D. H. Wilson, K. Schroder, C. Pinilla, S. Blondelle, R. A. Houghten, K. C. Garcia.
2004
. Specificity and degeneracy of T cells.
Mol. Immunol.
40
:
1047
-1055.
10
Garcia, K. C., M. Degano, L. R. Pease, M. Huang, P. A. Peterson, L. Teyton, I. A. Wilson.
1998
. Structural basis of plasticity in T cell receptor recognition of a self peptide-MHC antigen.
Science
279
:
1166
-1172.
11
Kjer-Nielsen, L., C. S. Clements, A. W. Purcell, A. G. Brooks, J. C. Whisstock, S. R. Burrows, J. McCluskey, J. Rossjohn.
2003
. A structural basis for the selection of dominant αβ T cell receptors in antiviral immunity.
Immunity
18
:
53
-64.
12
Reiser, J. B., C. Gregoire, C. Darnault, T. Mosser, A. Guimezanes, A. M. Schmitt-Verhulst, J. C. Fontecilla-Camps, G. Mazza, B. Malissen, D. Housset.
2002
. A T cell receptor CDR3β loop undergoes conformational changes of unprecedented magnitude upon binding to a peptide/MHC class I complex.
Immunity
16
:
345
-354.
13
Hare, B. J., D. F. Wyss, M. S. Osburne, P. S. Kern, E. L. Reinherz, G. Wagner.
1999
. Structure, specificity and CDR mobility of a class II restricted single-chain T-cell receptor.
Nat. Struct. Biol.
6
:
574
-581.
14
Davis, M. M., J. J. Boniface, Z. Reich, D. Lyons, J. Hampl, B. Arden, Y. Chien.
1998
. Ligand recognition by α β T cell receptors.
Annu. Rev. Immunol.
16
:
523
-544.
15
Willcox, B. E., G. F. Gao, J. R. Wyer, J. E. Ladbury, J. I. Bell, B. K. Jakobsen, P. A. van der Merwe.
1999
. TCR binding to peptide-MHC stabilizes a flexible recognition interface.
Immunity
10
:
357
-365.
16
Boniface, J. J., Z. Reich, D. S. Lyons, M. M. Davis.
1999
. Thermodynamics of T cell receptor binding to peptide-MHC: evidence for a general mechanism of molecular scanning.
Proc. Natl. Acad. Sci. USA
96
:
11446
-11451.
17
Ishizuka, J., G. B. Stewart-Jones, A. van der Merwe, J. I. Bell, A. J. McMichael, E. Y. Jones.
2008
. The structural dynamics and energetics of an immunodominant T cell receptor are programmed by its Vβ domain.
Immunity
28
:
171
-182.
18
Davis-Harrison, R. L., K. M. Armstrong, B. M. Baker.
2005
. Two different T cell receptors use different thermodynamic strategies to recognize the same peptide/MHC ligand.
J. Mol. Biol.
346
:
533
-550.
19
Ely, L. K., T. Beddoe, C. S. Clements, J. M. Matthews, A. W. Purcell, L. Kjer-Nielsen, J. McCluskey, J. Rossjohn.
2006
. Disparate thermodynamics governing T cell receptor-MHC-I interactions implicate extrinsic factors in guiding MHC restriction.
Proc. Natl. Acad. Sci. USA
103
:
6641
-6646.
20
Lee, J. K., G. Stewart-Jones, T. Dong, K. Harlos, K. Di Gleria, L. Dorrell, D. C. Douek, P. A. van der Merwe, E. Y. Jones, A. J. McMichael.
2004
. T cell cross-reactivity and conformational changes during TCR engagement.
J. Exp. Med.
200
:
1455
-1466.
21
Gagnon, S. J., O. Y. Borbulevych, R. L. Davis-Harrison, T. K. Baxter, J. R. Clemens, K. M. Armstrong, R. V. Turner, M. Damirjian, W. E. Biddison, B. M. Baker.
2005
. Unraveling a hotspot for TCR recognition on HLA-A2: evidence against the existence of peptide-independent TCR binding determinants.
J. Mol. Biol.
353
:
556
-573.
22
Mazza, C., N. Auphan-Anezin, C. Gregoire, A. Guimezanes, C. Kellenberger, A. Roussel, A. Kearney, P. A. van der Merwe, A. M. Schmitt-Verhulst, B. Malissen.
2007
. How much can a T-cell antigen receptor adapt to structurally distinct antigenic peptides?.
EMBO J.
26
:
1972
-1983.
23
Kranz, D. M., D. H. Sherman, M. V. Sitkovsky, M. S. Pasternack, H. N. Eisen.
1984
. Immunoprecipitation of cell surface structures of cloned cytotoxic T lymphocytes by clone-specific antisera.
Proc. Natl. Acad. Sci. USA
81
:
573
-577.
24
Udaka, K., T. J. Tsomides, P. Walden, N. Fukusen, H. N. Eisen.
1993
. A ubiquitous protein is the source of naturally occurring peptides that are recognized by a CD8+ T-cell clone.
Proc. Natl. Acad. Sci. USA
90
:
11272
-11276.
25
Sykulev, Y., A. Brunmark, T. J. Tsomides, S. Kageyama, M. Jackson, P. A. Peterson, H. N. Eisen.
1994
. High-affinity reactions between antigen-specific T-cell receptors and peptides associated with allogeneic and syngeneic major histocompatibility complex class I proteins.
Proc. Natl. Acad. Sci. USA
91
:
11487
-11491.
26
Udaka, K., K. H. Wiesmuller, S. Kienle, G. Jung, P. Walden.
1996
. Self-MHC-restricted peptides recognized by an alloreactive T lymphocyte clone.
J. Immunol.
157
:
670
-678.
27
Colf, L. A., A. J. Bankovich, N. A. Hanick, N. A. Bowerman, L. L. Jones, D. M. Kranz, K. C. Garcia.
2007
. How a single T cell receptor recognizes both self and foreign MHC.
Cell
129
:
135
-146.
28
Degano, M., K. C. Garcia, V. Apostolopoulos, M. G. Rudolph, L. Teyton, I. A. Wilson.
2000
. A functional hot spot for antigen recognition in a superagonist TCR/MHC complex.
Immunity
12
:
251
-261.
29
Holler, P. D., P. O. Holman, E. V. Shusta, S. O'Herrin, K. D. Wittrup, D. M. Kranz.
2000
. In vitro evolution of a T cell receptor with high affinity for peptide/MHC.
Proc. Natl. Acad. Sci. USA
97
:
5387
-5392.
30
Holler, P. D., L. K. Chlewicki, D. M. Kranz.
2003
. TCRs with high affinity for foreign pMHC show self-reactivity.
Nat. Immunol.
4
:
55
-62.
31
Holler, P. D., A. R. Lim, B. K. Cho, L. A. Rund, D. M. Kranz.
2001
. CD8 T cell transfectants that express a high affinity T cell receptor exhibit enhanced peptide-dependent activation.
J. Exp. Med.
194
:
1043
-1052.
32
Jones, L. L., S. E. Brophy, A. J. Bankovich, L. A. Colf, N. A. Hanick, K. C. Garcia, D. M. Kranz.
2006
. Engineering and characterization of a stabilized alpha1/alpha2 module of the class I major histocompatibility complex product Ld.
J. Biol. Chem.
281
:
25734
-25744.
33
Garboczi, D. N., D. T. Hung, D. C. Wiley.
1992
. HLA-A2-peptide complexes: refolding and crystallization of molecules expressed in Escherichia coli and complexed with single antigenic peptides.
Proc. Natl. Acad. Sci. USA
89
:
3429
-3433.
34
Shusta, E. V., P. D. Holler, M. C. Kieke, D. M. Kranz, K. D. Wittrup.
2000
. Directed evolution of a stable scaffold for T-cell receptor engineering.
Nat. Biotechnol.
18
:
754
-759.
35
Garcia, K. C., C. G. Radu, J. Ho, R. J. Ober, E. S. Ward.
2001
. Kinetics and thermodynamics of T cell receptor- autoantigen interactions in murine experimental autoimmune encephalomyelitis.
Proc. Natl. Acad. Sci. USA
98
:
6818
-6823.
36
Weber, K. S., D. L. Donermeyer, P. M. Allen, D. M. Kranz.
2005
. Class II-restricted T cell receptor engineered in vitro for higher affinity retains peptide specificity and function.
Proc. Natl. Acad. Sci. USA
102
:
19033
-19038.
37
Holler, P. D., D. M. Kranz.
2003
. Quantitative analysis of the contribution of TCR/pepMHC affinity and CD8 to T cell activation.
Immunity
18
:
255
-264.
38
Otwinowski, Z., W. Minor.
1997
. Processing of X-ray diffraction data collected in oscillation mode.
Methods Enzymol.
276
:
307
-326.
39
Read, R. J..
2001
. Pushing the boundaries of molecular replacement with maximum likelihood.
Acta Crystallogr. D Biol. Crystallogr.
57
:
1373
-1382.
40
Emsley, P., K. Cowtan.
2004
. Coot: model-building tools for molecular graphics.
Acta Crystallogr. D Biol. Crystallogr.
60
:
2126
-2132.
41
Brunger, A. T., P. D. Adams, G. M. Clore, W. L. DeLano, P. Gros, R. W. Grosse-Kunstleve, J. S. Jiang, J. Kuszewski, M. Nilges, N. S. Pannu, et al
1998
. Crystallography & NMR system: A new software suite for macromolecular structure determination.
Acta Crystallogr. D Biol. Crystallogr.
54
:
905
-921.
42
Laskowski, R. A., M. W. MacArthur, D. S. Moss, J. M. Thornton.
1993
. PROCHECK: a program to check the stereochemical quality of protein structures.
J. Appl. Cryst.
26
:
283
-291.
43
Morris, A. L., M. W. MacArthur, E. G. Hutchinson, J. M. Thornton.
1992
. Stereochemical quality of protein structure coordinates.
Proteins
12
:
345
-364.
44
Garcia, K. C., M. D. Tallquist, L. R. Pease, A. Brunmark, C. A. Scott, M. Degano, E. A. Stura, P. A. Peterson, I. A. Wilson, L. Teyton.
1997
. Alphaβ T cell receptor interactions with syngeneic and allogeneic ligands: affinity measurements and crystallization.
Proc. Natl. Acad. Sci. USA
94
:
13838
-13843.
45
Donermeyer, D. L., K. S. Weber, D. M. Kranz, P. M. Allen.
2006
. The study of high-affinity TCRs reveals duality in T cell recognition of antigen: specificity and degeneracy.
J. Immunol.
177
:
6911
-6919.
46
Li, Y., Y. Huang, J. Lue, J. A. Quandt, R. Martin, R. A. Mariuzza.
2005
. Structure of a human autoimmune TCR bound to a myelin basic protein self-peptide and a multiple sclerosis-associated MHC class II molecule.
EMBO J.
24
:
2968
-2979.
47
Li, Q. J., A. R. Dinner, S. Qi, D. J. Irvine, J. B. Huppa, M. M. Davis, A. K. Chakraborty.
2004
. CD4 enhances T cell sensitivity to antigen by coordinating Lck accumulation at the immunological synapse.
Nat. Immunol.
5
:
791
-799.
48
Purbhoo, M. A., D. J. Irvine, J. B. Huppa, M. M. Davis.
2004
. T cell killing does not require the formation of a stable mature immunological synapse.
Nat. Immunol.
5
:
524
-530.
49
Cho, B. K., K. C. Lian, P. Lee, A. Brunmark, C. McKinley, J. Chen, D. M. Kranz, H. N. Eisen.
2001
. Differences in antigen recognition and cytolytic activity of CD8+ and CD8 T cells that express the same antigen-specific receptor.
Proc. Natl. Acad. Sci. USA
98
:
1723
-1727.
50
Daniels, M. A., S. C. Jameson.
2000
. Critical role for CD8 in T cell receptor binding and activation by peptide/major histocompatibility complex multimers.
J. Exp. Med.
191
:
335
-346.
51
Manning, T. C., C. J. Schlueter, T. C. Brodnicki, E. A. Parke, J. A. Speir, K. C. Garcia, L. Teyton, I. A. Wilson, D. M. Kranz.
1998
. Alanine scanning mutagenesis of an αβ T cell receptor: mapping the energy of antigen recognition.
Immunity
8
:
413
-425.
52
Marrack, P., J. P. Scott-Browne, S. Dai, L. Gapin, J. W. Kappler.
2008
. Evolutionarily conserved amino acids that control TCR-MHC interaction.
Annu. Rev. Immunol.
26
:
171
-203.
53
Borg, N. A., L. K. Ely, T. Beddoe, W. A. Macdonald, H. H. Reid, C. S. Clements, A. W. Purcell, L. Kjer-Nielsen, J. J. Miles, S. R. Burrows, et al
2005
. The CDR3 regions of an immunodominant T cell receptor dictate the “energetic landscape” of peptide-MHC recognition.
Nat. Immunol.
6
:
171
-180.
54
Lee, P. U., H. R. Churchill, M. Daniels, S. C. Jameson, D. M. Kranz.
2000
. Role of 2CT cell receptor residues in the binding of self- and allo-major histocompatibility complexes.
J. Exp. Med.
191
:
1355
-1364.
55
Dai, S., E. S. Huseby, K. Rubtsova, J. Scott-Browne, F. Crawford, W. A. Macdonald, P. Marrack, J. W. Kappler.
2008
. Crossreactive T Cells spotlight the germline rules for αβ T cell-receptor interactions with MHC molecules.
Immunity
28
:
324
-334.
56
Feng, D., C. J. Bond, L. K. Ely, J. Maynard, K. C. Garcia.
2007
. Structural evidence for a germline-encoded T cell receptor-major histocompatibility complex interaction “codon”.
Nat. Immunol.
8
:
975
-983.
57
Huseby, E. S., F. Crawford, J. White, P. Marrack, J. W. Kappler.
2006
. Interface-disrupting amino acids establish specificity between T cell receptors and complexes of major histocompatibility complex and peptide.
Nat. Immunol.
7
:
1191
-1199.
58
Zhao, Y., A. D. Bennett, Z. Zheng, Q. J. Wang, P. F. Robbins, L. Y. Yu, Y. Li, P. E. Molloy, S. M. Dunn, B. K. Jakobsen, et al
2007
. High-affinity TCRs generated by phage display provide CD4+ T cells with the ability to recognize and kill tumor cell lines.
J. Immunol.
179
:
5845
-5854.
59
Foote, J., H. N. Eisen.
1995
. Kinetic and affinity limits on antibodies produced during immune responses.
Proc. Natl. Acad. Sci. USA
92
:
1254
-1256.
60
Foote, J., H. N. Eisen.
2000
. Breaking the affinity ceiling for antibodies and T cell receptors.
Proc. Natl. Acad. Sci. USA
97
:
10679
-10681.
61
Jimenez, R., G. Salazar, J. Yin, T. Joo, F. E. Romesberg.
2004
. Protein dynamics and the immunological evolution of molecular recognition.
Proc. Natl. Acad. Sci. USA
101
:
3803
-3808.
62
Kourentzi, K., M. Srinivasan, S. J. Smith-Gill, R. C. Willson.
2008
. Conformational flexibility and kinetic complexity in antibody-antigen interactions.
J. Mol. Recognit.
21
:
114
-121.