Simulation systems constructed based on PDB 1BD2 (Ding et al., 1998).

Extension is the distance between the harmonic potentials on the terminal restrained atoms (blue spheres in Figure 1A), which was selected for B7low and B7high to yield average low and high loads among simulations scanning different extensions (see Methods). Average load is calculated between 500-1000 ns. The standard deviation (std) in load as measured in 40-ns intervals between 500-1000 ns is shown in parentheses. Their values are consistent with those for A6 (Figure 1—figure Supplement 2). The extension and force (average±std) for A6 corresponding to B7low and B7high are 182.6 Å and 13.2±5.65 pN, and 187.7 Å and 18.2±9.17 pN, respectively (Chang-Gonzalez et al., 2024), which indicates that B7 is more compliant compared to A6. For B70, B7low, and B7high, simulations were further extended for additional time-dependent stability analysis. However, averaging was done for the 500-1000 ns interval for consistency.

B7 TCR-pMHC interface. (A) Overview of the base complex used in simulations. The 4 subdomains of TCRαβ are Vα, Vβ, Cα, and Cβ. Load was applied by holding the Cα atoms of terminal residues (blue spheres at the ends of “added strands”) at a given distance from each other. β2m: β2 microglobulin. (B) Magnified view of red box in panel A showing labeled CDR loops and side chains of peptide residues in stick representation. (C) Number of contacts with greater than 50% average occupancy and 80% maximum instantaneous occupancy from 500-1000 ns. Bars: std. Criteria for counting contacts and values for A6 are from Chang-Gonzalez et al. (2024). (D) Total contact occupancy measured in 40-ns overlapping intervals. TCR-pMHC (top; intermolecular) and intra-TCR (bottom; intramolecular) contacts are shown separately. Intra-TCR contacts exclude Cα-Cβ contacts (Methods). Circles with outline: B7high; without outline: B7low. Horizontal bar below each panel: B70. Figure 1—figure supplement 1. Model of TCRαβ with leucine zipper domain fused to the C-terminal added strands. Figure 1—figure supplement 2. Standard deviation vs. average of the applied force.

Load dependent behavior of the B7 TCR-pMHC interface. (A) Number of MHC-Vα, MHC-Vβ, peptide-Vα, and peptide-Vβ contacts in B7 and A6 TCRs between 500-1000 ns. Data for A6 are from Chang-Gonzalez et al. (2024). (B) Hamming distance ℍ. Histograms were calculated using data between 500-1000 ns (marked by vertical dashed lines), the interval between the dashed lines. (C) Location of V-module residues forming contacts with pMHC with greater than 50% average occupancy. The frame at 1000 ns is used for visualization. The backbone of the Tax peptide is shown as a purple tube. CDRs are labeled in the first panel. (D) Total (pink) and per-residue (blue) BSA for interfacial residues between 500-1000 ns. (E) pMHC residues forming contacts with the V-module with average occupancy greater than 50% in the high-load case. Left: B7high, right: A6high. MHC residues are shown as sticks and Cα atoms of the peptide residues are shown as spheres. Viewing direction is the same as in panel C. Figure 2—figure supplement 1. Additional characterization of the TCR-pMHC interface.

Vα-Vβ motion of B7. (A) Number of Vα-Vβ contacts with greater than 50% average occupancy and 80% maximum instantaneous occupancy between 500-1000 ns. Bars: std. (B) V-module triads {e1, e2, e3}. Arrows denote directions of the first 3 PC modes for B7high as an example. CDR3s are labeled. (C) Amplitudes for the first 6 PCs. PCA was performed between 500-1000 ns. Transparent bands: std for PCA performed in three overlapping intervals (500–800 ns, 600–900 ns, and 700–1000 ns). (D) Histograms of the V-module triad angles. (E) CDR3 distance vs. triad angles. Transparent bands: std. Both D and E were determined from data between 500-1000 ns.

V-C motion of B7. (A) Number of V-C contacts per chain measured with the same criteria as Figure 3A. (B) Average BOC for B7low and B7high. The V-module of B7high is less bent compared to B7low. The arrows for the first 3 V-C PC modes are shown, where PC1 corresponds to the V-C bending motion. (C) V-C PC amplitudes for the first 6 PC modes. Transparent bands: std measured in the same way as in Figure 3C. (D) Differences in amplitudes for the first 3 PCs between matching V- and H-beads of α and β chains. (E) Histograms of hinge angles defined in panel B. (F) CDR3 distance versus hinge angles. Transparent bands: std. All panels are generated with data between 500-1000 ns. Figure 4—figure supplement 1. Additional characterization of the V-C motion.

Model of TCRαβ with leucine zipper domain fused to the C-terminal added strands. Model was built based on PDB 1NFD (Wang et al., 1998). The construct was used in single-molecule optical tweezers experiments (Das et al., 2015). Blue spheres denote Cα atoms of residues corresponding to αT218 and βA259 of B7 TCR (Figure 1A, bottom). Their distance is 6.7 Å. The 10-Å flat-bottom harmonic distance restraint on those atoms (see Methods) mimic the presence of the leucine zipper that prevents large separation of the added strands.

Standard deviation vs. the average of the applied force. All data are measured between 500-1000 ns. Open circles are for A6 (Chang-Gonzalez et al., 2024). Red squares for B7low and B7high are from Table 1. Except for Y8Alow (an antagonist) and dFGlow (Cβ FG-loop deletion mutant) for A6, the near-linear relation between the std and average in force is a consequence of the force being applied to the restraining potential via random conformational fluctuation of the complex (Burgess, 1973). See Appendix 3 of Chang-Gonzalez et al. (2024) for further explanation. The fact that data for both B7 and A6 lie on approximately the same line also reflects consistency in our force measurement.

Additional characterization of the TCR-pMHC interface. (A–D) Contact occupancy heat maps. (A) B70, (B) B7low, (C) B7high, and (D) Vαβ-pMHC. Contacts with over-all occupancy greater than 30% and instantaneous occupancy greater than 80% are shown. hb: hydrogen bond, np: nonpolar contact. Arrow in panel D denotes the approximate time when the initial adjustment of contacts in Vαβ-pMHC happens (Figure 2B). (E) RMSF of peptide backbone Cα atoms between 500-1000 ns. Cα atoms were aligned to those at the beginning of the production run for RMSF calculation. (F) Distance between the V-module and pMHC. Histograms were calculated using data between 500-1000 ns, the interval between the vertical dashed lines. For Vαβ-pMHC and B7high, avg±std between 500-1000 ns are 32.5±0.31 Å and 31.9±0.30 Å, respectively. V-module to pMHC distance measured from

Additional characterization of the V-C motion. (A) Dot products computed between the BOC PC vectors of listed systems. Values closer to 1.0 denote similar V-C BOC direction of motion. (B) V-C hinge angle trajectories over time.