The α/β component of the T-cell antigen receptor (TCR) is a transmembrane heterodimer containing genetically variable (V) domains and constant domains within each polypeptide chain.1 The Vα and Vβ domains bind with a clonally unique (clonotypic) interface angle,2 and the complementary determining regions (CDR1, CDR2, and CDR3) manifest as ‘loops’ between anti-parallel β-strands of the overall sandwich of β-sheets with a central disulfide bond. As with immunoglobulins, the CDR1 and CDR2 loops are encoded in the germline via the particular V-region DNA segment involved in the RAG1/RAG2 recombination mechanism responsible for somatic construction, together with the D- and/or J-segment, of the third loop, CDR3.1,3 As shown by multiple solved X-ray crystallography structures, the CDRs contain the closest amino acid (a.a.) contacts with the peptide (p) plus the major histocompatibility complex (MHC) protein (together abbrev., pMHC) composite ligand. After rigorous selection in the thymus for only those TCRs that should function properly in the particular individual after birth (MHC restriction),4 T cells function clonally as well, i.e., different T-cell clones display differential responses to not only different pMHC ligands but also perhaps more importantly, to the same pMHC ligands.5,6 A long-standing question is how (not whether) TCR V-domain binding to pMHC is ultimately responsible for such vast biological outcomes as protection vs. susceptibility to infectious diseases, atopy, the onset of autoimmune diseases, and the success of various types of transplantation.

Groups have pursued studies that experimentally place a small force (≈10–20 pN) on isolated TCRs to show that binding to pMHC is surprisingly altered. Now, a September 2020 paper (Hwang et al.) suggests conventional physicochemistry for how this force operates to engender so-called catch bonds at the TCR:pMHC interface.7 One caveat is that the MD simulation approach is orders of magnitude too short compared to observed conformational change in TCR binding to pMHC. Nevertheless, the paper does, for the first time, identify how such an applied force might slow V-domain motions (dynamics), particularly twisting, tilting, and swaying over the pMHC ligand.7,8 Interface contact duration,7 hydrogen (H) bonding networks,8 together with entropy effects9 (e.g., resolvation and exclusion of water molecules10), might be altered if V-domain dynamics are inhibited by the load (force). Most importantly, the force that ostensibly occurs when a T cell surveys the antigen-presenting cell (APC) surface indirectly confronts the issues of TCR selectivity and specificity. Clearly, the TCR somehow recognizes antigenic (target) pMHC(s) on the APC despite thousands of ‘dummy’ (nonagonist) pMHC with ubiquitous micromolar-affinity TCR. Thus, this reported tangential force on the TCR might ‘specialize’ target (agonist) pMHC recognition without changing the equilibrium binding affinity a TCR possesses against nonagonist pMHC7.

A November 2020 paper (Bartleson et al.) further investigates so-called tonic TCR signaling, i.e., by post-thymic, self-peptide:self-MHC complexes.6 In adoptive transfer experiments, T cells from two different TCR-transgenic (Tg) mouse strains were analyzed after antigenic challenge in vivo. The two TCRs recognize the same target pMHC-II with the same affinity, but there is a large difference in IL-2 produced by the T cells from the two strains in response to target pMHC-II. One TCR delivers high tonic signaling (with high IL-2), while the other TCR shows low tonic signaling (with low IL-2). Previously, it was shown that this difference was not hard-wired but required thymic positive selection.4,6 Here, they pinpoint that these tonic interactions are key to preferential expansion of the follicular helper (TFH) subset (crucial to T:B collaboration leading to a strong antibody response). Interestingly, they determined that it is not the IL-2 production difference between the two TCR-Tg precursors that determines the difference in TFH outcomes but that a low-tonic TCR drives a higher frequency of TFH cells than a high-tonic TCR. Although the low-tonic TCR is associated with lower IL-4 on a cell population basis, they have a higher per cell production of IL-4—which is the hallmark cytokine of helper T-cell function—as described in previous model systems.11 A measured low density of antigenic pMHC-II presented on APCs in vivo can drive a Th1 to Th2 switch in clonal dominance, but only in permissive MHC genotypes.12 Thus, determinant selection at the level of peptide binding (including the binding register) to particular class-II MHC allotypes can control T-cell function,5,12 even within a repertoire of micromolar affinity TCR.7,13 Strikingly, what appears to make the recognition of the agonist pMHC-II special in turning on TFH function is the low-tonic TCR’s previously reduced engagements with self pMHC.6

Starting with the goal of a comprehensive geometrical analysis of V domains in solved TCR:pMHC-II crystallographic structures, we subsequently entered into the study of spherical coordinates and the multivariable calculus of V-domain dynamics.8,14 Each of a panel of 38 V domains in 19 structures was described in terms of three Euler angles defining their twist (ω), tilt (λ), and sway (σ) atop pMHC-II. Fixed a.a. positions were used across all the structures, such that what became obvious is that the exact geometry of each V domain is in fact clonotypic, i.e., each TCR in the panel had unique V-domain geometry. It was also evident that a fourth angle, pitch (φ), was also clonotypic. Here, pitch was measured by drawing a vector from the CDR2 a.a. making closest contact to the MHC α-helix across the peptide groove to the opposite MHC α-helix a.a. in closest contact with CDR3, then a second vector from this a.a. across the groove to the MHC α-helix a.a. in contact with said CDR2 a.a.—thus defining the measured pitch angle (φm). Remarkably, we could also calculate the pitch angle (φc) from the twist, tilt, and sway angles (orientation) by the general equation:

$$\small \varphi_{\mathrm{c}} = {\mathrm{k}}\varphi_{{\mathrm{m}}} = \left[ {\sigma \div \left( {\lambda + \sigma} \right)} \right]\omega,$$

where the value ‘k’ gives the difference between the calculated and measured pitch values. On average, k was close to 1.00, indicating the predictive nature of the equation. However, some V domains have significantly higher or lower k values, suggesting induced fit,10 specifically regarding pitch. The calculated pitch might be thought of as a given V domain’s optimum pitch, raising the hypothesis that it is a ‘hidden’ correlate of TCR positive selection in the fetal thymus4,8 (see below). Since the range of CDR2 closest contact positions on the MHC α-helix defined an ~10 a.a. stretch of the MHC α-helix ‘swept’ by CDR2, we modeled V domain rotation using trigonometry and conversion to spherical coordinates—solving separate triple integrals for each V domain; the general equation is:

$$\small \int \int \int _VdV = {\int_0^{\sin ^{ - 1}\left( {r/\rho } \right)}} {{\int_0^{\pi \cdot r/4}} {{\int_0^{z/\!\cos \phi }} {\rho ^2\sin \phi \,d\rho \,d\theta \,d\phi } } }$$

This is a ‘slice’ volume element of a cone, where for each of the 38 V domains, a measured distance was taken between the central C22/23 a.a. and the MHC α-helix a.a. used to define orientation—this is the rho (ρ) distance substituted into each equation. The other measure was the angle (ϕ), which is simply the difference from 180° of the previously determined tilt (angle λ) for each V-domain; other variables were derived by trigonometry.8,14 Thus, each cone-slice volume (dV) defines a probability of CDR2 sweeping/scanning of the MHC α-helix region (dθ), which was (as predicted) unique for each V domain. Interestingly, there was a linear relationship between the calculated pitch (φc) and dV across all the V domains. This indicates that a ‘flush’ (flat) V-domain pitch limits CDR2 scanning. From k, we interpret that V domains can alter the pitch and maintain dV, further suggesting that a precise ‘recognition’ dV might be set during the positive selection of a given TCR. Rare (3/38) V domains had significantly restricted dV values by ~75% of the average dV for the panel. H-bonding mechanisms revealed charge relays across all three components of the interface (Vβ, peptide, MHC-II), which distinguished these ‘highly restricted’ dV domains from all others. Thus, a CurtinHammett type15 of mechanism might operate for V-domain binding to target pMHC, where a stabilized transition state of CDR3 binding can be used to recapitulate dV and drive the overall reaction, i.e., a dynamics alternative to a relative TCR affinity-based mechanism.8

Today, immunology is built on a foundation that still misses the piece of molecular TCR function. These three papers introduce three distinct yet mutually nonexclusive ‘irons-in-the-fire’, but where it will go from here is hard to imagine.