T cell receptor recognition of CD1b presenting a mycobacterial glycolipid

Abstract

CD1 proteins present microbial lipids to T cells. Germline-encoded mycolyl lipid-reactive (GEM) T cells with conserved αβ T cell receptors (TCRs) recognize CD1b presenting mycobacterial mycolates. As the molecular basis underpinning TCR recognition of CD1b remains unknown, here we determine the structure of a GEM TCR bound to CD1b presenting glucose-6-O-monomycolate (GMM). The GEM TCR docks centrally above CD1b, whereby the conserved TCR α-chain extensively contacts CD1b and GMM. Through mutagenesis and study of T cells from tuberculosis patients, we identify a consensus CD1b footprint of TCRs present among GEM T cells. Using both the TCR α- and β-chains as tweezers to surround and grip the glucose moiety of GMM, GEM TCRs create a highly specific mechanism for recognizing this mycobacterial glycolipid.

Figure 1. TRAV1-2 TCRs recognize CD1b, MHC-I and MR1.

Overview of the TRAV1-2 TCRs in complex with CD1b (coloured white; a,b,g), MHC-I (coloured dark grey; c,d,h) and MR1 (coloured light grey; e,f,i) molecules. The top panels show the overview of each complex represented in cartoon format with the antigen in black spheres. The TRAV1-2 gene segment is coloured in light pink, the TRAJ gene segment in vibrant pink, and the β-chain in blue, orange and green for GEM42 TCR (a,b), ELS4 TCR (c,d) and MAIT TCR (e,f), respectively. A schematic of each TCR gene segment is represented as two rectangles for the α and β-chains, with TRAV1-2, TRAJ and the β-chains coloured as per the top panels. The middle panels show the footprint of each TCR on the surface of the CD1b-lipid (b), MHC-peptide (d) and MR1-metabolite (f). The black spheres on the middle panels represent the centre of mass of the Vα and Vβ domains, while the light grey spheres represent the antigen bound in each molecule. The atomic footprint is coloured according to the TCR segment making contact. The bottom panels show the TRAV1-2 gene segment (light pink) contact with (g) CD1b-lipid (white), (h) MHC-peptide (dark grey) and (i) MR1 (light grey).

Figure 2. GEM TCR footprint on CD1b–GMM.

(a) Footprint of the GEM42 TCR on the surface of CD1b (white) and GMM (pale orange spheres) is represented according to the atoms contacted and coloured as per the TCR segment making contact. Framework residue from the α-chain in pale pink, the CDR loops coloured in teal (CDR1α), green (CDR2α) and purple (CDR3α) for the α-chain and red (CDR1β), orange (CDR2β) and yellow (CDR3β) for the β-chain, respectively. Pink and blue spheres represent the centre of mass of the GEM42 TCR α and β-chains, respectively. The insert below the footprint represents the characteristic CDR3α loop sequence of the GEM TCR. GEM42 TCR interactions with the CD1b (panels b–f), with the CD1b in white, GMM in pale orange and the GEM 42 TCR coloured as per panel (a). The panels represent residues from the (b) the CDR1α (teal); (c) CDR2α (green) and framework residue from α-chain (pale pink); (d) CDR3α (purple); (e) CDR1/2β (red and orange) and framework β-chain (pale blue); (f) CDR3β (yellow) interacting with the CD1b molecule (white). Hydrogen bonds are shown as red dashed lines, and the sphere represents the Cα atom of the glycine 29α residue.

Figure 3. Molecular tweezers grip the glucose headgroup of GMM.

(a) Surface representation of the CDR3α (purple) and CDR3β (yellow) loops acting as ‘tweezers' grasping the GMM glucose head group (pale orange stick). (b) Schematic of the GMM analogues tested against the GEM T cells. Red shows the different hydroxyl orientation of Galactose monomycolate (MM) and Mannose MM. Each GMM analogue was used to stimulate IL-2 production upon presentation to GEM clone18 (black bar) and GEM clone 42 (pink bar), error bar representing triplicate wells. The experiments was performed twice with similar results. (c) Interaction network between the GMM (pale orange stick) and the GEM42 TCR (coloured as per Fig. 2a). The red dashed lines represent hydrogen bonds, while the blue dashed line represent hydrophobic interactions. (d) CDR3α loop sequences of the GEM TCR, element from the TRAV1-2 gene are highlighted in sand colour, N region in green and TRAJ9 gene in pink, while the CDR3β loop is coloured in yellow as per panel c.

Figure 4. CD1b–GMM plasticity upon GEM TCR engagement.

(a) Superposition of the GEM42 TCR free (blue) and bound to CD1b–GMM (pink) structures. (b–d) Superposition of the binary CD1b–GMM (green) and bound to GEM42 TCR (orange). GMM is represented as stick, and CD1b in cartoon format, while the sphere represents the Cα atom of the glycine 153 residue. The orange and green colours highlight the segment of the CD1b molecule that change conformation upon binding of the GEM42 TCR, namely residues 147–154 of the α2-helix (b,c) and the aromatic residues of the F′-pocket Tyr151, Phe144, Phe88 and Phe84 (d).

Figure 5. Energetic hotspot of GEM TCR–CD1b–GMM recognition.

(a) Surface plasmon resonance (SPR) sensogram of CD1b–GMM interacting with a range of GEM42 TCR concentrations (top graph represented by the purple and pink curves), as well as the sensogram of GEM42 TCR passed over an empty flow cell used as control (blue curve), CD1b loaded with endogenous lipid (orange curve) or GMM (pink curve). (b,c) SPR on the CD1b mutants loaded with GMM with the GEM42 TCR (b) and GEM21 TCR (c). The experiments were performed in duplicate (n≥2), with the error bar representing the standard deviation. The y-axis represents the ΔΔGeq (kcal mol⁻¹), the error bar represents standard deviation, the orange line indicates a 3-fold increase compared the wild type protein; the red line indicates a 5-fold increase; and the green line represents a 3-fold decrease. The x-axis indicates the position of CD1b mutation, and the bars are coloured according to the impact of each mutation on the affinity of the TCRs, ranging from 3-fold improvement (green), no impact (yellow), 3-fold decrease (orange) and more than 5-fold decrease (red). (d,e) Surface representation of CD1b (white) and GMM (pale orange stick) with the mutated residues coloured as per their impact on the GEM42 (d) and GEM21 TCR binding (e), coloured as per panels (b,c).

Figure 6. CD1b–GMM binding of Polyclonal GEM T-cells from TB patients.

(a) Mean fluorescence intensity (MFI) of GEM clone 42, stained with untreated and GMM-loaded wild type CD1b tetramers and with eleven GMM-treated CD1b proteins, where the indicated residue was mutated to alanine. T cells from latent tuberculosis donors C58 and A22 were stimulated for one round with GMM and stained. CD3⁺TRAV1-2⁺ stimulated T cells were then used against our panel of CD1b mutants tetramer loaded with GMM. FACS plot of wild-type CD1b without GMM (WT mock) and with GMM (WT GMM), as well as for eleven CD1b mutants loaded with GMM is represented for both donors. (b) MFI for the GEM clone 42 or for polyclonal T cells from two donors is represented for either wild type CD1b (WT) with (+) or without (−) the presence of GMM, as well as for each CD1b mutant with (+) GMM as per the x-axis description. MFI and percentage GEM T cells were determined and multiplied to obtain the integrated MFI (iMFI) from the both donors. Relative MFI or relative iMFI was calculated by dividing the value obtained with a mutant tetramer by the value obtained with the wild type CD1b tetramer.