Figure 1. Strategy for testing kinetic proofreading with optogenetic tools.

(A) Conventional methods of mutating the pMHC to alter the binding half-life also change the binding interface, which changes several parameters at once. By contrast, optogenetic control allows light intensity to control ligand binding half-life while keeping the binding interface constant. Therefore, no other aspects of the receptor-ligand interaction change. Red and blue lines highlight the binding interfaces. (B) Schematic of experimental setup. Jurkat cells expressing a live cell DAG reporter and a Zdk-CAR are exposed to an SLB functionalized with purified, dye-labeled LOV2. In the dark, LOV2 binds to and accumulates under the receptor, and stimulates DAG production, recruiting the reporter to the plasma membrane. Blue light excites LOV2, inducing its dissociation from the receptor to terminate signaling. (C) Montage from a time course in which cells were alternately stimulated in the presence or absence of blue-light. White arrows highlight cells with low to undetectable receptor occupancy, and red arrows highlight cells with high receptor occupancy. (D and E) Cells with very different levels of receptor occupancy (D), can have similar DAG levels (E), suggesting that receptor occupancy is not a good predictor of DAG levels. Top blue bars indicate the presence of blue light.

Figure 1—figure supplement 1. Light-based control of T cell signaling is durable for hours.

(A) Clonal Jurkat cells expressing both the Zdk-CAR and DAG reporter were exposed to periodic pulses of blue-light (blue bars near the x-axis) for five hours while on SLBs functionalized with LOV2. Plotting data from cells that were signaling at the end of the time course shows no apparent diminishment in the per-cell activity of ligand binding or DAG signaling over time. (B) Plotting data from all cells in the time course shows a modest reduction in ligand binding and DAG signaling. This was driven mostly by gradual cell contraction and withdrawing from contact with the SLB. Note that all experimental time courses in the paper are no longer than 60 min, a duration over which no appreciable decrease in ligand binding or DAG signaling is observed. DAG levels reported as mean TIRF561 instead of fold-above PP2 background, as cells were not treated with PP2 after the time course. (C) Cells adhered to SLBs without LOV2 show no light-dependent changes in either CAR occupancy or DAG levels upon blue light illumination. Cells were passively adhered to the SLB with a biotinylated anti-β2 microglobulin antibody. Error bars for CAR occupancy are too small to be visible. The full time course is shown in Video 4. All plots are mean with a 95% CI (two-sided Student’s t-test).

Figure 1—figure supplement 2. Cells spread in response to blue-light illumination.

In addition to increasing DAG production, antigen stimulation leads to increased cell spreading in T cells. We observe similar cell spreading following optogenetic stimulation of the CAR. Cell area was measured over time using reflection interference contrast microscopy (RICM). Cells show larger increases in cell area in response to longer LOV2 binding half-lives, providing another indicator of cell stimulation in addition to DAG production. Because cells slowly contract over the imaging time course, we show the change in cell area during each pulse of blue light. The baseline drift was measured by measuring the absolute cell area during the strong ‘control’ pulses of blue light regularly spaced throughout the time course (see Figure 2—figure supplement 2). This baseline drift was subtracted from the absolute cell area to yield the change in cell area. See Video 5 for the complete animated time course. n = 31 cells. Mean with 95% CI (two-sided Student’s t-test).

Figure 1—figure supplement 3. Colocalization of ligand binding and downstream signaling.

Higher magnification TIRF images show how ligand binding to the CAR spatially relates to downstream signaling reporters. Jurkat cells expressing the CAR and either a ZAP70-mCherry or DAG reporter were imaged on supported lipid bilayers functionalized with LOV2. All LOV2-Alexa488 molecules are revealed by TIRF, however brighter green regions indicate where LOV2 accumulates because of binding the CAR. A ZAP70-mCherry reporter (purple) colocalizes well with sites of LOV2 binding, which is expected as ZAP70 directly binds the phosphorylated ITAMs of the CAR. The DAG reporter (purple) shows a roughly diffuse localization throughout the cell which is consistent with DAG diffusing away from sites of production. The LOV2-CAR interactions and reporter activities are only observed within the cell footprint (RICM). Imaged with a Nikon 1.49NA 100x Apo TIRF objective.

Figure 1—figure supplement 4. Absolute quantification of cell surface TCRs and CARs.

The number of cell surface exposed TCRs and CARs was measured using beads with known antibody binding capacities and flow cytometry. (A) To ensure measurements were made at saturating antibody concentrations, high concentrations of antibodies were used. Saturation was verified by serially diluting each antibody individually with resulting negligible changes in MFI. Staining under saturating antibody conditions are shown for the primary antibodies in red and the secondary antibody in blue. The primary antibodies OKT3 and OKT8 were used to stain for the TCR and CAR, respectively, while a goat anti-mouse Alexa488 antibody was used as the secondary. Three 2-fold serial dilutions of each antibody are shown in lighter shades, which nearly overlap the saturating condition. Light gray indicates no primary antibody was used. (B) The absolute number of cell surface TCRs or CARs were calculated for each cell line based on the MFI of the population. Because the bead standards bind a known number of IgG primary antibodies and each primary antibody could bind between one to two receptors, we report a two-fold range for the number of each receptor.