The ZAP-70 Double Src Homology 2 (SH2) Domain-GFP Fusion Protein Allows Reliable Visualization of the Localization of Ligand-Engaged P14 T Cell Receptor (TCR). To specifically determine the location of phosphorylated TCR, we used a ZAP-70 tandem SH2-domain-GFP fusion construct (Z2SH2-GFP) retrovirally expressed in primary P14 T cells. In the presence of 10 m M gp33 agonist peptide, Z2SH2-GFP rapidly accumulated at the center of the T cell/antigen-presenting cell (APC) interface in 68% of the cell couples (Fig. 2 A and B). Such accumulation was dramatically reduced at lower agonist peptide concentrations (Fig. 2A). A similar construct of the SHP-1 2SH2 domain did not show any accumulation (data not shown), establishing specificity. No Z2SH2-GFP accumulation was seen when the APCs were loaded with 10 m M of the partial agonist peptide A4Y (Fig. 2A). The pattern and kinetics (Fig. 9) of Z2SH2-GFP accumulation, the frequency of central accumulation, and the dependence on agonist peptide thus were comparable to our own and published data (1–3) on TCR accumulation using TCR-z -chain-GFP and MHC-GFP fusion proteins (Fig. 2A), corroborating Z2SH2-GFP functionality. One potential concern by using Z2SH2-GFP is the dominant negative effects seen in overexpression studies (4). Comparing effector functions of GFP- versus Z2SH2-GFP-transduced P14 T cells, we found indistinguishable target cell lysis (Fig. 11) and IFN-g secretion (data not shown) over a wide range of agonist peptide concentrations. In contrast, a similar analysis of HY target cell killing did show a dominant negative effect of Z2SH2-GFP at all agonist peptide concentrations (data not shown), consistent with the lower affinity of the HY TCR. For the determination of HY TCR localization, a TCR-z -chain-GFP fusion protein was therefore used instead. In summary, a retrovirally expressed ZAP-70 double SH2-GFP fusion protein allowed the reliable determination of the localized TCR signaling in primary P14 T cells.

Data Normalization and Probes for the Visualization of TCR/MHC Accumulation. A direct comparison of various T cell functions requires normalization. Normalization of T cell effector functions was straightforward because they were analyzed numerically in bulk assays. Raw data and normalization routines are given in Table 1. For comparability of results between T cells from different TCR transgenic systems, effector functions of T cells from different TCR transgenic mouse strains were often analyzed in a single assay. No significant differences in effector functions trigged by maximal/minimal agonist peptide between transgenic models were observed. The normalization of single-cell imaging data were more difficult. TCR accumulation was defined here and in refs. 2 and 5 as fluorescence intensity at the T cell/APC interface ³ 40% above cellular background. This threshold reliably distinguishes random cell surface fluorescence fluctuations from APC contact-induced accumulation. Each cell couple was scored for the occurrence of accumulation and its geometrical pattern (as defined in refs. 2 and 5) at 1, 3, and 5 min after cell couple formation. A final time point is taken as late as possible between 7 and 15 min after cell couple formation. Central TCR accumulation either occurred at least once (scored central), or only other patterns (other) or no accumulation (none) were observed, thus generating a digital readout for each cell couple that can be compiled and analyzed statistically. Because our digital scoring disregards the dynamics of TCR accumulation, they are separately presented in Fig. 9 for all conditions where substantial central accumulation occurred. A potential alternate numerical enumeration of TCR accumulation, e.g., through percentage increase of fluorescence intensity at the interface, is currently not feasible for a study involving hundreds of cell couples because of limitations in the acquisition of suitable raw data and its quantitative 4D analysis.

The scoring disregards the nature of the imaging probe used. To investigate the localization of TCR signaling, probes have to selectively visualize ligand-engaged TCR. MHC-GFP (6) and ZAP-70 2SH2-GFP accomplish this. Unfortunately, ZAP-70 2SH2 interferes with TCR signaling in HY and 5C.C7 T cells (DO11.10 was not evaluated) and H2-D^b-GFP shows substantial intracellular expression in EL4 cells. This intracellular expression is likely responsible for a reduced percentage of P14/EL4 cell couples with detectable TCR clustering (Fig. 2A). TCR-z -GFP was therefore used as an alternate approach. By generating data from probes that are selective for ligand-engaged TCR (ZAP-70 2SH2/P14 and HY; H2-D^b–GFP/P14; I-E^k-GFP/5C.C7) and showing that they are similar to data from TCR-z -GFP (HY, DO11.10) (Fig. 2A), we support that centrally accumulated TCR is actively signaling in all TCR transgenic models used. Comparability of the different probes was ascertained by repeatedly applying several to the same TCR transgenic T cells with similar results (Fig. 2A and data not shown). In addition, fluorescence intensities of the various probes used were similar. The alternate approach, employing z -chain–GFP as a single probe to all TCR transgenic systems, would visualize TCR localization without regard for its engagement by peptide/MHC.

Visualizing the Accumulation of Ligand-Engaged CD28/CTLA-4 with B7-2–GFP. To specifically visualize the localization of ligand-engaged CD28 in 5C.C7 T cell/APC couples, we used a strategy (6) similar to the visualization of ligand-engaged TCR. Instead of tagging the receptor, CD28, itself, as has already been done (7), we labeled the ligand B7. However, two complications require consideration. First, CD28 has two ligands, B7-1 and B7-2, both of which also bind to CTLA-4 on the T cell surface. Lack of B7-2 on the APC surface interfered with the accumulation of CD28–GFP in 5C.C7 T cell/APC couples, the lack of B7-1 did not (8). Although CTLA-4 binds B7-1 and B7-2 more firmly than CD28, the affinity advantage of CTLA-4 over CD28 is smaller in binding to B7-2 (2.6 m M vs. 20 m M K_d) than in binding to B7-1 (0.2 m M vs. 4 m M K_d) (9). We therefore used a B7-2–GFP fusion protein (C.W., unpublished data) to visualize ligand-engaged CD28. To assess the contribution of CTLA-4 to B7-2–GFP localization, we blocked CTLA-4 with 10 m g/ml Fab fragments from the anti-CTLA-4 antibody 4F10. B7-2–GFP accumulation at the 5C.C7 T cell/CH27-B7-2–GFP APC interface (Fig. 7) was reduced but not blocked, indicating a partial contribution. Second, in contrast to the A20/I-E^k–GFP APCs (2) used to determine the localization of ligand-engaged TCR that don’t express endogenous I-E^k, CH27 cells express endogenous B7-1 and B7-2. Competition of B7-2–GFP with endogenous B7 for CD28 binding in combination with the low affinity of the B7-2/CD28 interaction (20 m M K_d; ref. 9) was likely responsible for the smaller percentage of 5C.C7 T cell/APC couples showing B7-2–GFP accumulation (Fig. 7) than that showing CD28–GFP accumulation (7). Nevertheless, central accumulation of ligand-engaged CD28 at the 5C.C7 T cell/APC interface was observed, allowing a reliable determination of an MCC agonist peptide dose–response (Fig. 1A).

1. Monks, C. R., Freiberg, B. A., Kupfer, H., Sciaky, N. & Kupfer, A. (1998) Nature 395, 82–86.

2. Wülfing, C., Sumen, C., Sjaastad, M. D., Wu, L. C., Dustin, M. L. & Davis, M. M. (2002) Nat. Immunol. 3, 42–47.

3. Krummel, M. F., Sjaastad, M. D., Wülfing, C. & Davis, M. M. (2000) Science 289, 1349–1352.

4. Qian, D., Mollenauer, M. N. & Weiss, A. (1996) J. Exp. Med. 183, 611–620.

5. Tskvitaria-Fuller, I., Rozelle, A. L., Yin, H. L. & Wülfing, C. (2003) J. Immunol. 171, 2287–2295.

6. Wülfing, C., Tskvitaria-Fuller, I., Burroughs, N. J., Sjaastad, M. D., Klem, J. & Schatzle, J. D. (2002) Immunol. Rev. 189, 64–83.

7. Egen, J. G. & Allison, J. P. (2002) Immunity 16, 23–35.

8. Pentcheva-Hoang, T., Egen, J. G., Wojnoonski, K. & Allison, J. P. (2004) Immunity 21, 401–413.

9. van der Merwe, P. A. & Davis, S. J. (2003) Ann. Rev. Immunol. 21, 659–684.