F5 Lck^off CTLs lack killing ability. (A) Western blot of CTL lysates (7 d after stimulation) for F5 WT (1) and Lck^off (2) CTLs probed with antibodies as shown. (B) Immunofluorescence images displaying flattened z stacks of conjugated F5 WT and Lck^off CTLs with NP68-pulsed EL4 targets labeled for Lck (green, i and v), nuclei (blue, ii and vi), and CD8 (white, iii and vii). (C) Killing assay of WT (diamonds) and Lck^off (squares) F5 CTLs using NP68-pulsed EL4 as targets. Error bars show standard deviation from the means of triplicates. The assay is representative of over three independent experiments. Also see Fig. S1. Bars, 10 µm.

Lck is required for complete polarization and docking of the centrosome to the immunological synapse. (A) Merged immunofluorescence projections of centrosome polarization phenotypes seen in F5 CTL conjugates with the centrosome distal, proximal, or docked at the synapse (i.e., with γ-tubulin contacting the plasma membrane marker CD8), with cells labeled with antibodies against talin (green), γ-tubulin (red), and CD8 (white). Nuclei are stained with Hoechst (blue). (B) Representative images of centrosome polarization observed in WT and Lck^off CTLs. Every image (A and B) is a merge of four channels collected from z stacks. (C) Quantitation of centrosome polarization for F5 WT (n = 112) and F5 Lck^off (n = 103) from data analyzed in 3D showing the percentage of conjugates with the centrosome distal (blue), proximal (red), or docked (green) at the synapse. A two-tailed Student’s t test for loss of centrosome docking in Lck^off samples compared with WT gave a statistical significance of P = 10⁻⁴. Centrosome polarization was also quantitated in three independent experiments without 3D reconstruction (Fig. S3). Bars, 5 µm.

Actin clearance into the dSMAC is impaired in the absence of Lck F5 WT and Lck^off CTL conjugates with NP68-pulsed EL4 targets. (A) Quantitation of forming dSMACs from 3D reconstructions labeled for actin (WT, n = 121 and Lck^off, n = 110). (B) Representative 3D images of actin staining in WT and Lck^off CTLs in which actin has failed to clear (ii) or has partially reorganized (iii). (C) Quantitation from 3D reconstructions (WT, n = 117 and Lck^off, n = 119). (D) 3D images for pSMACs labeled for talin at the synapse. All 3D images are displayed at 0, 45, and 90° rotations. Bars, 10 µm.

Lck- and Fyn-deficient CTLs cannot polarize their centrosome toward antigen-presenting targets or anti-CD3 beads. (A) Killing assay of F5 WT (diamonds) and Lck^off Fyn^−/− (squares) CTLs using NP68-pulsed EL4 as targets. Error bars show standard deviations from triplicates. (B) Representative immunofluorescence images of centrosome polarization observed in WT and Lck^off Fyn^−/− F5 CTLs with NP68-pulsed EL4 targets. Cells are labeled with antibodies against talin (green), γ-tubulin (red), and CD8 (white). Nuclei were stained with Hoechst (blue). Bars, 10 µm. (C) Quantitation of centrosome position relative to the synapse in conjugates formed between EL4-pulsed targets and F5 Lck^off Fyn^−/− (n = 103) CTLs from 3D reconstructions from the same experiment shown in Fig. 3 C showing distal (blue), proximal (red), or docked (green) centrosomes at the synapse. A two-tailed Student’s t test for loss of centrosome docking in Lck^off samples compared with WT gave a statistical significance of P = 0.007. Lck^off Fyn^−/− formed conjugates with similar frequency (31%; n = 317) to WT CTLs (33%; n = 424). Centrosome polarization was also quantitated in three independent experiments without 3D reconstruction (Fig. S3). (D) Quantitation of centrosome position relative to anti-CD3 antibody–coated latex beads for F5 WT (n = 400), Lck^off Fyn^−/− (n = 400), and Lck^off Fyn^−/− + PP2 (n = 300) CTL conjugates. Error bars show the standard deviation from the means from at least three independent experiments. A two-tailed Student’s t test for loss of centrosome docking in Lck^off and Lck^off Fyn^−/− samples compared with WT gave statistical significances of P = 0.0006 and P = 0.007, respectively. (E) Representative images of F5 CTLs from WT, Lck^off Fyn^−/−, and Lck^off Fyn^−/− + PP2 conjugated to anti-CD3 antibody–coated latex beads labeled with Hoechst (blue) and antibodies against γ-tubulin (red) and anti–rat-FITC (green) viewed by differential interference contrast (DIC) and merged. Also see Fig. S2. Bars, 5 µm.

Lck is essential for degranulation and lytic granule polarization toward the immunological synapse. (A) Merged immunofluorescence images of three lytic granule polarization phenotypes seen in F5 CTL conjugates with lytic granules: (i) >90% of granules distal, (ii) dispersed, or (iii) tightly clustered at the synapse. Conjugates are labeled with antibodies against LAMP1 (red) or CD8 (white). Nuclei are stained with Hoechst (blue). (B) Representative images of granule polarization observed in (i) WT and (ii) Lck^off F5 CTL conjugates. (C) Quantitation of granule polarization in F5 WT (n = 465) and Lck^off (n = 484) CTLs for granules that were distal (blue), dispersed (red), or tightly polarized (green) at the immune synapse. Error bars show the standard deviation from the means of at least three independent experiments. A two-tailed Student’s t test for loss of granule polarization in Lck^off samples compared with WT gave a statistical significance of P = 5 × 10⁻⁶. (D) Histograms of a CTL degranulation assay displaying a CD8⁺ cell count against a LAMP1-PE signal upon activation with pulsed (shaded) or unpulsed (unshaded) EL4 targets. The percent increase in LAMP1 staining with a pulsed versus an unpulsed target is given. Data are representative of more than three independent experiments. Bars, 5 µm.

The centrosome fails to dock at the immunological synapse in Lck^off CTLs. (A–C) EM micrographs from thin (50–100 nm) lead-stained sections of nonmotile (A), motile (B), and target-conjugated (C) F5 WT or Lck^off CTLs loaded with HRP to reveal the endocytic pathway. The Golgi complex (G), lytic granules (white asterisks), secretory cleft (black asterisks), nuclei (N), and centrosome (arrows) are indicated in each image. Insets are magnified images of the centrosome area marked by white boxes. Bars: (main panels) 2 µm; (insets) 0.5 µm.

Lck-deficient CTLs can polarize their centrosome toward antigen-presenting targets or anti-CD3 beads. (A) Cartoon defining distal and proximal positions of the centrosome relative to beads or targets. (B) Representative images of conjugates formed between rat anti-CD3 latex beads and F5 CTLs from WT, Lck^off, and Lck^off + PP2 labeled with antibodies to γ-tubulin (red) and anti–rat-FITC (green) in separate channels, viewed by differential interference contrast (DIC), and merged. (C) Quantitation of conjugates from F5 WT (n = 400), Lck^off (n = 400), or Lck^off + PP2 inhibitor (n = 300) CTL conjugates. A two-tailed Student’s t test for differences between WT and Lck^off samples + PP2 revealed a statistical significance (P = 0.01). (D) Quantitation of centrosome polarization from OT-I CTL-bead conjugates. Error bars show the standard deviation from the means from at least three independent experiments. n > 300 conjugates. Bars, 5 µm.