Nature Immunology homepage
Advanced search
 Journal home > Archive > Table of Contents > Article > Full text
Journal home
Advance online publication
Current issue
Archive
Press releases
Focuses
Guide to authors
Online submissionOnline submission
For referees
Free online issue
Contact the journal
Subscribe
Advertising
work@npg
Reprints and permissions
About this site
For librarians
 
NPG Resources
Nature
Nature Reviews Immunology
Nature Medicine
Nature Cell Biology
NI Tutorial: Finding regulatory DNA regions
Signaling Gateway
Immunology & Cell Biology
Mucosal Immunology
Nature Conferences
Nature Stem Cells
NPG Subject areas
Biotechnology
Cancer
Chemistry
Clinical Medicine
Dentistry
Development
Drug Discovery
Earth Sciences
Evolution & Ecology
Genetics
Immunology
Materials Science
Medical Research
Microbiology
Molecular Cell Biology
Neuroscience
Pharmacology
Physics
Browse all publications
Article
Nature Immunology  5, 524 - 530 (2004)
Published online: 28 March 2004; | doi:10.1038/ni1058

There is a Corrigendum (June 2004) associated with this document.

T cell killing does not require the formation of a stable mature immunological synapse

Marco A Purbhoo1, Darrell J Irvine1, 2, Johannes B Huppa1 & Mark M Davis1

1 The Department of Microbiology and Immunology and the Howard Hughes Medical Institute, Stanford University School of Medicine, Stanford, California 94305, USA.

2 Present address: Department of Material Science & Engineering/Biological Engineering Division, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA.

Correspondence should be addressed to Mark M Davis mdavis@cmgm.stanford.edu
A notable feature of T lymphocyte recognition on other cell surfaces is the formation of a stable mature immunological synapse. Here we use a single-molecule labeling method to directly measure the number of ligands a cytotoxic T cell engages and track the consequences of that interaction by three-dimensional video microscopy. Like helper T cells, cytotoxic T cells were able to detect even a single foreign antigen but required about ten complexes of peptide−major histocompatibility complex (pMHC) to achieve full calcium increase and to form a mature synapse. Thus, cytotoxic T cells and helper T cells are more uniform in their antigen sensitivities than previously thought. Furthermore, only three pMHC complexes were required for killing, showing that stable synapse formation and complete signaling are not required for cytotoxicity.
Antigen recognition by T lymphocytes is characterized by the interaction of the T cell receptor (TCR) and its CD4 or CD8 coreceptor with major histocompatibility complex (MHC) molecules presenting peptide antigen on the surface of antigen-presenting cells (APCs). After antigenic stimulation, many adhesion, costimulatory, cytoskeletal and signal transduction components rearrange to the T cell−APC interface in a concerted effort to effect and direct T cell activation. This region of dynamic, supramolecular organization is commonly referred to as the immunological synapse1, 2, 3.

Although T cell responses to antigen are dose dependent, it has been difficult to determine the minimum number of peptide-MHC (pMHC) complexes required to induce T cell activation and thus to show the ultimate sensitivity of the immune response. Previous estimates of the minimum number of pMHC complex ligands required to stimulate T lymphocytes have ranged from 60 to 400 for CD4+ T cells and from 1 to 400 for cytotoxic T lymphocytes (CTLs)4, 5, 6, 7, 8, 9, 10. These results represent the average number of pMHC complexes per APC calculated to exist when they are pulsed with the minimum exogenous antigen concentration required for T cell stimulation. Although this approach allows confinement of the minimum pMHC complex requirements for T cell activation, it precludes determination of a more precise value, as it relies on the average, not actual, number of pMHC complexes per APC and does not consider that the T cell contacts only a fraction of the APC surface. Studies of T cell interactions with antigen-pulsed B cells have shown that the T cell−APC interface may comprise only 10% of the total APC surface and that peptide loading of individual APCs may vary by half a log from the average loading at any given concentration. Thus, the actual number of pMHC complex ligands 'seen' by a T cell may vary up to 50-fold from the average pMHC complex number calculated per APC. These limitations have been overcome by direct imaging on live cells of individual fluorescence-labeled pMHC class II complexes at the T cell−APC interface11. Correlation of the number of pMHC complexes at the interface with the extent of T cell stimulation (with intracellular calcium flux as a marker of early T cell activation) showed that full calcium mobilization and immunological synapse formation (as measured by the redistribution of the adhesion molecule ICAM-1) by CD4+ T cells requires engagement with as few as ten pMHC ligands, with even a single pMHC ligand capable of inducing a measurable biological response in the form of a submaximal intracellular calcium flux11. Furthermore, blocking of CD4 prevented calcium flux unless 25 or more specific pMHC complexes were present at the interface, thus demonstrating that CD4 acts to increase T cell sensitivity at low, but not high, ligand density11.

Here we adapted this procedure to peptides presented by MHC class I molecules and investigated antigen sensitivity in two H-2Kb-restricted systems: a cytotoxic T cell clone derived from mice transgenic for the 2C TCR, and T cell blast populations expanded from mice transgenic for the OT-1 TCR. By determining the relationship between the number of pMHC molecules at the interface and T cell activation and immunological synapse formation, we were able to delineate the minimum number of antigenic pMHC complexes necessary for the induction of their effector function, cytotoxicity. We further investigated the CD8 dependency of these processes and characterized the unique dynamics of this molecule at the immunological synapse.

Results
CTLs recognize wild-type and extended antigens equally
The N terminus of an antigenic, MHC class II−binding peptide was previously biotinylated and extended to achieve quantitative labeling with phycoerythrin-conjugated streptavidin (SAv-PE)11 when such peptides are bound to MHC molecules on cell surfaces. We adapted this procedure to MHC class I molecules using C-terminal antigen extensions, which we found bound MHC class I more efficiently than peptides extended at the N terminus. We compared the relative affinity of wild-type and extended peptides for MHC class I as well as their recognition by CTLs in cytotoxicity assays. The 2C CTLs recognized the wild-type and extended antigen to an identical extent (Fig. 1a), with both peptide forms showing a similar affinity for the MHC. In contrast, the OT-1 CTL response to the extended antigen was less potent than it was for wild-type peptide (Fig. 1b). The MHC class I stabilization assay showed that unlike 2C antigen (Fig. 1c), the extended OT-1 antigen bound less well to MHC than the wild-type peptide did (Fig. 1d), with the shift between wild-type and extended antigen in the MHC stabilization curve equal in extent to the shift in the dose-response curve of the cytotoxicity assay. Thus, taking into account the reduced MHC loading, OT-1 CTLs recognized both pMHC complexes with equal efficacy. For both 2C and OT-1, the extended pMHC class I complexes could be quantitatively labeled with SAv-PE (Supplementary Fig. 1 online), as demonstrated for moth cytochrome c peptide (amino acids 88−103) in a MHC class II−restricted system11.

Figure 1. Both 2C and OT-1 CTLs recognize wild-type and extended (and biotinylated) antigen-MHC complexes with equal efficacy.
Figure 1 thumbnail

(a,b) Cytotoxicity assay to determine 2C (a) and OT-1 (b) sensitivity to wild-type () and extended (^) forms of antigen. CTLs were incubated overnight with H-2Kb-transfected CH27 B cells pulsed with various concentrations of wild-type or extended antigen. Specific lysis was determined by flow cytometry. Specific lysis is plotted against peptide concentration. (c,d) RMA-S MHC-stabilization assay for the peptides recognized by 2C (c) and OT-1 (d) CTLs to determine relative binding of wild-type () and extended (^) antigen to H-2Kb. H-2Kb molecules on RMA-S cells were stabilized by incubation of cells with various concentrations of wild-type or extended antigen. The extent of H-2Kb stabilization was determined by flow cytometry. Relative H-2Kb stabilization (as the shift in mean fluorescence intensity over background) is plotted against peptide concentration. Assays were done in triplicate.



Full FigureFull Figure and legend (31K)
Antigen sensitivity of early CTL activation
To assess their peptide sensitivity, we 'loaded' CTLs with the intracellular calcium indicator dye FURA-2, mixed them with peptide-labeled APCs (H-2Kb-transfected CH27 B cells) and analyzed them by three-dimensional fluorescence video microscopy. We monitored intracellular calcium flux by comparing the differentially calcium-dependent FURA fluorescence in response to illumination at 340 nm and 380 nm. We quantified the dose response of calcium signaling to specific numbers of pMHC complexes by integrating the FURA 340:380 ratio data over 4 min after the initial calcium increase. For both 2C and OT-1, a single pMHC complex induced a detectable calcium signal, and additional ligands resulted in greater calcium increase, with maximum responses first noted with approximately ten pMHC complexes (Fig. 2a,b, filled circles), exactly as seen before with two helper T cell lines11. Although rapid photobleaching of phycoerythrin restricted determination of pMHC complex density at the immunological synapse to a single time point, the calcium flux data were consistent whether the number of peptides was determined within a few minutes after the initial calcium flux or at later time points (Supplementary Fig. 2 online). As noted before11, this result suggests that the number of pMHC complexes at the immunological synapse remains constant after initial CTL activation and that it is unlikely that pMHC complexes continue to be recruited to the immunological synapse.

Figure 2. Dose response of T cell calcium signals to antigenic pMHC complexes in the T cell−APC interface.
Figure 2 thumbnail

(a,b) Response curves for the 2C CTL clone (a) and OT-1 blasts (b). The coreceptor function for CTLs in early T cell activation was assessed by comparison of responses to APCs expressing H-2Kb or H-2Kb(D227K) (a,b). For quantification of early calcium signal, FURA-2AM ratios were measured every 20 s and were integrated for 4 min from the time of initial calcium increase. Integrated FURA ratios are shown versus the number of pMHC complexes detected at the CTL-APC interface. FURA values are normalized to the plateau value of the curve representing wild-type interaction in each graph. (c,d) The relationship of pMHC complexes at the CTL-APC interface to the stability of ICAM-1 redistribution in 2C (c) and OT-1 (d) CTLs. After imaging of pMHC complexes at the CTL-APC interface, the distribution of GFP-labeled ICAM-1 at the interface was determined by fluorescence microscopy. ICAM-1−GFP was imaged at 1-minute intervals for a further 5 min to assess the stability of ICAM-1 redistribution. (e) Dose response curves for 5C.C7 T cells in CD4-blocked or unblocked conditions. Measurements were made as described in a and b above except that FURA ratios were integrated over 7 min.



Full FigureFull Figure and legend (38K)
Formation of the immunological synapse is characterized by a complex pattern of receptor redistribution at the T cell−APC interface12, 13, a hallmark of which is the redistribution of the adhesion molecule ICAM-1 to the interface periphery. We investigated the antigen sensitivity of ICAM-1 redistribution, and thus immunological synapse formation, by determining the distribution at the CTL-APC interface of green fluorescent protein (GFP)−labeled ICAM-1 for at least 5 min after the imaging of pMHC complexes. We classified the distribution over time of ICAM-1 into three categories: 'stable rings', in which peripheral distribution was maintained over at least 5 min; 'transient rings', in which peripheral distribution degraded into clusters; and 'no rings', showing only random clustering or no ICAM-1 accumulation. For both 2C (Fig. 2c) and OT-1 (Fig. 2d) CTLs, although a single pMHC complex could induce transient redistribution of ICAM-1, eight to ten pMHC complexes were required for stable immunological synapse formation.

As reported before for helper T cells11, a single pMHC complex induced a small calcium flux and resulted in a longer residence time. For OT-1 CTLs, only 15% (n = 14) of CTL-APC conjugates lacking antigenic pMHC complexes were stable over 30 min. The percentage of stable CTL-APC conjugates increased to 60% (n = 10) when a single pMHC complex was engaged, increased further to 75% (n = 8) with two pMHC complexes and reached 100% (n = 60) with three or more pMHC complexes at the immunological synapse. Thus, even a single pMHC complex substantially extends conjugate stability, presumably by increasing the affinity of the adhesion molecule LFA-1 for ICAM-1 (refs. 14, 15, 16).

To investigate the contribution of the CD8 molecule to peptide sensitivity, we transfected APCs with H-2Kb(D227K), a mutant unable to interact with CD8 (refs. 17,18). The lack of CD8 engagement with pMHC complexes greatly reduced the response of 2C and OT-1 CTLs to any number of pMHC complexes, with the calcium flux on average equal to the amount induced by one to two pMHC complexes with wild-type MHC (Fig. 2a,b, open circles).

The pMHC requirements of CTLs for prolonged APC engagement, calcium signal and immunological synapse formation were essentially identical to data described for 2B4 (ref. 11) and 5C.C7 CD4+ T cells (Fig. 2e), indicating a similar relationship between pMHC complex number and these early activation markers among T lymphocytes. In contrast, the dependence of CTL sensitivity on CD8 differs to CD4+ T cells in which CD4 acts to increase T cell sensitivity at low, but not high, pMHC complex densities. This indicates that there are considerable differences between CTLs and helper T cells in the requirement for coreceptor to initiate TCR signal transduction.

Antigen sensitivity of the CTL cytotoxic response
We next determined the minimum number of antigens required to induce CTL effector function (cytotoxicity), which is difficult to measure for helper T cells because their effector functions (cytokine secretion and proliferation) take many hours to manifest. After imaging pMHC complexes at the interface, we observed CTL-APC conjugates for up to 230 min and determined APC death visually by monitoring for extensive APC membrane blebbing (characteristic of apoptosis) or severe cellular deformation and rupture (Supplementary Fig. 3 online). Although completion of cell death may take hours, the earliest sign of cell death (extensive membrane blebbing) was usually apparent shortly (5−15 min) after CTL-APC engagement, indicating that the 'decision' to proceed with the cytotoxic response is dependent on the events during the early stage of CTL-APC contact. Engagement with zero, one or (in most cases) two pMHC complexes was insufficient to induce target cell death, even though conjugates were often stable for the entire experiment (data not shown). In contrast, 2C and OT-1 CTLs readily induced target cell death when three or more antigens were engaged (Table 1 and Supplementary Table 1 online). Thus three pMHC complexes are sufficient to induce cytotoxicity.

Table 1. Target cell death in response to pMHC complexes at the T cell−APC interface for OT-1 and 2C CTLs
Table 1 thumbnail

Full TableFull Table
Our single-time-point images of pMHC complex distribution at the immunological synapse in the experiment described above did not indicate that agonist peptides require colocalization to effect cytotoxicity, ruling out models of activation that require dimerization or trimerization of agonist pMHC complex ligands (Fig. 3) and favoring those in which monomeric interactions or heterologous multimers formed between antigenic and endogenous pMHC complexes are the crucial trigger11, 19, 20. Although we relied on single 'snapshots' of pMHC complex distribution, comparison of different cells obtained at diverse time points also did not show a tendency of pMHC complexes to aggregate as the immunological synapse matures (data not shown).

Figure 3. Correlation of cytotoxicity with spatial distribution of pMHC complexes at the immunological synapse.
Figure 3 thumbnail

The distribution of SAv-PE−labeled pMHC complexes at the immunological synapse was imaged for all conjugates (2C and OT-1) described in Table 1 containing one to three pMHC complexes at the interface. The distribution of individual pMHC complex was classified as either distinct from others (resolved) or colocal (unresolved). Thus, the configuration of three pMHC complexes could appear as three separate pMHC complexes (1 + 1 + 1), one single and two colocal pMHC complexes (1 + 2) or three colocal pMHC complexes (3). The configuration of pMHC complexes at the immunological synapse was correlated with progress of the APC to cell death. Color bar (right) indicates brightness levels.



Full FigureFull Figure and legend (19K)
Dynamics of the CD8 coreceptor at the immunological synapse
Unlike the case with some CD4+ T cells, CTL activation and immunological synapse formation21 depend on coreceptor engagement with pMHC complex regardless of antigen density. To investigate this effect further, we used APCs (CH27 B cells) expressing either H-2Kb or H-2Kb(D227K) to ascertain by video microscopy at which stage of antigen recognition coreceptor engagement is essential. We assigned scores to cells for the number of CTL-APC contacts, number of stable conjugates and number of cells progressing toward cell death. CD8 engagement was most essential early during antigen recognition, for the formation and maintenance of tight CTL-APC conjugates (Table 2). For the few stable conjugates that formed without the CD8-pMHC complex interaction, progression to target cell death was still around 50% within 30 min (compared with 87% for wild-type MHC), indicating that CD8-pMHC complex interaction is less essential for induction of cytotoxicity than for the initial recognition of antigen.

Table 2. Dependence of early CTL-APC contact on the CD8-pMHC complex interaction
Table 2 thumbnail

Full TableFull Table
To determine whether CD8 acts mainly during the early events of activation, like CD4 (ref. 22), we investigated the dynamics of fluorescence-tagged CD8, MHC and the TCR-CD3 complex at the immunological synapse of 2C CTLs. CD8 accumulated rapidly at the immunological synapse, with peak intensity 1 min after CTL-APC contact (Fig. 4a,b). CD8 clustering preceded and did not extensively coincide with clustering of its MHC ligand or the TCR. Instead, CD8 was rapidly internalized from the center of the immunological synapse, with only remnant surface accumulation persisting in association, but not colocalization, with MHC clusters. In contrast, MHC and TCR clustered with similar kinetics and demonstrate good spatial overlap for around 15 min, after which clusters frequently dissociated (Fig. 4c). Unlike the dissimilar profiles of antigen receptors (MHC and TCR) and coreceptor at the immunological synapse, CD8 accumulation seemed concurrent with onset of key signal transduction events, such as activation of the phosphatidyl inositol secondary messenger pathway (as measured by the redistribution of the protein kinase B−pleckstrin homology domain (Akt-PH domain)−yellow fluorescent protein (YFP) construct to the immunological synapse after formation of its phosphoinositol-3,4,5-trisphosphate ligand), intracellular calcium flux and accumulation of the LAT adaptor protein at the immunological synapse (Fig. 4d). These data suggest that CD8 accumulation acts mainly to initiate signal transduction and is distinct from the (as-yet-unknown) purpose of MHC and TCR clustering as the synapse matures.

Figure 4. Distribution of antigen receptors and signal transduction components at the MHC class I−restricted immunological synapse.
Figure 4 thumbnail

The interaction between 2C transfectants expressing fluorescence-tagged CD8beta, CD3zeta, the linker for activation of T cells (LAT) or the Akt-PH domain with agonist-pulsed APCs (CH27 cells) transfected with fluorescence-tagged H-2Kb was monitored by three-dimensional video microscopy. Minutes after initial CTL-APC contact are shown at the top of the DIC images; color bars (right margin of images) indicate 'fold' over background intensity. (a) Temporal and spatial resolution of CD8beta-YFP and H-2Kb−CFP within a single immunological synapse formed between a 2C CTL and an antigen-pulsed APC. Top row, DIC images of the CTL-APC interaction; middle rows, en face view of the distribution of the indicated fluorescence-labeled receptor species at the immunological synapse; bottom row, overlay the images for both fluorescent species imaged at the immunological synapse. Arrowheads indicate nonoverlapping clusters. (b) Intensity of CD8beta-YFP accumulation at the immunological synapse compared with that of antigen receptors (MHC and TCR-CD3zeta) over 26 min. (c) Temporal and spatial resolution of CD3zeta-CFP and H-2Kb−YFP within a single immunological synapse formed between a 2C CTL and an antigen-pulsed APC. Images organized as in (a). (d) Intensity of CD8beta-YFP accumulation at the immunological synapse compared with that of signal transduction molecules (LAT-YFP; Akt-PH domain−YFP) over 26 min. Vertical axes (b,d) represent average intensity of clusters over background. For CD8beta, CD3zeta, H-2Kb and Lat fluorescent constructs, the membrane intensity in clusters at the immunological synapse is compared with average intensity outside the immunological synapse. For the Akt-PH domain fluorescent construct, the intensity at the plasma membrane is compared with the average intensity within the cytoplasm. Calcium flux is presented as the ratio of FURA emission when excited at 340 nm or 380 nm.



Full FigureFull Figure and legend (126K)
As MHC clustering on the APC is CD8 dependent and is actively directed by the T cell19, 23, we investigated whether CD8 directly sequesters MHC. The 2C CTLs engaging peptide-pulsed APCs (CH27 B cells) expressing only H-2Kb(D227K) were not activated and did not induce clustering of MHC class I at the immunological synapse (Fig. 5a). However, when 2C CTLs engaged antigen-pulsed APCs (CH27 B cells) coexpressing both wild-type H-2Kb−cyan fluorescent protein (CFP), and mutant H-2Kb(D227K)−YFP, there was equal and colocal clustering of wild-type and mutant MHC (Fig. 5b). As H-2Kb(D227K) does not interact with CD8, MHC clustering by activated CTLs may be driven solely by the TCR. Results for targets expressing H-2Kb constructs with reversed fluorescent tagging or for OT-1 CTLs were identical (data not shown).

Figure 5. Antigen-induced clustering of H-2Kb(D227K) in presence of wild-type H-2Kb.
Figure 5 thumbnail

Reconstructed interface of 2C CTLs engaging antigen-pulsed targets expressing YFP-tagged H-2Kb or H-2Kb(D227K) for single-transfectant CH27 B cells (a) or coexpressing YFP-tagged H-2Kb (red in merge) and CFP-tagged H-2Kb(D227K) (green in merge) for cotransfectant CH27 B cells (b). Color bar (bottom) indicates 'fold' over background intensity. WT, wild-type.



Full FigureFull Figure and legend (32K)
 Top
Discussion
Most previous estimates of the minimum number of ligands required to stimulate T lymphocytes have been based on the calculated average number of pMHC complex presented per APC4, 5, 6, 7, 8, 9, 10. Here we directly visualized the minimum antigenic requirements at the T cell−APC interface for various stages during CTL activation. We found that a single pMHC complex was sufficient to arrest T cell procession, prolong scanning of the APC, stimulate a suboptimal calcium flux and induce transient immunological synapse formation. Stable immunological synapse formation and complete calcium flux, however, required engagement with about ten pMHC complexes. We extended this analysis of T cell activation to include the antigenic requirements for induction of CTL effector function. Although CTL engagement with a single pMHC complex promoted conjugate formation, it was insufficient to induce cytotoxicity in either of the two model systems surveyed here4. Instead, a minimum of three agonist pMHC complexes were required at the T cell−APC interface to induce target cell death. The nearly identical results with either CTL type (2C and OT-1) suggest that this is representative of mature CTLs in general, at least for those dependent on CD8. This extreme peptide sensitivity of CTLs may be essential, as it is likely that not very many antigenic peptides are presented per APC, given that the efficiency of MHC class I peptide generation from degraded molecules is estimated to be as low as 1 in 10, 000 (refs. 24,25). It was also unexpected that although the affinities of the TCR-pMHC complex interactions of the four MHC class I− and MHC class II−restricted systems investigated (2C, OT-1, 5C.C7 (ref. 11) and 2B4 (ref. 11)) vary by ninefold (from 6 muM for OT-1 (ref. 26) to 54 muM for 2C27), the T cell sensitivity to antigen was almost identical. This suggests that thymic and peripheral selective forces, presumably driven by interactions with specific self antigens28, 29 may act to 'tune' T cells to have very uniform sensitivities.

At the lowest antigen densities sufficient to induce cytotoxicity, we noted formation of only a transient immunological synapse at the T cell−APC interface and induction of a calcium flux that was well short of the maximum, indicating that cytotoxicity requires neither a stable immunological synapse nor a maximum calcium response. However, we did find formation of a stable immunological synapse at higher antigen densities. For OT-1 CTLs, target cell death in response to less than 30 pMHC complexes was always accompanied by intense blebbing of the APC membrane, whereas cytotoxicity induced by higher antigen density in most cases (14 of 17) was not associated with membrane blebbing. Membrane blebbing is a general marker of apoptosis (including Fas mediated), and blocking of Fas-mediated killing with a Fas ligand monoclonal antibody reduced the occurrence of membrane blebbing in OT-1 CTL-APC conjugates containing 3−30 pMHC complexes at the interface from 100% to 15% (n = 14). It has been suggested that Fas ligand−mediated cytotoxicity is more antigen sensitive than perforin-mediated cytolysis30; thus, Fas ligand−mediated cytotoxicity may predominate at the lowest antigen density and may be independent of immunological synapse stability. In contrast, a stable immunological synapse at higher antigen densities may allow continued lytic granule secretion and thus increase the contribution to cell death of perforin-mediated cytolysis. Furthermore, cytotoxicity has been determined to be the most antigen sensitive of the CTL effector responses31, suggesting that a stable mature synapse is required to direct less sensitive responses such as cytokine production and proliferation.

Early CTL activation was dependent on the interaction of CD8 with pMHC complexes and, unlike the situation with CD4+ T cells, this could not be compensated for by higher antigen densities. In the absence of the CD8-MHC interaction, the calcium response in general was just below that induced by three pMHC complexes in wild-type conditions (Fig. 2a,b). CD8 thus supplies an essential additional signal that increases partially activated T cells to full function. CD4, in contrast, acts to enhance antigen sensitivity to low ligand densities, which otherwise would induce no detectable T cell activity11.

CD8 engagement with pMHC complexes was most essential for CTL-APC conjugate formation and, to a lesser extent, for subsequent induction of cytotoxicity. Thus, CD8 clusters only briefly after CTL-APC contact. This clustering is concurrent with activation of key signal transduction events, suggesting that coreceptor clusters mainly in its capacity to induce TCR signal transduction. However, as basal CD8 numbers are likely to exceed the low initial number of pMHC-TCR complexes requiring activation at the CTL-APC interface, bulk CD8 accumulation after antigen recognition not only may serve to activate specific pMHC-TCR-CD3 complexes but also could conceivably lead to the nonspecific activation of TCR-CD3 complexes, thus amplifying the signal from the few antigenic pMHC-TCR complexes. Alternatively, an early rise in CD8 concentration may simply increase the probability that the few initial antigenic pMHC-TCR-CD3 complexes formed before synapse reorganization are activated. In contrast, by the time the mature synapse forms and TCR-MHC clustering is complete, signal transduction events have mostly ebbed. In the absence of pronounced signaling, MHC-TCR clustering is unlikely to represent continued antigen recognition, as is the case for CD4+ T cells32. Prolonged antigen recognition seems superfluous, because CTLs rapidly destroy their targets. After clustering, pMHC complexes may be transferred to the CTL surface to serve as target for 'fratricide', thus aiding in CTL population control33, 34. TCRs in turn may cluster at the immunological synapse as a 'beacon' directing continued effector protein release, perhaps acting as a focal point for cytoskeletal rearrangements.

In summary, we have visualized the antigenic requirements for MHC class I−restricted T cell activation and determined the minimum number of antigens required for induction of CTL effector functions, showing that at least one form of cytotoxicity does not require the formation of a stable immunological synapse. We have further shown that, unlike the case with CD4+ T cells, recruitment of coreceptor and other key TCR signal transduction elements is restricted to the early immunological synapse, suggesting that only antigen recognized immediately after CTL-APC contact contributes to CTL activation and effector response. In contrast, we found TCR accumulation only as the immunological synapse matured, indicating a function for antigen receptor clustering during the effector phase of CTL activation.

 Top
Methods
Cells and constructs.
The 5C.C7 or OT-1 primary cells were isolated from lymph nodes of transgenic mice and were stimulated with either 1 muM moth cytochrome c peptide (amino acids 88−103; for 5C.C7) or irradiated (3,000 rad) APCs pulsed with 1 muM peptide of amino acid sequence SIINFEKL (for OT-1). Cells were maintained in R-10 medium (RPMI-1640 medium with 10% FCS, 2 mM L-glutamine, 50 muM beta-mercaptoethanol and penicillin plus streptomycin). The 2C CTLs were stimulated weekly with irradiated (3,000 rad) P815 mastocytoma cells and were maintained for no longer than 4 months after which new cells were grown from frozen stock of the original clone. All T cells were grown in the presence of 15 U/ml of recombinant interleukin 2 (Sigma) and were used on day 5 after stimulation or later.

Genes of interest were amplified from thymocyte- or 2C CTL−derived cDNA22, 32 and were fluorescence tagged by ligation into the plasmid EGFP-N1 or spectral variants thereof (Clontech). The H-2Kb(D227K) mutant was generated by site-directed mutagenesis. Untagged and fluorescence-tagged (GFP, YFP and CFP) constructs of H-2Kb and H-2Kb(D227K) were generated. Constructs were subcloned into mouse Moloney murine leukemia virus−based or murine stem cell virus−based retroviral vectors. CTLs and APCs were stably transfected with the Phoenix retroviral transfection system35, and transfectants were selected in the presence of 20 mug/ml of blasticidin (Invitrogen) or 100 mug/ml of zeocin (Invitrogen), as appropriate. CH27 B cells transfected with GFP-labeled ICAM-1 have been described36.

B cell lymphoma cells (CH27) were used as APCs for 5C.C7 T cells or were transfected with the various H-2Kb constructs described above for use as APCs to CTLs. Different H-2Kb transfectants (wild-type, mutant or fluorescence-tagged H-2Kb) were clonally expanded for selection of clones with similar H-2Kb expression as determined by flow cytometry (Cytomics FC500; Beckman Coulter) with phycoerythrin-linked H-2Kb monoclonal antibody AF6-88.5 (PharMingen).

Peptides and reagents.
We prepared 2C and OT-1 antigenic peptides (with sequences SIYRYYGL and SIINFEKL, respectively) by solid-phase synthesis with a biotinylated C terminus connected to the MHC binding sequence by a flexible linker of sequence GGGSGGGSGGGSGGGSK-biotin. The linker contained D-stereoisomers of serine to avoid possible proteolytic cleavage of residues extending outside the MHC binding groove. Fab fragments of CD4 monoclonal antibody GK1.5 were prepared as instructed with an ImmunoPure Fab generation kit (Pierce). SAv-PE (Pharmingen) was confirmed to be monomeric, with a 1:1 ratio of SAv to phycoerythrin11.

Peptide labeling.
APCs were pulsed with biotinylated peptide (10−1,000 nM) for 1 h at 37 °C in 5% CO2 in R-10 medium at a density of 4 times 106 cells/ml, were washed twice with 10 ml PBS and were resuspended in 0.5% BSA (Promega) in PBS. Fc receptors were blocked by incubation of cells on ice for 10 min with 10 mug/ml of Fc block (Pharmingen). SAv-PE was added at a concentration of 5 mug/ml for 20 min at 21 °C. Cells were washed twice with 10 ml PBS, resuspended in phenol red−deficient RPMI-1640 medium and placed on ice until used. We quantified SAv-PE labeling of peptide by flow cytometry and calibration with Quantibrite phycoerythrin calibration beads (Pharmingen).

RMA-S MHC stabilization assay.
RMA-S cells were grown overnight at 21 °C to allow cell surface expression of 'empty' MHC class I molecules (not in complex with peptide). Cells were incubated with peptide in 96-well V-bottomed plates (Nunc) at a density of 4 times 105 cells/well. After cells were incubated for 1 h at 21 °C (to allow peptide stabilization of MHC), they were moved to 37 °C for 2 h to allow degradation of MHC not in complexes. Cells were washed twice with 200 mul PBS, stained on ice with 50 mug/ml of phycoerythrin-linked H-2Kb monoclonal antibodies AF6-88.5 or CTKb for 1 h and rewashed twice. Staining was analyzed by flow cytometry.

Cytotoxicity assay.
APCs were washed twice with 10 ml PBS and stained with 1 muM CFSE (carboxyfluorescein diacetate succinimidyl diester; Molecular Probes) for 10 min at 37 °C at a density of 5 times 107 cells/ml. Cells were washed twice with imaging media and pulsed with peptide for 1 h at 21 °C in 96-well U-bottomed plates at a density of 5 times 104 cells/well. Cells were washed twice and were resuspended in 200 mul imaging media, and CTLs were added at a density of 1 times 105 to 1 times 106 cells/well. Assays were left at 37 °C for 16 h and were terminated by being moved to 4 °C and by the addition of EDTA to a concentration of 2.5 mM. Cell death was determined by death-associated changes in the forward- and side-scatter properties among the CFSE-positive APC population. Specific lysis was determined as [1 - (% live cells(sample, with peptide)/ % live cells(sample, no peptide))] times 100.

Microscopy and image analysis.
A Zeiss/Universal Imaging three-dimensional time-lapse system with a 40times objective was used for microscopy and data analysis as described11, 22, 32. Eight-well chambered coverslips (Nunc) with minimal imaging medium (R-10 lacking phenol red) were used for imaging. Samples were provided with humidified air at 37 °C and 5% CO2 during imaging. APCs were pulsed with wild-type antigen at a concentration of 1 muM unless indicated otherwise. Data were analyzed as described11.

For early T cell activation and synapse formation experiments, before imaging, T cells were loaded with 10 muM of the calcium indicator dye FURA-2AM (Molecular Probes) in R-10 medium for 25 min at 21 °C. FURA-loaded T cells and APCs were mixed together and images were obtained every 20 s for FURA-2AM emission and differential interference contrast (DIC). In addition to single plane DIC-FURA imaging, the distribution of fluorescence-tagged proteins was tracked by z-stack acquisition (21 individual planes 1 mum apart) at appropriate wavelengths at 1-minute intervals. However, rapid photobleaching of phycoerythrin limited three-dimensional observation of this fluorophore (and thus labeled pMHC complexes) to one time point. Furthermore, in experiments tracking multiple fluorescence-tagged proteins, overlap between the excitation spectra of phycoerythrin and GFP or YFP restricted imaging of GFP- and YFP-labeled proteins to after the phycoerythrin 'snapshot' (to avoid photobleaching). In CD8-MHC or CD3zeta-MHC colocalization experiments, CFP and YFP z-stacks were acquired interleaved for simultaneous imaging of these fluorophores. For assessment of the effects of CD4 receptor blockage, blocking Fab was added to the appropriate T cells at a concentration of 10 mug/ml for 20 min at 4 °C before imaging.

For cytotoxicity experiments, DIC images of T cells and labeled APCs were obtained every 20 s for up to 230 min. To avoid phototoxicity associated with (short wavelength) fluorescent excitation light, FURA-2AM was not imaged in these experiments and fluorescence acquisition was mostly restricted to a single phycoerythrin z-stack for determination of pMHC complex distribution soon after conjugate formation (in a few experiments the distribution of GFP- or YFP-tagged proteins was determined for up to 10 min after acquisition of the single phycoerythrin stack). After completion of experimental time course, APCs were stained with annexin V and propidium iodide to visualize specific stages of apoptosis. Wells were supplemented with annexin V−fluorescein isothiocyanate (Pharmingen) and 2.5 mM CaCl2 and were imaged and then were supplemented with 50 mug/ml propidium iodide and imaged again.

Note: Supplementary information is available on the Nature Immunology website.

 Top
Received 10 November 2003; Accepted 28 January 2004; Published online: 28 March 2004.

REFERENCES
  1. Bromley, S.K. et al. The immunological synapse. Annu. Rev. Immunol. 19, 375–396 (2001). | Article | PubMed | ISI | ChemPort |
  2. Delon, J. & Germain, R.N. Information transfer at the immunological synapse. Curr. Biol. 10, R923–933 (2000). | Article | PubMed | ISI | ChemPort |
  3. Davis, M.M. et al. Dynamics of cell surface molecules during T cell recognition. Annu. Rev. Biochem. 72, 717–742 (2003). | Article | PubMed | ISI | ChemPort |
  4. Sykulev, Y., Joo, M., Vturina, I., Tsomides, T.J. & Eisen, H.N. Evidence that a single peptide-MHC complex on a target cell can elicit a cytolytic T cell response. Immunity 4, 565–571 (1996). | PubMed | ISI | ChemPort |
  5. Kimachi, K., Croft, M. & Grey, H.M. The minimal number of antigen-major histocompatibility complex class II complexes required for activation of naive and primed T cells. Eur. J. Immunol. 27, 3310–3317 (1997). | PubMed | ISI | ChemPort |
  6. Demotz, S., Grey, H.M. & Sette, A. The minimal number of class II MHC-antigen complexes needed for T cell activation. Science 249, 1028–1030 (1990). | PubMed | ISI | ChemPort |
  7. Harding, C.V. & Unanue, E.R. Quantitation of antigen-presenting cell MHC class II/peptide complexes necessary for T-cell stimulation. Nature 346, 574–576 (1990). | Article | PubMed | ISI | ChemPort |
  8. Christinck, E.R., Luscher, M.A., Barber, B.H. & Williams, D.B. Peptide binding to class I MHC on living cells and quantitation of complexes required for CTL lysis. Nature 352, 67–70 (1991). | Article | PubMed | ISI | ChemPort |
  9. Reay, P.A. et al. Determination of the relationship between T cell responsiveness and the number of MHC-peptide complexes using specific monoclonal antibodies. J. Immunol. 164, 5626–5634 (2000). | PubMed | ISI | ChemPort |
  10. Brower, R.C. et al. Minimal requirements for peptide mediated activation of CD8+ CTL. Mol. Immunol. 31, 1285–1293 (1994). | Article | PubMed | ISI | ChemPort |
  11. Irvine, D.J., Purbhoo, M.A., Krogsgaard, M. & Davis, M.M. Direct observation of ligand recognition by T cells. Nature 419, 845–849 (2002). | Article | PubMed | ISI | ChemPort |
  12. Monks, C.R., Freiberg, B.A., Kupfer, H., Sciaky, N. & Kupfer, A. Three-dimensional segregation of supramolecular activation clusters in T cells. Nature 395, 82–86 (1998). | Article | PubMed | ISI | ChemPort |
  13. Grakoui, A. et al. The immunological synapse: a molecular machine controlling T cell activation. Science 285, 221–227 (1999). | Article | PubMed | ISI | ChemPort |
  14. Dustin, M.L., Singer, K.H., Tuck, D.T. & Springer, T.A. Adhesion of T lymphoblasts to epidermal keratinocytes is regulated by interferon gamma and is mediated by intercellular adhesion molecule 1 (ICAM-1). J. Exp. Med. 167, 1323–1240 (1988). | PubMed | ISI | ChemPort |
  15. Shimaoka, M. et al. Reversibly locking a protein fold in an active conformation with a disulfide bond: integrin alphaL I domains with high affinity and antagonist activity in vivo. Proc. Natl. Acad. Sci. USA 98, 6009–6014 (2001). | Article | PubMed | ChemPort |
  16. Lu, C. et al. An isolated, surface-expressed I domain of the integrin alphaLbeta2 is sufficient for strong adhesive function when locked in the open conformation with a disulfide bond. Proc. Natl. Acad. Sci. USA 98, 2387–2392 (2001). | Article | PubMed | ChemPort |
  17. Potter, T.A., Rajan, T.V., Dick, R.F., 2nd & Bluestone, J.A. Substitution at residue 227 of H-2 class I molecules abrogates recognition by CD8-dependent, but not CD8-independent, cytotoxic T lymphocytes. Nature 337, 73–75 (1989). | Article | PubMed | ISI | ChemPort |
  18. Purbhoo, M.A. et al. The human CD8 coreceptor effects cytotoxic T cell activation and antigen sensitivity primarily by mediating complete phosphorylation of the T cell receptor zeta chain. J. Biol. Chem. 276, 32786–32792 (2001). | Article | PubMed | ISI | ChemPort |
  19. Wülfing, C. et al. Costimulation and endogenous MHC ligands contribute to T cell recognition. Nat. Immunol. 3, 42–47 (2002). | Article | PubMed | ISI | ChemPort |
  20. van der Merwe, P.A. Do T cell receptors do it alone? Nat. Immunol. 3, 1122–1123 (2002). | Article | PubMed | ISI | ChemPort |
  21. Potter, T.A., Grebe, K., Freiberg, B. & Kupfer, A. Formation of supramolecular activation clusters on fresh ex vivo CD8+ T cells after engagement of the T cell antigen receptor and CD8 by antigen-presenting cells. Proc. Natl. Acad. Sci. USA 98, 12624–12629 (2001). | Article | PubMed | ChemPort |
  22. Krummel, M.F., Sjaastad, M.D., Wülfing, C. & Davis, M.M. Differential clustering of CD4 and CD3zeta during T cell recognition. Science 289, 1349–1352 (2000). | Article | PubMed | ISI | ChemPort |
  23. Wülfing, C. & Davis, M.M. A receptor/cytoskeletal movement triggered by costimulation during T cell activation. Science 282, 2266–2269 (1998). | Article | PubMed | ISI | ChemPort |
  24. Princiotta, M.F. et al. Quantitating protein synthesis, degradation, and endogenous antigen processing. Immunity 18, 343–354 (2003). | Article | PubMed | ISI | ChemPort |
  25. Yewdell, J.W. Not such a dismal science: the economics of protein synthesis, folding, degradation and antigen processing. Trends Cell Biol. 11, 294–297 (2001). | Article | PubMed | ISI | ChemPort |
  26. Alam, S.M. et al. Qualitative and quantitative differences in T cell receptor binding of agonist and antagonist ligands. Immunity 10, 227–237 (1999). | PubMed | ISI | ChemPort |
  27. Degano, M. et al. A functional hot spot for antigen recognition in a superagonist TCR/MHC complex. Immunity 12, 251–261 (2000). | PubMed | ISI | ChemPort |
  28. Stefanova, I., Dorfman, J.R. & Germain, R.N. Self-recognition promotes the foreign antigen sensitivity of naive T lymphocytes. Nature 420, 429–434 (2002). | Article | PubMed | ISI | ChemPort |
  29. Ernst, B., Lee, D.S., Chang, J.M., Sprent, J. & Surh, C.D. The peptide ligands mediating positive selection in the thymus control T cell survival and homeostatic proliferation in the periphery. Immunity 11, 173–181 (1999). | PubMed | ISI | ChemPort |
  30. Kojima, H., Toda, M. & Sitkovsky, M.V. Comparison of Fas- versus perforin-mediated pathways of cytotoxicity in TCR- and Thy-1-activated murine T cells. Int. Immunol. 12, 365–374 (2000). | Article | PubMed | ISI | ChemPort |
  31. Valitutti, S., Muller, S., Dessing, M. & Lanzavecchia, A. Different responses are elicited in cytotoxic T lymphocytes by different levels of T cell receptor occupancy. J. Exp. Med. 183, 1917–1921 (1996). | PubMed | ISI | ChemPort |
  32. Huppa, J.B., Gleimer, M., Sumen, C. & Davis, M.M. Continuous T cell receptor signaling required for synapse maintenance and full effector potential. Nat. Immunol. 4, 749–755 (2003). | Article | PubMed | ISI |