The Size of the Synaptic Cleft and Distinct Distributions of Filamentous Actin, Ezrin, CD43, and CD45 at Activating and Inhibitory Human NK Cell Immune Synapses

Cells and Abs

NK cell transfectants, YTS/Mock (YTS, mock-transfected), and YTS/KIR2DL1 (YTS transfected to express KIR2DL1) were previously described (60). YTS is a subclone of the human NK tumor line YT (61). An EBV-transformed B cell line 721.221 (hereafter referred to as 221), selected to lack surface expression of endogenous class I MHC protein (62), was transfected to express enhanced green fluorescent protein (GFP)-tagged HLA-Cw3, -Cw4, or -Cw6 (i.e., 221/Cw6-GFP, etc.), as previously described (14). Cytotoxicity assays show that YTS/KIR2DL1 is effectively inhibited from killing 221 transfected with HLA-Cw6 (i.e., 221/Cw6) or 221/Cw4, but not 221/Cw3 or 221/Cw7, target cells (60). An activating NK cell IS is created when the intercellular contact between the NK and target cell results in lysis of the target cell. In contrast, an inhibitory NK cell IS is created at the intercellular contact of noncytolytic conjugates between the NK cell and target cell. Thus, conjugates between YTS/KIR2DL1 and 221/Cw6-GFP create an inhibitory NK cell IS and conjugates of YTS/Mock and 221/Cw6-GFP or YTS/KIR2DL1 and 221/Cw3-GFP create an activating NK cell IS.

Mouse anti-human ezrin mAb (clone 18, IgG1; BD Transduction Laboratories, Lexington, KY), rabbit polyclonal anti-human ezrin Ab (Upstate Biotechnology, Lake Placid, NY), mouse anti-human CD43 mAb (IG10, IgG1; BD PharMingen, San Diego, CA), and mouse anti-human CD45 (HI30, IgG1; BD PharMingen) were used at 5 μg/ml. Mouse anti-CD158a (EB6, IgG1; Serotec, Oxford, U.K.) was used at 10 μg/ml. Alexa Fluor 568 goat anti-mouse IgG, Alexa Fluor 633 goat anti-mouse IgG, and Alexa Fluor 568 goat anti-rabbit IgG (Molecular Probes, Eugene, OR) were used at a concentration of 2 μg/ml. To phenotype cells for expression of the various CD45 isoforms, the following FITC-conjugated mAbs (Caltag Laboratories, Burlingame, CA) were used: UCHLI (IgG2a) for CD45RO, MEM56 (IgG2b) for CD45RA, and MEM55 (IgG1) for CD45RB. For immunogold labeling, anti-GFP mAb (JL-8, IgG2a; Clontech Laboratories, Palo Alto, CA) was used at 20 μg/ml and detected with 12 μg/ml goat anti-mouse IgG conjugated with 10 nm gold (TAAB Laboratories, Berkshire, U.K.). All other reagents were purchased from Sigma-Aldrich (St. Louis, MO) unless stated otherwise.

Generation of NK clones

PBMCs were isolated from the buffy coat residues of healthy donors by centrifugation on Ficoll-Paque plus gradient (Amersham Pharmacia Biotech, Piscataway, NJ) according to the manufacturer’s instructions. Freshly isolated PBMCs and RPMI 8866 cells were irradiated (6,000 and 12,000 rad, respectively) and mixed in a ratio of 20:1 in DMEM supplemented with 10% human serum (type AB; Sigma-Aldrich), 30% nutrient mixture F-12 (Ham), 2 mM l-glutamine, 1× nonessential amino acids, 1 mM sodium pyruvate, 50 μM 2-ME, 50 U/ml penicillin-streptomycin, and 100 U/ml human recombinant IL-2 (Roche, Basel, Switzerland), referred to hereafter as DMEM plus supplements, (all from Invitrogen, Carlsbad, CA), with 1 μg/ml PHA (Sigma-Aldrich). The irradiated cells were plated into the wells of U-bottom 96-well plates (10⁵ PBMC and 5 × 10³ RPMI 8866 cells per well) and incubated overnight at 37°C/5% CO₂. CD3 PBL derived from human buffy coat residues that were cultured for 1 wk with irradiated 721.221 cells (20,000 rad), in a 4:3 ratio in RPMI 1640 medium containing with 10% human serum (type AB; Sigma-Aldrich) with supplements, were sorted by FACS into the 96-well plates containing the irradiated PBMCs and RPMI 8866 cells at 1 cell per well. After sorting, 100 μl of DMEM plus supplements was added to each well. Cells in wells that had undergone expansion were split every 4 to 5 days in a 1:1 ratio in DMEM plus supplements. NK clones were periodically monitored to be CD3⁻ and CD56⁺ by mAb staining and their phenotype was determined by mAb staining and cytotoxicity assays against a panel of target cells. Cytotoxicity assays and laser scanning confocal microscopy were performed at least 4 days after restimulation with fresh human recombinant IL-2.

Phenotyping of NK clones

Abs used for flow cytometry were CyChrome-conjugated anti-human CD56 (B159), CyChrome-conjugated anti-human CD3 (UCHT1), CyChrome-conjugated IgG1 control (MOPC-21), anti-human CD94 (HP-3D9), and PE-conjugated anti-KIR2DL2 (DX27), all purchased from BD PharMingen. Anti-KIR2DL1 (HP3E4) was purchased from the American Type Culture Collection (Manassas, VA) and anti-Ig-like transcript-2 (HP-F1) was a kind gift from M. López-Botet (Universitat Pompeu Fabra, Barcelona, Spain). All of these Abs were used at a concentration of 1 μg/ml. FITC-conjugated rabbit anti-mouse IgG (Sigma-Aldrich) and FITC-conjugated rat anti-mouse IgM (R6-60.2; BD PharMingen) were used at 2 μg/ml. The cytolytic activity of NK clones against various target cells was assessed in 5-h ³⁵S-release assays as described previously (63). Assays were performed in triplicate and data values differed by ∼5% of the mean. Spontaneous release of ³⁵S was <25% of the maximal release.

Imaging the IS

221 and YTS cell transfectants (10⁶ of each) were coincubated for 45 min at 37°C/5% CO₂ in conical end tubes containing 1 ml of culture medium comprising RPMI 1640 supplemented with 10% FBS, 2 mM l-glutamine, 1 mM sodium pyruvate, 1× nonessential amino acids, 50 μM 2-ME, 50 U/ml penicillin-streptomycin, (all purchased from Invitrogen), after which the cells had fallen under gravity. Excess medium was removed and the sedimented cells were fixed in Cytofix/Cytoperm (BD PharMingen) for 12 min at 4°C. The fixed cells were then washed twice in 0.1% Tween 20/PBS before cell staining. For NK clones, the same method was used except that 2 × 10⁵ cells of each type were used and the conjugates were briefly centrifuged before fixation to ensure that the cells had sedimented before decanting excess medium. Following cell fixation, f-actin was labeled using Alexa Fluor 633-conjugated phalloidin (Molecular Probes) at 5 U/ml in 1% BSA/PBS, for 1 h at 4°C. Cells were washed three times in 0.1% Tween 20/PBS before imaging.

For imaging the NK cell IS after different times of coincubation, 10⁶ of both YTS and 221 transfectants were incubated in a V-bottom 96-well culture plate at 37°C/5% CO₂ in 100 μl of prewarmed culture medium. The cells were then fixed after 2, 10, 20, or 45 min of coincubation by adding 200 μl of Cytofix/Cytoperm directly to the well at 4°C for 15 min.

For Ab staining, fixed cells were initially incubated in a blocking solution comprising 5% horse serum (Sigma-Aldrich) and 3% BSA in Perm/Wash buffer (BD PharMingen) for 1 h at 4°C. Next, cells were incubated with primary Ab diluted in blocking solution for 90 min at 4°C. Following three washes in 0.1% Tween 20/PBS, secondary Ab, also diluted in blocking solution, was added to the cells for 1 h at 4°C. The stained cells were then washed three times in 0.1% Tween 20/PBS and 7 μl of the resuspended cell pellet was placed between a microscope slide and a 24 × 24 mm coverslip (thickness no. 1) for microscopy. Staining procedures were altered to incorporate the use of buffers containing alternative detergents such as 0.05% saponin though no effect on staining was observed (data not shown). For double immunofluorescence labeling, both primary Abs followed by both secondary Abs were added. It was confirmed that there was no species cross-reactivity between Alexa Fluor-conjugated goat anti-mouse IgG and goat anti-rabbit IgG and also that none of the secondary Abs used showed any nonspecific binding to the cells in the absence of primary Ab (data not shown).

Confocal microscopy image analysis

Cell conjugates were imaged under a 63× oil immersion objective using a confocal laser scanning microscope (TCS SP2; Leica, Deerfield, IL) equipped with argon/krypton and helium/neon lasers using excitation wavelengths of 488, 568, and 633 nm. Conjugates were scanned in the xy-direction every 0.3 μm throughout the z-plane. The face of the IS was then reconstructed using a maximum intensity projection (Confocal Software; Leica). To quantify the relative amounts of HLA-C-GFP, ezrin, and CD43 at the inhibitory IS within a single optical slice, their mean fluorescence intensities were obtained from a 15-μm² area at the intercellular contact. The mean intensities were then measured from the same area marked onto the unconjugated membranes of both the NK and B cell and a percentage of mean fluorescence intensity at the inhibitory IS relative to the sum of the unconjugated membranes was calculated.

Electron microscopy

Live YTS and 221 transfectants were coincubated for 45 min at 37°C/5% CO₂ as previously described for IS formation. For negative staining, cells were fixed in 2.5% glutaraldehyde/2.5% paraformaldehyde in PBS for 2 h at 4°C after which the cells were washed three times in PBS. Cells were then postfixed in 1% osmium tetroxide/0.5 M veronal acetate, pH 7.2, for 1 h at room temperature, then taken through a graded series of alcohols and embedded in Spurr’s resin. Ultrathin sections were cut using a diamond knife and stained with 5% uranyl acetate at 60°C for 30 min followed by Reynolds lead citrate for 10 min at room temperature. Sections were examined with a transmission electron microscope (Philips 410; Endoven, The Netherlands) operating at 80 kV and images were recorded on Kodak 4489 film (Rochester, NY). Distances across the synapse were measured from the negatives between the outer edges of the outer membranes at the intercellular contact between YTS and 221 transfectants, using image analysis software (Leica Q500 MC).

Immunogold labeling

Cell conjugates were fixed in 4% paraformaldehyde/0.2% glutaraldehyde in PBS for 10 min on ice, followed by fresh 4% paraformaldehyde/0.2% glutaraldehyde in PBS on ice for 1 h. The fixed cells were then washed three times in PBS and dehydrated through a series of alcohols as the temperature was progressively lowered to −30°C in a freeze-substitution unit (EM AFS; Leica). Next, the cells were processed in Lowicryl K4 M embedding medium (TAAB Laboratories) infiltrated overnight in 1:1 ethanol/K4 M, followed by incubation with K4 M for 24 h and finally a further incubation with fresh K4 M that was polymerized under UV light for 8 h. Ultrathin sections were cut using an ultramicrotome (UCT; Leica) and collected on Formvar support grids. The grids were blocked before labeling with 0.02 M glycine for 15 min followed by 10% FCS in PBS for 30 min at room temperature. The grids were then incubated with 20 μg/ml mouse anti-GFP (JL-8; Clontech Laboratories) in 10% FCS in PBS for 1 h at room temperature. After washing three times in PBS, the grids were incubated for 1 h in 12 μg/ml goat anti-mouse IgG conjugated with 10 nm of gold (TAAB Laboratories) in the same diluent as before. Following a further 6 washes in PBS and 10 washes in ddH₂O, the grids were contrasted with uranyl acetate and lead citrate before examination with a transmission electron microscope (Philips CM100).

F-actin clusters at the activating, but not the inhibitory, NK cell IS

Conjugates of YTS/KIR2DL1 (YTS transfected to express KIR2DL1) or YTS/MOCK (mock-transfected) and 221/Cw6-GFP, 221/Cw4-GFP, or 221/Cw3-GFP were fixed and labeled with Alexa Fluor 633-conjugated phalloidin, a stain for f-actin, and were imaged by laser scanning confocal microscopy. At the intercellular contact between conjugates lacking either KIR2DL1 or its cognate MHC ligand, thus forming an activating NK cell IS, f-actin clearly clustered (Fig. 1⇓b). Throughout this study, clustering at the IS was confirmed by projecting a single optical slice through the IS with fluorescence intensity plotted on a third axis. Where the fluorescence intensity at the IS was more than the sum of the intensity of the unconjugated membranes, that label was defined as being clustered at the IS. The fluorescence intensity of phalloidin at the activating IS (Fig. 1⇓b) is clearly more than the sum of unconjugated membranes, whereas at the inhibitory IS (Fig. 1⇓a) no increase in fluorescence intensity of phalloidin above the sum of unconjugated membranes was observed. At the inhibitory IS, HLA-C clustered ∼70% of the time whereas f-actin accumulated only 30% of the time. In contrast, in activating synapses, f-actin was found to accumulate 70–80% of the time, while HLA-C-GFP clustered <15% of the time (Fig. 1⇓c).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

F-actin clusters at the activating, but not inhibitory, NK cell IS. Conjugates of 221 and YTS transfectants were fixed and labeled with Alexa Fluor 633 phalloidin to stain for f-actin. In both a and b, bright-field images of conjugates followed by HLA-C-GFP and phalloidin staining with the corresponding plots of fluorescence intensity are shown. Higher peaks represent areas of higher fluorescence intensity. a, An inhibitory NK cell IS between 221/Cw6-GFP and YTS/KIR2DL1 where HLA-Cw6-GFP, but not Alexa Fluor 633 phalloidin, clusters at the intercellular contact. b, YTS/KIR2DL1 forming an activating NK cell IS with a 221/Cw3-GFP cell. In contrast to the inhibitory NK cell IS, Alexa Fluor 633 phalloidin, but not HLA-C-GFP, accumulates. Scale bars represent 10 μm. c, The distribution of HLA-C-GFP and Alexa Fluor 633 phalloidin in activating and inhibitory conjugates was assessed and the mean percentages from six independent experiments are shown. Error bars are the SEM and n is the total number of conjugates scored.

Ezrin and CD43 are excluded from the inhibitory, but not activating, NK cell IS

Conjugates between 221/Cw6-GFP or 221/Cw3-GFP and YTS/KIR2DL1 or YTS/Mock were fixed and either stained for ezrin and f-actin (Fig. 2⇓) or ezrin and CD43 (Fig. 3⇓). Consistent with analysis by flow cytometry (data not shown), the 221 transfectant often stained brighter for ezrin than the YTS cell inferring a higher level of expression in 221 cells. At the inhibitory NK cell IS, a zone devoid of ezrin and CD43 staining corresponding to the region where HLA-C clustered at the IS was evident (Figs. 2⇓ and 3⇓). In contrast, at the activating IS, ezrin and CD43 were evenly distributed across the membranes of apposing cells within the conjugate (Figs. 2⇓ and 3⇓). At the inhibitory NK cell IS, ezrin and CD43 was excluded in 50–60% of inhibitory conjugates whereas at the activating NK cell IS, exclusion of either protein occurred <10% of the time (Figs. 2⇓d and 3⇓f). CD43 and ezrin were equally likely to be excluded from an inhibitory NK cell IS whether each NK cell simultaneously maintains one or several synapses (data not shown). Clustering of ezrin and CD43 at either the inhibitory or activating NK cell IS was observed in <10% of all conjugates assessed. In contrast to anti-CD43 staining, an isotype-matched mAb against KIR2DL1, EB6, was observed to brightly stain the NK cell juxtaposed to where HLA-C clusters on the target cell (Fig. 3⇓, c and e). This indicated that the staining procedure used does permit access of Abs to proteins within the inhibitory NK cell IS.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Ezrin is excluded from the inhibitory, but not activating, NK cell IS. Conjugates of 221 transfectants and YTS/KIR2DL1 cells were fixed and labeled with anti-ezrin mAb followed by Alexa Fluor 633 phalloidin to stain for f-actin. The distribution of HLA-C-GFP, ezrin, and f-actin at (a) the inhibitory NK cell IS formed between 221/Cw6-GFP and YTS/KIR2DL1 and (b) the activating NK cell IS formed between 221/Cw3-GFP and YTS/KIR2DL1 cells is demonstrated. c, A reconstruction of the face of an inhibitory IS is shown. Ezrin, but not f-actin, is excluded from the region where HLA-C-GFP clusters. Scale bars in a and b represent 10 μm, and 5 μm in c. d, Each inhibitory and activating NK cell IS was assessed for ezrin being clustered, excluded, or distributed homogeneously around the membranes of apposing cells. The mean percentages of conjugates exhibiting each ezrin distribution are shown. Error bars are the SEM and n is the total number of conjugates assessed.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

CD43 is excluded from the inhibitory, but not the activating, NK cell IS. Conjugates of 221 transfectants and YTS/KIR2DL1 were labeled with rabbit polyclonal anti-ezrin and mouse monoclonal anti-CD43 Abs. Columns a and b show the distribution of HLA-C-GFP, ezrin, and CD43 at the inhibitory and activating NK cell IS between (a) 221/Cw6-GFP and YTS/KIR2DL1 and (b) 221/Cw3-GFP and YTS/KIR2DL1. c, EB6 (anti-CD158a/KIR2DL1) staining at an inhibitory NK cell IS between 221/Cw6-GFP and YTS/KIR2DL1 is shown. d, Reconstruction of the face of the inhibitory NK cell IS shows that ezrin and CD43 are excluded from the inhibitory NK cell IS where HLA-C-GFP clusters. e, Reconstruction of the face of the IS demonstrates that KIR2DL1 clusters at the inhibitory NK cell IS, opposing HLA-C-GFP on the target cell. Scale bars represent 10 μm in a–c and 5 μm in d and e. f, Cells were fixed after coincubation for the times shown and the mean percentages of conjugates where ezrin or CD43 were clustered or excluded from the intercellular contact is shown. Error bars are SEM from at least three independent experiments. At least 98 conjugates were counted at each time point for both activating and inhibitory conjugates.

The time taken for the accumulation of HLA-C and the exclusion of ezrin and CD43 at the inhibitory NK cell IS was assessed by imaging cells that were fixed after different times of being incubated together (Fig. 3⇑f). Interestingly, after 10 min of coincubation of cells, HLA-C is clustered at ∼60% of inhibitory NK synapses, whereas ezrin and CD43 are excluded from only ∼20% of these synapses. This suggests that accumulation of HLA-C precedes exclusion of ezrin and CD43 at the inhibitory NK cell IS.

The extent to which ezrin and CD43 are excluded and HLA-C-GFP is clustered at the inhibitory IS was quantified using software to analyze mean fluorescence intensities from defined regions at the inhibitory NK cell IS relative to regions on the unconjugated cell membranes, as described in Materials and Methods. From a total of 40 conjugates counted, the fluorescence intensity of ezrin and CD43 was reduced by 68% ± 12.5 and 56.7% ± 10.2, respectively, at the inhibitory NK cell IS relative to the amount on the two unconjugated membranes. Interestingly, there was no apparent trend between the extent of HLA-C-GFP clustering and the fluorescence intensity of ezrin or CD43 at the IS (Table I⇓). Comparison of the amount of ezrin and CD43 staining on the NK cell and target cell with the amount at the inhibitory IS suggests that neither cell alone can account for the amount of ezrin and CD43 excluded from the IS. Thus, ezrin and CD43 from both the NK and target cell are excluded from the inhibitory NK cell IS.

Exclusion of ezrin and CD43 from inhibitory, but not activating, NK cell IS in conjugates involving human NK clones

Two human NK clones, 5H4 (KIR2DL1⁺KIR2DL2⁻) and 6H5 (KIR2DL1⁻KIR2DL2⁺), were selected for microscopy. By flow cytometry, both clones were found to be CD56⁺CD3⁻Ig-like transcript-2⁻CD94⁺. In agreement with the specificity of KIR for HLA-C alleles, lysis by NK clone 5H4, expressing KIR2DL1, was effectively inhibited by HLA-Cw6 while NK clone 6H5, expressing KIR2DL2, was inhibited by HLA-Cw3 expressed on 221 transfectants (Fig. 4⇓a).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Exclusion of ezrin and CD43 from the inhibitory, but not activating, NK cell IS in conjugates involving human NK clones. a, The cytotoxicity of two human clones, 5H4 (KIR2DL1⁺KIR2DL2⁻) and 6H5 (KIR2DL1⁻KIR2DL2⁺), was assessed in a 5-h ³⁵S release cytotoxicity assay. b, Inhibitory conjugates of 5H4 and 221/Cw6-GFP and (c) activating conjugates of 5H4 and 221/Cw3-GFP were stained for ezrin and CD43. Ezrin and CD43 are excluded from the inhibitory IS where HLA-C clusters but not the activating IS. d, Reconstructions of the face of two inhibitory NK cell IS, stained for ezrin or CD43, are shown. A region devoid of mAb staining for ezrin and CD43 corresponds to the region where HLA-C clusters. Scale bars represent 10 μm in b and c and 5 μm in d. e, The distribution of CD43 and ezrin at the inhibitory and activating NK cell IS was assessed in n conjugates.

Conjugates of NK clones 5H4 or 6H5 with 221/Cw6-GFP or 221/Cw3-GFP targets were fixed and stained for CD43 and ezrin using mAbs. Both CD43 and ezrin were excluded from the inhibitory IS between clone 5H4 (KIR2DL1⁺KIR2DL2⁻) and 221/Cw6-GFP (Fig. 4⇑b). At the activating IS, ezrin and CD43 were homogeneously distributed, shown in conjugates between 5H4 and 221/Cw3-GFP (Fig. 4⇑c). A reconstruction of the face of the inhibitory NK cell IS shows exclusion of CD43 and ezrin where HLA-Cw6-GFP clusters (Fig. 4⇑d). For both NK clones, ezrin and CD43 were excluded from the inhibitory NK cell IS 60–80% of the time (Fig. 4⇑e). This demonstrates that exclusion of ezrin and CD43 is a general property of the inhibitory NK cell IS rather than specific to an IS involving one type of KIR/MHC protein. At the activating NK cell IS, ezrin or CD43 was homogeneously distributed at the intercellular contact in ∼80% of conjugates (Fig. 4⇑e). Thus, the distribution of ezrin and CD43 at the inhibitory and activating IS in conjugates involving human NK clones derived from peripheral blood corresponds to that observed in conjugates of YTS and 221 transfectants (Figs. 2⇑ and 3⇑).

CD45 is excluded from the inhibitory, but not the activating, NK cell IS

With a view to further assessing the role of protein size in assembly of the NK cell IS, we imaged the distribution of CD45 in inhibitory and activating conjugates. Because alternative splicing produces various CD45 isoforms with extracellular domains ranging from 20 to 50 nm, we first established their respective distributions on the cells used for imaging. Abs specific for CD45RO (20–28 nm), CD45RA, and CD45RB (each ∼50 nm) were used to identify the expression of the various isoforms on the 221 and YTS transfectants as well as NK clones derived from peripheral blood. By flow cytometry, it was determined that YTS/KIR2DL1 expressed only CD45RO, 221/Cw6-GFP expressed only CD45RA, and the NK clones analyzed expressed CD45RO, CD45RA, and CD45RB (data not shown). This distribution of CD45 isoforms on human NK clones is consistent with previous observations (64). Conjugates involving effectors YTS/KIR2DL1 or NK clone 5H4 (KIR2DL1⁺KIR2DL2⁻) were fixed and stained with anti-CD45 mAb (HI30), which recognizes CD45RO and CD45RB. This mAb did not stain 221 cells, inferring that it does not detect CD45RA (our unpublished observations). At inhibitory conjugates, CD45RO expressed on YTS/KIR2DL1 (Fig. 5⇓, a and c) or CD45RO and CD45RB expressed on the NK clone 5H4 (Fig. 5⇓d) are clearly excluded from the region where HLA-C clusters (Fig. 5⇓, c and d). At the activating IS between YTS/KIR2DL1 and HLA-Cw3-GFP (Fig. 5⇓b), CD45RO was evenly distributed across the face of the IS in 87% of conjugates. At the inhibitory NK cell IS between YTS/KIR2DL1 and 221/Cw6-GFP, CD45RO was found to be excluded from the IS in 67% of conjugates (Fig. 5⇓e).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Exclusion of CD45 from the inhibitory, but not activating, NK cell IS. Inhibitory and activating conjugates of YTS/KIR2DL1 and 221/Cw6-GFP or 221/Cw3-GFP were fixed, stained with anti-CD45 mAb (HI30), and the distribution at the IS was determined. Inhibitory conjugates where CD45 is excluded from the IS are shown in a and c. b, At the activating IS, between YTS/KIR2DL1 and 221/Cw3-GFP, CD45 is evenly distributed across the intercellular contact. c, Reconstruction of the face of the inhibitory IS formed between YTS/KIR2DL1 and 221/Cw6-GFP demonstrates a region devoid of CD45 staining corresponding to that where HLA-C clusters. Scale bars represent 10 μm in a and b, and 5 μm in c. d, Reconstruction of the inhibitory IS formed between human NK clone 5H4 and 221/Cw6-GFP. Exclusion of CD45 from the region where HLA-C-GFP clusters is apparent. e, The distribution of CD45 at the inhibitory and activating IS was assessed in numerous conjugates involving YTS/KIR2DL1. n, The number of conjugates analyzed.

The size of the synaptic cleft at the inhibitory and activating NK cell IS

If proteins are segregated at the NK cell IS according to their size, it is necessary that the size of the synaptic cleft, defined in this study as the distance between the outer membranes of apposing effector and target cells at the IS, would vary accordingly. Thus, we assessed the ultrastructure of the activating and inhibitory NK cell IS in conjugates of 221 transfectants and YTS/KIR2DL1 cells by electron microscopy. NK cells were easily distinguished from target cells by the presence of large granules within their cytoplasm. The distance across the synaptic cleft in both activating and inhibitory conjugates varied such that distinct wide and narrow domains were evident (Fig. 6⇓a). Distances across the synaptic cleft of conjugates between 221/Cw6-GFP and YTS/KIR2DL1 and conjugates of 221/Cw3-GFP and YTS/KIR2DL1 cells were measured at 300-nm intervals along each intercellular contact. Although the size of the synaptic cleft in activating and inhibitory conjugates was found to range from 10 to 55 nm and 10 to 42 nm, respectively, the majority of regions measured between 10 and 30 nm at both the activating and inhibitory IS (Fig. 6⇓, c and d).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

The size of the synaptic cleft at the inhibitory and activating NK cell IS. Inhibitory and activating conjugates between 221 transfectants and YTS/KIR2DL1 cells were processed for electron microscopy and the distance across the synaptic cleft was measured. a, Negative staining of intercellular contacts between NK cells and 221 transfectants reveals how the distance across the synaptic cleft varies along the length of the IS. b, Immunogold labeling of HLA-Cw6-GFP at the inhibitory NK cell IS demonstrates 10-nm gold beads clustering at regions along the IS. All scale bars represent 0.5 μm. The distance across the synaptic cleft was measured along (c) the activating and (d) inhibitory NK cell IS and the percentage of measurements occurring within 5-nm intervals between 0 and 55 nm are shown. e, Following localization of HLA-Cw6-GFP at the inhibitory IS by immunogold labeling, the distance across the synaptic cleft at each gold particle was measured and the percentage occurrence of distances between 0 and 55 nm in 5-nm intervals is shown. Data was obtained from over 100 measurements across 20 synapses in each of c and d, and 30 measurements across 3 synapses in e.

To characterize the distribution of HLA-C at the inhibitory NK cell IS, immunogold labeling was used to locate HLA-C-GFP. HLA-C-GFP was detected on the surface, and to a lesser extent within the cytoplasm, of the target cell. Staining was occasionally observed on the apposing NK cell, which could reflect the intercellular transfer of HLA-C-GFP from the target cell to the NK cell (24). Strikingly, accumulation of the gold beads, marking the location of HLA-C-GFP, was observed only at the tighter intercellular contacts (Fig. 6⇑b). Although, it may be important to note that a few gold beads were located where the distance between the cells was ∼100 nm, which may or may not be part of the intercellular contact. The distance across the synaptic cleft at each gold bead that was clearly at the IS i.e., HLA-C-GFP at the IS, was measured and found to be 14.5 ± 2.2 nm (Fig. 6⇑e), which is the size of the extracellular domains of HLA-C bound to KIR (54, 55).