The NK cell receptor protein 1 (NKR-P1) (CD161) molecules represent a family of type II transmembrane C-type lectin-like receptors expressed predominantly by NK cells. Despite sharing a common NK1.1 epitope, the mouse NKR-P1B and NKR-P1C receptors possess opposing functions in NK cell signaling. Engagement of NKR-P1C stimulates cytotoxicity of target cells, Ca2+ flux, phosphatidylinositol turnover, kinase activity, and cytokine production. In contrast, NKR-P1B engagement inhibits NK cell cytotoxicity. Nonetheless, it remains unclear how different signaling outcomes are mediated at the molecular level. Here, we demonstrate that both NKR-P1B and NKR-P1C associate with the tyrosine kinase, p56lck. The interaction is mediated through the di-cysteine CxCP motif in the cytoplasmic domains of NKR-P1B/C. Disrupting this motif leads to abrogation of both stimulatory and inhibitory NKR-P1 signals. In addition, mutation of the consensus ITIM (LxYxxL) in NKR-P1B abolishes both its Src homology 2-containing protein tyrosine phosphatase-1 recruitment and inhibitory function. Strikingly, engagement of NKR-P1C on NK cells obtained from Lck-deficient mice failed to induce NK cytotoxicity. These results reveal a role for Lck in the initiation of NKR-P1 signals, and demonstrate a requirement for the ITIM in NKR-P1-mediated inhibition.
Natural killer cells are a subpopulation of large granular lymphocytes that have the ability to recognize and lyse certain virally infected or neoplastic cells (1, 2). Both stimulatory and inhibitory receptors are involved in the regulation of NK cell function (2, 3). Among these, the NK cell receptor protein 1 (NKR-P1)3 family of receptors has been of considerable interest, due to its conservation in rats, mice, and humans (4, 5, 6, 7).
One of the first NK cell receptors described (8), the original NK1.1 alloantigen (mouse NKR-P1C (9)) has been widely used as a marker to identify NK cells (10) and NKT cells (11) in C57BL/6 (B6) mice. Analysis of NK cells from different mouse strains revealed that the NK1.1 alloantigen epitope is shared by the products of two distinct genes, Nkr-p1b and Nkr-p1c (9, 12, 13). This discovery led to the first demonstration of the inhibitory function of mouse NKR-P1B (12, 13, 14) and to the cDNA cloning of a novel related inhibitory receptor, the B6-derived NKR-P1D (GenBank accession numbers, AF338321–2) (12, 13). Since then, there have been a total of five NKR-P1 proteins identified to date: three stimulatory receptors, NKR-P1A/C/F, which possess a charged transmembrane arginine (R) residue thought to be important for association with the FcRγ adaptor protein (6, 12, 15, 16); and two inhibitory receptors, NKR-P1B/D (6, 12, 16, 17, 18), which possess a consensus cytoplasmic ITIM (L/VxYxxL/I/V) (19).
With respect to the signaling pathways for NKR-P1 receptors, it has been shown that cross-linking of the stimulatory rat NKR-P1A or mouse NKR-P1C molecules stimulates phosphatidylinositol turnover and Ca2+ flux (20), as well as NK cell-mediated cytotoxicity and cytokine production (4, 21, 22). The functional consequences of treating human NK cells with anti-NKR-P1 mAb are more complex, resulting in no effect, activation, or inhibition, depending on the NK cell population studied (7, 23, 24). These diverse responses elicited by anti-NKR-P1 mAb suggest that additional, functionally distinct, isoforms or alleles of NKR-P1 may exist in humans.
Sequence analysis reveals that all murine NKR-P1 proteins possess the Cys-X-Cys-Pro (CxCP) motif also found in the cytoplasmic domains of CD4 and CD8 that mediates association with the Src-related nonreceptor protein tyrosine kinase, p56lck (25). Campbell and Giorda (26) have demonstrated a physical association between the rat NKR-P1A cytoplasmic CxCP motif and p56lck; however, no functional requirements for this association were shown. A feature of the inhibitory NKR-P1 receptors is the presence of an ITIM in the cytoplasmic domains of mouse NKR-P1B/D (12, 13, 14). We and others have shown that, like other ITIM-bearing receptors expressed by NK cells, mouse NKR-P1B binds Src homology 2 (SH2)-containing protein tyrosine phosphatase-1 (SHP-1) in a phosphorylation-dependent manner, suggesting a molecular mechanism for the inhibition of NK cell cytotoxicity through NKR-P1B (12, 13). However, a requirement for SHP-1 recruitment to the cytoplasmic ITIM in mediating NKR-P1 inhibition has not been shown.
In the present study, we analyzed the functional requirements for mouse NKR-P1 signaling. We found that both mouse NKR-P1B and NKR-P1C functionally associate with p56lck. In addition, mutation of the putative Lck-recruitment CxCP motif abolishes signal transduction through NKR-P1B/C. Furthermore, NK cells from lck−/− B6 mice lacked efficient NK1.1-mediated redirected lysis of FcR+ targets, demonstrating a direct functional requirement for Lck in NKR-P1C activation. Finally, mutation of the consensus ITIM of NKR-P1B was found to abolish both SHP-1 association and inhibitory signaling, providing direct evidence for the importance of the cytoplasmic ITIM in inhibitory NKR-P1 receptor function. Our data implicate Lck in the initiation of NKR-P1 signaling, and SHP-1 in the effector function of the inhibitory NKR-P1 receptors.
Materials and Methods
C57BL/6 mice were purchased from the NCI-Frederick Animal Production Area (Frederick, MD). Lck−/− mice (C57BL/6 background) were obtained from T. W. Mak (Ontario Cancer Institute, Toronto, Ontario, Canada).
The following cell lines were used: Jurkat (N. Berinstein, Aventis Pasteur, Toronto, Ontario, Canada), J.CaM1.6 (Dr. M. Julius, Sunnybrook and Women’s College Health Sciences Centre, Toronto, Ontario, Canada), P815 (B. Barber, University of Toronto, Toronto, Ontario, Canada), YTSeco (G. Cohen, Massachusetts Institute of Technology, Boston, MA), GP+E.86 (P. Ohashi, Ontario Cancer Institute, Toronto, Ontario, Canada), and PT-67 (BD Clontech). Jurkat and YTSeco cells were grown in complete RPMI 1640 (12.5% FCS). YTSeco cells were maintained with 2 mg/ml G418 added. GP+E.86 and PT67 cells were grown in complete DMEM (10% FCS). All cells were maintained at 37°C in a humidified 5% CO2 atmosphere.
Expression constructs and NKR-P1 variants
Mouse NKR-P1B (Swiss: Sw) and NKR-P1C (C57BL/6: B6) constructs were previously cloned (12). Cloned NKR-P1 genes were excised from pcDNA3.1/V5/His-TOPO using BglII and PmeI, then subcloned into the BglII/SnaBI cloning site of pMIEV retroviral vector (R. Hawley, American Red Cross, Rockville, MD). This construct was engineered to express NKR-P1 genes 5′ of the internal ribosomal entry site, allowing bicistronic expression of NKR-P1 and GFP.
Variant constructs were generated using PCR-based mutagenesis (Stratagene). The following primers were used: YF8, CAA CAA CAC TGG TCT TTG CAG ATT TAA ACC TAG; LS6, GGA TTC AAC AAC ATC GGT CTA TGC AGA T; YF7, ACA GCA AGT ATC TTC CTC GGT TTA AAG CC; P1C SXCP, CAG ATG CCT CTC GGT GCC CAC GTT CAC; P1B SXCP, TTT CCC CAG ACA CCT CTC GGT GCC CTC. PCR products were cloned into pcDNA3.1/V5/His-TOPO, excised with BglII, and inserted into the BglII site of pMIEV. All variants were confirmed by sequencing.
Flow cytometry and cell sorting
Cells were stained as previously described (27). Briefly, single-cell suspensions were stained for surface expression of various markers using PE- or biotin-conjugated mAbs obtained from BD Biosciences or eBioscience in staining buffer (HBSS with 1% BSA and 0.05% NaN3). Cells were stained in 100 μl for 30 min on ice and washed twice before analysis. Analysis was performed using a FACSCalibur flow cytometer (BD Biosciences). For cell sorting, single-cell suspensions were prepared and stained for FACS as described above, except that no NaN3 was added to staining buffer. Cells were sorted using a FACSDiVa (BD Biosciences). Sorted cells were >99% pure, as determined by postsort analysis. Staining was not altered in the presence of blocking FcγRII/III Ab (2.4G2).
PT-67 and GP+E.86 cells were used as previously described (28). YTSeco cells were spin-infected with retroviral supernatants from stably transfected GP+E.86 cells (plus 4 μg/ml polybrene) at 2200 rpm for 90 min at 32°C, then sorted the next day. Jurkat cells were infected with retroviral supernatants from transient triple-transfected 293T cells as previously described (18), then sorted at day 3–4 following infection.
Cells were washed twice in complete RPMI medium and precoated with biotin-conjugated anti-NK1.1 (PK136; mouse IgG2aκ) or anti-CD3ε (OKT3; mouse IgG2aκ) mAb (1 μg/ml mAb/106 cells) on ice for 30 min, then washed with warm medium and resuspended at 1.5 × 106 cells/ml. Cells were loaded with 1-[2-amino-5-(6-carboxyindol-2-yl)phenoxy]-2-(2′-amino-5′-methylphenoxy)ethane-N,N,N′,N′-tetraacetic acid peutaacetoxymethylester (Molecular Probes) at final concentration of 5 μM for 45 min at 37°C in the dark. Following loading, the cells were washed and resuspended at concentration of 2 × 106 cells/ml in complete RPMI. Each sample was prewarmed for 3 min at 37°C in a water bath before being run on a FACSDiVa flow cytometer. After measurement of baseline Ca2+ signal for 1 min, 60 μg of avidin (Sigma-Aldrich) was added to the tube (0.5 ml) to cross-link the primary mAb. Data were analyzed using FlowJo software (Tree Star). Percent response was calculated according to the following equation: percent Ca2+ response = [experimental Ca2+ response − baseline Ca2+ response]/[index Ca2+ response − baseline Ca2+ response] × 100%; experimental response represents that in the presence of α-NK1.1 Ab for NKR-P1C variants, and α-NK1.1 plus α-CD3ε Abs for NKR-P1B variants; baseline response represents Ca2+ response in the absence of cross-linking; index response represents that in the presence of α-CD3ε Ab. Response values represent integrated areas of the magnitude of Ca2+ responses over time.
IFN-γ production by NKT cells
NKT cells were isolated by sorting NK1.1+CD3+ cells from spleens and bone marrows of 6- to 8-wk-old CD-1 mice. A total of 2.4 × 105 NKT cells were plated on 96-well plates precoated with 10 μg/ml of either α-CD3 (clone 145-2C11) or α-CD3 plus α-NK1.1 (clone PK136) mAbs. The cells were left on plates for 72 h and then intracellular staining was performed for IFN-γ. Intracellular staining for IFN-γ was done with the Cytofix/Cytoperm with GolgiStop kit (BD Biosciences) according to the manufacturer’s instructions.
51Cr-release Ab-induced redirected lysis (AIRL) assays
For B6 and lck−/− effectors, we used bone marrow cells that were NK1.1 enriched by complement-mediated CD8 depletion using YTS169 mAb (rat IgG2b; obtained from Dr. P. Poussier, Sunnybrook and Women’s College Health Sciences Centre) and Low-Tox complement (Cedarlane Laboratories). These cells were then grown in high rhIL-2 (1000 U/ml) for 7 days, generating lymphokine-activated killers (LAKs) that were >98% NK1.1+. Target cells (5 × 106 cells/ml) were loaded with 50 μCi of Na251CrO4 (Amersham) at 37°C for 1.5 h in complete RPMI with 60% FBS. Radiolabeled targets were resuspended at 5 × 104 cells/ml and placed in triplicate wells in 100-μl aliquots (5 × 103 cells/well) in V-bottom microtiter plates. Effector cells were mixed with 51Cr-labeled targets at various E:T ratios with 0.5 μg per well of purified Abs (α-NK1.1, PK136 (mouse IgG2aκ); α-2B4, c1.7 (mouse IgG1); α-trinitrophenyl isotype control, G155–178 (mouse IgG2aκ); α-I-Ek isotype control, 14.4.4S (mouse IgG2aκ). Plates were centrifuged for 2 min at 1200 rpm then placed in culture for 4 h at 37°C. After lysis, 100 μl of culture supernatants were collected and radioactivity was measured in a γ counter (Cobra Quantum, Packard BioScience Company). Supernatant from target cells cultured alone or target cells plus 1% SDS gave the spontaneous or maximal release counts, respectively. Percent specific lysis was calculated according to the following formula: percent SL = (experimental − spontaneous)/(maximal − spontaneous) × 100%. The error bars in the cytotoxicity assays represent SEM of the specific lysis. Each experiment was repeated at least three times.
Immunoprecipitation and Western blotting
A total of 107 YTSeco cells were left unstimulated or stimulated with pervanadate for 20 min, as described previously (12), washed once in serum-free medium, then lysed for 15 min in ice-cold lysis buffer (0.5% TNTE (0.5% Triton X-100, 150 mM NaCl, and 20 mM Tris-Cl, pH 7.4) plus 50 mM NaF, 10 mM sodium pyrophosphate, 1 mM sodium orthvanadate, and protease inhibitors). Lysates were centrifuged at 14,000 × g for 10 min to remove cell debris, then precleared for 2 h at 4°C on protein G-Sepharose beads that had been loaded with 5 μg of control mAb. A total of 5 μg of biotin-conjugated α-NK1.1 or biotin-conjugated isotype control mAb (α-trinitrophenyl, G155–178) was added to precleared lysate and incubated with gentle agitation for 2–3 h. A total of 20 μl of streptavidin agarose beads (Pierce) was added to lysate and incubated for an additional 2 h with gentle agitation. Beads were washed four times with wash buffer (0.1% TNTE plus 10 mM NaF, 10 mM sodium pyrophosphate, and 1 mM sodium orthovanadate), boiled for 4 min in protein-loading buffer containing DTT, then proteins were resolved on 10% SDS-PAGE gels and electroblotted onto polyvinylidene difluoride membranes. After blocking with 3% BSA in TBS-T (50 mM Tris-Cl (pH 8.0), 500 mM NaCl, and 0.1% Tween 20), membranes were incubated with monoclonal anti-human-SHP-1 Ab (clone 52, mouse IgG1 specific for the C-terminal tail of human SHP-1; kind gift of Dr. F. Tsui, University of Toronto) at 1/500 dilution in TBS-T for 2–3 h at room temperature, then washed. Blots were visualized using 1/4000 anti-mouse HRP in TBS-T and an ECL Plus Kit (Amersham).
For Lck coimmunoprecipitation studies, 3 × 107 Jurkat cells were precoated with biotinylated α-NK1.1 (PK136; 1 μg/106 cells) for 30 min at 4°C in a 1-ml volume. Cells were washed twice with ice-cold wash buffer (PBS plus 3% FBS), pelleted, and resuspended in a total volume of 50 μl of wash buffer, then prewarmed to 37°C for 1 min. Cells were left unstimulated (0′ stimulation; see Fig. 5) or activated by addition of streptavidin to a final concentration 50 μg/ml at 37°C for the indicated times. Cells were then lysed in 1 ml of ice-cold lysis buffer (TNE plus 50 mM Tris, pH 8, 20 mM EDTA, 200 μM Na3VO4, 50 mM NaF, 1% Nonidet P-40, 20 μg/ml leupeptin, 20 μg/ml aprotinin). After centrifugation, the postnuclear lysates were precleared for 1 h at 4°C using protein A-Sepharose beads. NKR-P1 complexes were immunoprecipitated from the lysates with protein A-Sepharose beads precoated for 4 h at 4°C with 5 μg of α-NK1.1 (PK136) Ab. Immunoprecipitates were then washed six times before resolution by 10% SDS-PAGE followed by electroblotting onto polyvinylidene difluoride membranes. After blocking with 5% milk in TBS-T, membranes were incubated with rabbit anti-mouse Lck Ab (29), followed by HRP-conjugated protein A (BioRad) and developed with an ECL kit (Amersham). A 5-μl aliquot from each cell lysate was analyzed directly for the presence of Lck by Western blotting as described above.
Generation of NKR-P1 variants and stable transductants
Attempts to define NK cell receptor signaling pathways in the mouse are hampered by the lack of available mouse NK cell lines. Because NK and T lymphocytes are derived from a common progenitor cell (30, 31, 32) and share numerous cell surface receptors and signaling pathways, we used the immortalized human T cell (Jurkat) and NK cell (YTSeco) lines as a tools to study the NKR-P1 signaling pathways (33). To this end, various constructs encoding native and genetically engineered forms of the mouse NKR-P1B/C molecules were generated (Fig. 1, A and B). One approach involved PCR-based site-directed mutagenesis to introduce coded amino acid substitutions in the cytoplasmic domains of NKR-P1B (Y8F, L6S, and SxCP), and NKR-P1C (Y7F and SxCP) (Fig. 1, A and B). These NKR-P1 variants were transduced into Jurkat and YTSeco cells in the form of a bicistronic retroviral vector (pMIEV). Transduced cells were sorted based on both GFP fluorescence and expression of the shared NK1.1 (NKR-P1B/C) epitope for GFP+ NK1.1+ cells. See Figs. 1,C and 3F1 A for a depiction of expression levels of GFP and NK1.1 on the sorted Jurkat and YTSeco cells, respectively. These cells were used in subsequent experiments.
NKR-P1B/C cytoplasmic residues are crucial for mAb-induced Ca2+ flux
It has been shown that mAb-induced cross-linking of the stimulatory murine NKR-P1A/C molecules induces calcium flux in NK cells (13, 20). To study the effects of amino acid substitutions in the cytoplasmic domains of the NKR-P1 receptors on signaling, we tested the ability of various Jurkat transductants to induce Ca2+ mobilization. Biotinylated α-NK1.1 (anti-NKR-P1B/C) and control biotinylated α-CD3ε mAb’s were used in combination with secondary avidin to induce cross-linking and subsequent Ca2+ flux in transductants loaded with Indo-1. As shown in Fig. 2, A and B, α-NK1.1 cross-linking induces Ca2+ mobilization in Jurkat cells expressing a native NKR-P1C molecule. In fact, α-NK1.1 cross-linking of the native NKR-P1C receptor induces 70% of Ca2+ mobilization relative to that of α-CD3ε cross-linking (Fig. 2,B, dashed line). The NKR-P1 cytoplasmic YxxL sequence has been suggested to meet functional constraints for a phospholipase-Cγ binding motif (14). However, it has yet to be determined whether the NKR-P1C YxxL motif is involved in function, as it is absent in the stimulatory NKR-P1F receptor sequence (16). As shown in Fig. 2 B, cross-linking of the NKR-P1C variant, Y7F, produces only ∼55% of the Ca2+ mobilization signal relative to that of native NKR-P1C. This suggests that although the YxxL motif may play a partial role in NKR-P1C signaling, it is not strictly required for function.
It has been reported that the cytoplasmic domain of the rat NKR-P1A receptor interacts with p56lck via a di-cysteine motif in its cytoplasmic domain (CxCP motif) (26). Substitution of both cysteine residues with alanines abrogated binding of rat NKR-P1A to Lck (26). However, the functional relevance of this interaction was not ascertained. Because the Lck-binding motif in the murine NKR-P1 receptors (CxCPR/H) closely resembles that of CD4 (CxCPH) and CD8α (CxCPR) (26), we generated an NKR-P1C variant in which the first cysteine in the CxCP motif was replaced by serine (SxCP variant). As shown in Fig. 2, A and B, cross-linking of the NKR-P1C SxCP variant abolishes Ca2+ mobilization. This suggests that the CxCP motif in the stimulatory NKR-P1C receptor is required for initiation of Ca2+ flux, perhaps through recruitment of Lck.
Next, we tested whether NKR-P1B-mediated signals could inhibit α-CD3ε-induced signaling. To this end, NKT cells were isolated from NKR-P1B-expressing CD-1 mice and tested by intracellular staining for IFN-γ production upon α-CD3ε cross-linking. Co-cross-linking of NKR-P1B and CD3ε blunted the IFN-γ response of NKT cells relative to CD3ε cross-linking alone (data not shown). To extend these findings, several Jurkat transductants expressing NKR-P1B variants were generated and tested for their ability to induce Ca2+ mobilization. As shown in Fig. 2,B, both the native and SxCP variants of NKR-P1B elicited minimal Ca2+ flux, as expected, because NKR-P1B represents an inhibitory NK cell receptor (12, 13). We next tested the ability of these variants to inhibit the Ca2+ mobilization induced by α-CD3ε cross-linking. Notably, co-cross-linking of native NKR-P1B with CD3ε resulted in a consistently blunted (∼65%) Ca2+ flux response relative to that induced by α-CD3ε alone (Fig. 2,C). In contrast, none of the NKR-P1B variants tested (Y8F, L6S, and SxCP) were able to inhibit α-CD3ε induced Ca2+ mobilization in three independent trials (Fig. 2 C). The inability of the ITIM mutants of NKR-P1B (Y8F and L6S) to inhibit α-CD3ε induced Ca2+ flux suggests that an intact ITIM may be required for inhibitory function. In addition, the inability of the SxCP variant to inhibit α-CD3ε induced Ca2+ mobilization suggests that NKR-P1B may also require Lck for initiating phosphorylation of the ITIM before recruitment of SHP-1 phosphatase.
To further explore a requirement for Lck in NKR-P1 signaling, we examined Ca2+ flux of the native NKR-P1B/C receptors in the Lck-deficient mutant Jurkat cell line, J.CaM1.6 (34). Importantly, both parental Jurkat and Lck-deficient J.CaM1.6 cells had similar α-CD3-induced Ca2+ responses (data not shown). However, unlike their parental Jurkat counterparts, J.CaM1.6 cells transduced with a native NKR-P1C receptor are unable to induce Ca2+ mobilization upon anti-NK1.1 cross-linking (Fig. 2,D; black bars). Moreover, J.CaM1.6 cells transduced with a native NKR-P1B molecule possessed a similar α-CD3ε-induced Ca2+ mobilization in response to α-NK1.1 co-cross-linking relative to cross-linking of α-CD3ε alone (Fig. 2 D; gray bars). These results support a role for Lck in the initiation of NKR-P1B/C signaling.
The NKR-P1B cytoplasmic ITIM is essential for inhibitory function
To directly examine the requirements for NKR-P1-mediated inhibition, we examined the functional response of variant NKR-P1B receptors in AIRL assays (35) using human YTSeco transductants (Fig. 3 A) and the FcR+ target cell line, P815. As a positive control for lysis, we tested AIRL using the α-2B4 Ab. 2B4 (CD244) is a receptor expressed on all human NK cells and a subset of CD8+ T cells, and ligation of 2B4 (using c1.7 mAb) induces redirected lysis of FcR-bearing target cells (36).
Fig. 3,B demonstrates that ligation of human 2B4 on the surface of YTSeco cells induces lysis of murine P815 targets. Therefore, we tested the ability of NKR-P1B variants to mediate redirected inhibition in the presence of the stimulatory 2B4 signal. As expected, co-cross-linking of the native NKR-P1B receptor by PK136 mAb significantly inhibits 2B4-induced NK cell cytotoxicity (Fig. 3 B). This supports previous results using mouse NK cells (12, 13) and demonstrates that mouse NKR-P1B can dominantly inhibit the 2B4-mediated stimulatory signal in human YTSeco NK cells.
To examine the role of the cytoplasmic ITIM of NKR-P1B, we tested redirected inhibition of YTSeco transductants expressing an NKR-P1B variant in which the ITIM was mutated (Y8F variant). Fig. 3 B demonstrates that co-cross-linking of the Y8F variant has no effect on 2B4-induced redirected lysis of P815 targets. These results formally demonstrate that an intact ITIM in the NKR-P1B cytoplasmic domain is required for NKR-P1B-mediated inhibition.
Mutation of the NKR-P1B ITIM abrogates recruitment of SHP-1
A number of inhibitory receptors have been shown to exert their inhibitory effects on NK cell function by interrupting the early tyrosine phosphorylation pathways responsible for NK cell activation (37, 38). We have previously demonstrated that the NKR-P1B receptor binds to SHP-1 in a phosphorylation-dependent manner (12). However, to directly demonstrate the importance of the NKR-P1B ITIM for recruitment of SHP-1 phosphatase, we performed immunoprecipitation experiments on YTSeco transductants. As previously shown (12), Western blot analysis on α-NK1.1 immunoprecipitates reveals that the native NKR-P1B molecule associates with SHP-1 upon pervanadate stimulation (Fig. 3,C; pervanadate+/IP α-NK1.1+). This association is phosphorylation-dependent, as SHP-1 does not coimmunoprecipitate with NKR-P1B in unstimulated cells (Fig. 3,C; pervanadate−/IP α-NK1.1+) (12). In contrast, SHP-1 recruitment is completely abrogated in the ITIM-substituted Y8F variant (Fig. 3 C; pervanadate+/IP α-NK1.1+). These data show that an intact ITIM is required for phosphorylation-dependent recruitment of SHP-1 by the inhibitory NKR-P1B receptor.
Lck functionally associates with mouse NKR-P1B and NKR-P1C
To examine Lck recruitment to the mouse NKR-P1B/C receptors, Jurkat cells transduced with either native NKR-P1B/C or variant P1B-SxCP or P1C-SxCP molecules were analyzed for association with p56lck. As shown in Fig. 4, A and B, Western blot analysis for Lck on α-NK1.1 immunoprecipitates reveals that the stimulatory NKR-P1C molecule is constitutively associated with Lck, and this association increases with α-NK1.1-induced stimulation in a time-dependent manner. This association was not observed in parental Jurkat control cells (Fig. 4,A). Furthermore, Lck also associates with the inhibitory NKR-P1B receptor, and recruitment of Lck is enhanced with α-NK1.1 stimulation (Fig. 4,B). In contrast, association of NKR-P1C with Lck is disrupted when only the first residue in the cytoplasmic di-cysteine CxCP motif is substituted with a serine residue (Fig. 4,B, C-SxCP). The association of NKR-P1B with Lck in the SxCP variant is also significantly reduced, although to a lesser extent than NKR-P1C (Fig. 4,B, B-SxCP). This latter finding might be due to only a partial disruption of the CxCP motif, as it has been previously shown that mutation of both cysteine residues abolishes Lck binding in yeast two-hybrid analysis (26). Importantly, the disruption of only a single cysteine in NKR-P1B CxCP motif not only led to a reduction in Lck coimmunoprecipitation, but also abrogated the induced recruitment of SHP-1 to the receptor (Fig. 4 C). This supports a role for Lck activity in mediating SHP-1 recruitment. These results demonstrate that both the inhibitory NKR-P1B and stimulatory NKR-P1C receptors functionally associate with p56lck, and that this association involves their cytoplasmic di-cysteine CxCP motifs.
Lck is necessary for efficient initiation of NKR-P1C signaling
Although lck−/− mice have previously been reported to possess normal NK cell development, natural killing, and Ab-dependent cellular cytotoxicity (ADCC) (39), the involvement of Lck in NKR-P1-mediated redirected lysis has not been assessed. Therefore, to investigate the physiological function of p56lck in NKR-P1 signaling, we examined NK1.1-mediated redirected lysis using NK cells from p56lck-deficient mice (B6 lck−/− mice). As a positive control for NK cell cytolytic function, we tested α-2B4-mediated AIRL, as 2B4 signaling has been shown to signal through Fyn (40). As shown in Fig. 5,A, NK cells from lck−/− mice possess normal α-2B4-mediated redirected lysis when compared with their wild-type B6 counterparts. Strikingly, however, lck−/− NK cells failed to respond to α-NK1.1-mediated redirected lysis, while NK cells from wild-type B6 mice responded normally (Fig. 5,B). However, because both CD16 and NKR-P1C are thought to signal through FcRγ, we re-examined ADCC mediated by lck−/− NK cells using the anti-CD16 mAb, 2.4G2. In contrast to previously published results (39), we found a defect in 2.4G2-mediated ADCC in lck−/− NK cells (Fig. 5,D) relative to control 2B4-mediated AIRL (Fig. 5 C). These results demonstrate that the Lck tyrosine kinase plays a nonredundant role in FcRγ-mediated signaling through NKR-P1C and CD16 on mouse NK cells.
Previous biochemical studies have yielded some valuable insights into the downstream signaling pathways triggered through the NKR-P1B and NKR-P1C receptors in mouse NK cells (12, 13, 15, 22). However, the study of NK cell signaling has been restricted by the lack of available mouse NK cell lines or an alternate suitable system to manipulate NK cell functionality. In this study, we have used retroviral expression of mouse stimulatory and inhibitory NKR-P1 receptor variants in human T and NK cell lines to investigate the role of downstream signals (33). Our analysis provides direct evidence that Lck functionally associates with a CxCP motif in the cytoplasmic tails of mouse NKR-P1B/C and is necessary for the initiation of NKR-P1 signaling. In addition, we demonstrate that an intact ITIM (LxYxxL) located in the cytoplasmic domain of NKR-P1B is essential for the recruitment of the SHP-1 phosphatase and inhibitory function.
NKR-P1C has been reported to associate with the ITAM-containing FcRγ chain, which was shown to be critical for redirected lysis through NKR-P1C (15). The noncovalent association of FcRγ is thought to involve a charged arginine (R) residue found only in the transmembrane domains of stimulatory NKR-P1 receptors (Fig. 1 A; this R is absent in the inhibitory NKR-P1 sequences) (12). Engagement of NKR-P1C is thought to result in phosphorylation of the FcRγ chain ITAMs with subsequent recruitment of the SH2 domain-containing Syk-family protein tyrosine kinase (41). Downstream events include phosphoinositide turnover, mobilization of intracellular calcium, tyrosine kinase activation, degranulation, and release of cytokines. Insight into the mechanism by which NKR-P1C activates NK cells can be inferred from the sequence motifs within the NKR-P1C cytoplasmic domains. Such analysis of the murine NKR-P1 receptors reveals a consensus p56lck tyrosine kinase-binding motif characterized by the sequence CxCP. This motif has been studied in T cells, where it has been shown to be required for the constitutive binding of p56lck by the cytoplasmic domains of CD4 and CD8 (25, 42). Although recent reports have demonstrated a physical association of Lck with NKR-P1A in rat NK cells (26), the functional significance of these observations was not assessed.
Our current study provides a detailed examination of the involvement of Lck in signaling through the mouse NKR-P1B/C receptors. We demonstrate that Lck associates with both NKR-P1B and NKR-P1C (Fig. 4). Moreover, this association is abrogated or at least partially disrupted by substitution of a single serine (S) residue into the cytoplasmic CxCP motif of NKR-P1C or NKR-P1B, respectively (Fig. 4,B). This finding has direct functional consequences, as Jurkat cells transduced with the SxCP variant of NKR-P1C are incapable of mobilizing intracellular calcium compared with Jurkat cells expressing a native NKR-P1C receptor (Fig. 2, A and B). Furthermore, Ca2+ mobilization is absent upon cross-linking of a native NKR-P1C receptor on the surface of J.CaM1.6 cells, which lack endogenous Lck activity (Fig. 2,D). Finally, although NK cells from lck−/− mice were reported to possess normal natural killing activity and ADCC (39), we found that lck−/− NK cells failed to efficiently induce both NK1.1 (NKR-P1C)-mediated redirected lysis and CD16-mediated ADCC (Fig. 5, B and D). This loss of function is not due to a lack of NK cell activity, as lck−/− NK cells possess normal redirected killing through 2B4 (Fig. 5 A). Therefore, inability of lck−/− NK cells to efficiently redirect killing through the stimulatory NKR-P1C receptor directly demonstrates that Lck plays a nonredundant role in NKR-P1 signaling. Moreover, our data showing a defect in lck−/− NK cell-mediated ADCC shows that Lck also plays a nonredundant role in FcRγ-dependent signaling.
In addition to the Lck requirements shown for the stimulatory NKR-P1C and CD16 receptors, as mentioned above, we also show that Lck plays a role in the initiation of NKR-P1B-mediated inhibitory signals. It was previously shown that cross-linking of NKR-P1B on mouse NK cells leads to inhibition of function, and that NKR-P1B recruits the SHP-1 phosphatase in a phosphorylation-dependent manner (12). Here, we directly demonstrate that disruption of the cytoplasmic ITIM of NKR-P1B abrogates both SHP-1 recruitment and inhibitory function of the receptor (Fig. 3, B and C). These results demonstrate that an intact ITIM is required for SHP-1 recruitment and subsequent inhibition of NK cell function. Consistent with the notion that Lck plays a role in the initiation of NKR-P1B-mediated inhibitory signals, we show that mutating the Lck docking site in NKR-P1B disrupts the induced association with SHP-1 (Fig. 4 C).
Thus, we propose a model for the initiation and effector signaling through the stimulatory and inhibitory mouse NKR-P1 receptors (Fig. 6, A and B). Presumably, the initial phosphorylation of the cytoplasmic ITIM tyrosine in NKR-P1B is mediated by Lck, providing a docking site for the SH2 domain of the SHP-1 phosphatase, leading in turn to dephosphorylation and inhibition of proximal kinases. In contrast, association of Lck with the stimulatory NKR-P1C receptor leads, upon cross-linking, to transphosphorylation of the cytoplasmic ITAM tyrosines in the FcRγ adaptor protein, in turn leading to recruitment of the Syk kinase and activation of downstream second messengers. Thus, our findings support a stepwise model for the signaling requirements of the stimulatory and inhibitory NKR-P1 receptors (Fig. 6). The recent identification of cognate ligands for the NKR-P1 receptors (17, 18) will lead to new insight into the physiology of NKR-P1-mediated recognition of target cells by this functionally dichotomous receptor family.
The authors have no financial conflict of interest.
We thank Andre Veillette for the anti-Lck Ab. We gratefully acknowledge Gisele Knowles for expert assistance in cell sorting.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by funds from the Canadian Institutes of Health Research. B.L. was supported by an Ontario Graduate Scholarship. J.R.C. was supported by a Long Term Fellowship from the Human Frontier Science Program. J.C.Z.-P. is supported by an Investigator Award from the Canadian Institutes of Health Research.
Abbreviations used in this paper: NKR-P1, NK cell receptor protein 1; ADCC, Ab-dependent cellular cytotoxicity; AIRL, Ab-induced redirected lysis; SH2, Src homology 2; SHP-1, SH2-containing protein tyrosine phosphatase-1; LAK, lymphokine activated killer.