The two isoforms of phospholipase C (PLC)-γ couple immune recognition receptors to important calcium- and protein kinase C-dependent cellular functions. It has been assumed that PLC-γ1 and PLC-γ2 have redundant functions and that the receptors can use whichever PLC-γ isoform is preferentially expressed in a cell of a given hemopoietic lineage. In this study, we demonstrate that ITAM-containing immune recognition receptors can use either PLC-γ1 or PLC-γ2, whereas the novel NK cell-activating receptor NKG2D preferentially couples to PLC-γ2. Experimental models evaluating signals from either endogenous receptors (FcR vs NKG2D-DAP10) or ectopically expressed chimeric receptors (with ITAM-containing cytoplasmic tails vs DAP10-containing cytoplasmic tails) demonstrate that PLC-γ1 and PLC-γ2 both regulate the functions of ITAM-containing receptors, whereas only PLC-γ2 regulates the function of DAP10-coupled receptors. These data suggest that specific immune recognition receptors can differentially couple to the two isoforms of PLC-γ. More broadly, these observations reveal a basis for selectively targeting the functions initiated by distinct immune recognition receptors.
Multisubunit immune recognition receptors are used by lymphoid effector cells to initiate cellular activation. Many of these receptor complexes (including the TCRs, BCRs, and FcRs) are composed of ligand binding subunits noncovalently associated with one or more signal-transducing, transmembrane adapter molecules. A common structural feature of the adapter molecules is one or more copies of an ITAM motif, the loose consensus sequence of which is YxxL/I(x6–8)YxxL/I (where x denotes any amino acid). NKG2D is a different kind of immune recognition receptor (1). The ligand-binding portion of this receptor recognizes “stress-induced,” MHC class I-like structures (2), and the signal-transducing subunit DAP10 lacks an ITAM motif. Instead, DAP10 possesses a PI3K binding motif (YINM) (3). ITAM-containing receptor complexes and the YINM-containing NKG2D-DAP10 complex couple to phospholipase C (PLC) 3-γ-dependent calcium signaling mechanisms in cytotoxic lymphocytes to generate granule exocytosis and cell-mediated killing.
Phosphoinositide-specific PLC isoforms hydrolyze the membrane lipid phosphatidylinositol 4,5-bisphosphate into inositol-1,4,5-trisphosphate and sn-1,2-diacylglycerol, which, in turn, mediate the mobilization of intracellular calcium and the activation of protein kinase C, respectively (4). The PLC-γ isoforms couple proximal protein tyrosine kinase activation to downstream calcium signaling. PLC-γ1 and PLC-γ2 have similar subdomain structures (i.e., an N-terminal pleckstrin homology domain, two Src homology 2 domains, a Src homology 3 domain, a C2 domain, several EF hand domains, a catalytic region, and potential tyrosine phosphorylation sites) and overall have 67% homology. These two isoforms are differentially expressed in hemopoietic cells (e.g., predominantly PLC-γ1 in T cells, and predominantly PLC-γ2 in B cells), and separate targeted disruption of a specific PLC-γ isoform can differentially influence the activation of certain hemopoietic cells (e.g., abnormal B cell activation in PLC-γ2 knockouts) (5, 6). However, from knockout data, it has been difficult to know whether the functional abnormalities observed are due to the loss of the predominantly expressed isoform or whether there is selective use of PLC-γ isoforms by specific receptors. NK cells express both PLC-γ1 and PLC-γ2 (7), and targeted disruption of PLC-γ2 alone results in reduced Ab-dependent cellular cytotoxicity (ADCC) and reduced natural cytotoxicity against Yac tumor targets (5). Yet, the previous analyses do not address whether any of the NK-activating receptors selectively couple to a specific PLC-γ isoform.
In this report, we extend previous analyses by evaluating whether specific multisubunit immune recognition receptors differentially couple to the two PLC-γ isoforms. We show that endogenously expressed, ITAM-containing receptor complexes (e.g., FcRs on NK cells) and ectopically expressed chimeric receptors with ITAM-containing cytoplasmic tails biochemically and functionally couple to both PLC-γ1 and PLC-γ2. In contrast, endogenously expressed, DAP10-coupled NKG2D receptors or chimeric receptors with DAP10-containing cytoplasmic tails biochemically and functionally couple selectively to PLC-γ2. These data indicate that whereas certain receptors can use either PLC-γ isoform (e.g., ITAM-containing receptor complexes), other receptors are selective in their use of specific PLC-γ isoforms (e.g., use of PLC-γ2 by the NKG2D-DAP10 receptor complex).
Materials and Methods
Reagents, cells, Abs
Unless otherwise stated, all chemicals were obtained from Sigma-Aldrich. The human PLC-γ1-deficient Jurkat subline J.gamma 1 was kindly provided by R. T. Abraham (Burnham Institute, San Diego, CA). The murine mastocytoma P815 was obtained from American Type Culture Collection. Human NK cells and CD8+ T cells were cloned and passaged as described previously (8). We purified the 3G8 mAb to human FcγRIIIA by affinity chromatography over protein A-agarose. We also used mAb to NKG2D (R&D Systems). PLC-γ1 and PLC-γ2 peptide fragments were used to create rabbit polyclonal Abs as described previously (7).
The Flag-tagged CD4 chimeric receptors were generated by PCR and standard molecular biology techniques as described previously (9). Recombinant vaccinia viruses were generated using the various pSP11 CD4 chimeric constructs as described previously (9, 10). The recombinant vaccinia viruses encoding PLC-γ1 and PLC-γ2 were generously provided by A. M. Scharenberg (University of Washington, Seattle, WA) and J.-P. Kinet (Harvard, Boston, MA).
Ca2+ mobilization assays
Changes in the levels of intracellular Ca2+ were assessed in Indo-1-loaded cells by flow cytometry as described previously (11). Briefly, J.gamma 1 cells or human NK cells were infected with the indicated nonrecombinant (WR) or recombinant vaccinia viruses at a multiplicity of infection of 10:1 for 3 h (J.gamma 1) or 20:1 for 5 h (NK) in a humidified 37°C incubator. For the last 30 min of the infection, the cells were loaded with 5 μM Indo-1 (Calbiochem-Novabiochem). After washing, the cells were stimulated with a combination of anti-CD4 mAb and goat anti-mouse IgG F(ab′)2. The samples were analyzed immediately by flow cytometry using a UV laser for excitation with violet (390 nM) and blue (500 nM) fluorescence emissions recorded.
The 51Cr release assays were done as described previously (8). Before cytotoxicity assays, Some CTL clones were removed from the cloning medium and incubated in RPMI 1640 with 10% human serum, 1% l-glutamine, 1% sodium pyruvate, and 40–100 U/ml IL-2 at 37°C for 72 h. In redirected cytotoxicity assays, NK clones or CTL clones killed the FcR+P815 target cells only in the presence of the designated triggering mAb. We calculated lytic units per 106 cells on the basis of 20% cytotoxicity.
Cell stimulation and immunoblot analysis
In experiments involving vaccinia infection, NK clones or J.gamma 1 cells were infected with control vaccinia (WR) or recombinant vaccinia at a multiplicity of infection of 20:1 for 5 h (NK clones) or 10:1 for 3 h (J.gamma 1 cells). In experiments involving specific cell surface receptor cross-linking, cells were incubated for 3 min on ice with the indicated mAb. Washed cells were then incubated with goat anti-mouse IgG F(ab′)2 at 37°C for the indicated period of time. After stimulation, the cells were lysed on ice for 10 min in 1 ml of buffer containing 20 mM Tris-HCl, 40 mM NaCl, 5 mM EDTA, 50 mM NaF, 30 mM Na4P2O7, 0.1% BSA, 1 mM Na3VO4, 1 mM PMSF, 5 μg/ml aprotinin, 10 μg/ml leupeptin, and 1% Triton X-100. Cellular debris was removed by centrifugation at 20,800 × g for 5 min at 4°C. Immunoprecipitations were done from cell lysates at 4°C using a combination of PLC-γ1 and PLC-γ2 antisera bound to protein A-Sepharose beads. Protein complexes were then eluted in 40 μl of SDS sample buffer, resolved by SDS-PAGE, and transferred to Immobilon-P membranes (Millipore). Tyrosine-phosphorylated proteins were detected using the anti-pTyr mAb 4G10 (Upstate Cell Signaling Solutions), followed by goat anti-mouse IgG coupled to HRP (Santa Cruz Biotechnology) and the SuperSignal detection system (Pierce Biotechnology). PLC-γ1 and PLC-γ2 were detected using a combination of PLC-γ1 and PLC-γ2 antisera, followed by protein A coupled to HRP (Amersham Biosciences) and the SuperSignal detection system.
Tyrosine phosphorylation of PLC-γ after stimulation of FcR vs NKG2D in NK cells
NK cells can initiate ADCC after FcR ligation and natural cytotoxicity after NKG2D ligation (1). PLC-dependent calcium signaling in NK cells is required for granule exocytosis and killing, and previous results have highlighted the role of the tyrosine phosphorylation of PLC-γ in its activation (7, 9). Because both PLC-γ1 and PLC-γ2 are expressed in NK cells, we first evaluated which endogenously expressed isoforms are tyrosine phosphorylated during the development of each mode of cytotoxicity. For these studies, we used homogeneous, cloned populations of nontransformed human NK cells. The NK cells were stimulated with cross-linked anti-FcγRIII mAb, and then PLC-γ1 and PLC-γ2 were immunoprecipitated with their respective antiserum. The immunoprecipitates were separated by SDS-PAGE, transferred, and immunoblotted with the antiphosphotyrosine mAb, 4G10. Consistent with our previous analyses (7), selective FcR ligation on these NK cells induced the tyrosine phosphorylation of both PLC-γ1 and PLC-γ2 (Fig. 1 A). The proportion of FcR-induced tyrosine-phosphorylated PLC-γ1 vs tyrosine-phosphorylated PLC-γ2 was proportional to the amounts of total PLC-γ1 vs total PLC-γ2. In contrast, selective NKG2D cross-linking resulted predominantly in a rapid increase in the tyrosine phosphorylation of PLC-γ2 alone, being detectable within 1 min, remaining elevated for at least 10 min, and returning to basal level by 20 min. Although no detectable tyrosine-phosphorylated PLC-γ1 was observed, the sensitivity of this approach was limited by the following: 1) the weaker signal induced by NKG2D cross-linking relative to the signal seen after FcR stimulation; and 2) the significantly higher proportion of PLC-γ2 in NK cells as compared with PLC-γ1.
Tyrosine phosphorylation of PLC-γ after stimulation of ITAM-containing receptors vs DAP10-containing receptors
Because of the limited sensitivity for analyzing differential tyrosine phosphorylation of PLC-γ1 vs PLC-γ2 after stimulation of endogenously expressed receptors in NK cells, we developed an experimental model that 1) would allow us to control the strength of the signal and 2) allow equivalent amounts of PLC-γ1 and PLC-γ2 to be accessible as substrates for tyrosine phosphorylation. To control signal strength, we generated recombinant vaccinia viruses that express Flag-tagged CD4 chimeric receptors with either ITAM-containing cytoplasmic tails from zeta (F.CD4-zeta2) or gamma (F.CD4-gamma), or DAP10-containing cytoplasmic tails (F.CD4-gamma DAP10) (Fig. 1,B). We have shown previously that when these chimeric constructs are expressed in human NK cells, they will act as triggering receptors to induce killing after receptor ligation (9). To develop a cell model where we could have equivalent amounts of PLC-γ1 and PLC-γ2, we used the PLC-γ-deficient subline of Jurkat (J.gamma 1) (12) in which we expressed PLC-γ1 and PLC-γ2 using recombinant vaccinia virus. Ligation of the F.CD4-zeta chimeric rapidly induced the approximately equal tyrosine phosphorylation of PLC-γ1 and PLC-γ2 (Fig. 1, C and D). This finding is consistent with the observation that ITAM-containing receptors function effectively in cells predominantly expressing either PLC-γ1 (e.g., stimulating the TCR on T cells) or PLC-γ2 (e.g., stimulating the BCR on B cells). In contrast, stimulation of the F.CD4-DAP10 chimeric receptors induced a slower tyrosine phosphorylation of the PLC-γ isoforms, and although there was detectable tyrosine phosphorylation of PLC-γ1, there was significantly greater tyrosine phosphorylation of PLC-γ2 (Fig. 1, C and D). Taken together, these results suggested that whereas ITAM-containing receptors can effectively induce the tyrosine phosphorylation of either PLC-γ1 or PLC-γ2, the DAP10 signal-transducing subunit preferentially tyrosine phosphorylates PLC-γ2. However, both PLC-γ1 and PLC-γ2 have multiple potential tyrosine phosphorylation sites. Thus, it remained unclear whether quantitative differences in global tyrosine phosphorylation of the two isoforms would lead to qualitative differences in the DAP10-mediated calcium flux.
Influence of PLC-γ use on calcium signaling after stimulation of either ITAM-containing receptors or DAP10-containing receptors
It was shown previously that a calcium signal is necessary for NKG2D-mediated cytotoxicity (9). To assess whether the quantitative differences seen in the tyrosine phosphorylation of PLC-γ1 vs PLC-γ2 after DAP10 stimulation result in qualitative differences in calcium signaling, we again used the experimental model where the PLC-γ-deficient J.gamma 1 cells expressing either chimeric receptor was reconstituted with either PLC-γ1 or PLC-γ2. The cells were then loaded with Indo-1, and calcium fluxes were measured by flow cytometry. Stimulation of F.CD4-zeta-expressing cells infected with the control WR virus (instead of recombinant virus encoding either PLC-γ1 or PLC-γ2) showed a reduced but detectable calcium signal (Fig. 2,A). This result is consistent with the previous study (12), indicating that the small amount of residual PLC-γ2 in J.gamma 1 cells can induce a suboptimal calcium signal. Reconstitution of the J.gamma 1 cells with either PLC-γ1 or PLC-γ2 resulted in equivalent enhancement of the F.CD4 zeta-induced calcium signal. In contrast, F.CD4-DAP10 stimulation induced a delayed calcium signal (consistent with the delayed kinetics of F.CD4-DAP10-induced PLC-γ tyrosine phosphorylation as seen in Fig. 1 C). The small amount of PLC-γ2 in J.gamma 1 cells enabled the F.CD4-DAP10 chimeric to initiate a suboptimal calcium signal. However, whereas increased expression of PLC-γ2 markedly enhanced the calcium signal, PLC-γ1 did not.
We next tested whether this pattern of PLC-γ use would also be seen in primary NK cells. Because calcium signals are much more difficult to detect in nontransformed cells than in transformed cells like Jurkat, we overexpressed the chimeric F.CD4-gamma or F.CD4-DAP10 receptors with either PLC-γ1 or PLC-γ2 in human NK cell clones (Fig. 2 B). Stimulation through either chimeric receptor triggered a calcium signal in NK cells that were infected with the control vaccinia, indicating that endogenous levels of PLC-γ were sufficient to induce a calcium signal, albeit at a much lower level than Jurkat. As with Jurkat, both PLC-γ1 and PLC-γ2 overexpression enhanced the F.CD4-gamma-initiated signal, whereas only PLC-γ2 overexpression was able to enhance the F.CD4-DAP10-inititated signal. These results suggest that there are significant qualitative differences in the way that DAP10 couples to PLC-γ1 vs PLC-γ2.
Regulation of NK cell-mediated cytotoxicity by the different isoforms of PLC-γ
To determine whether the qualitative differences in DAP10-induced calcium signals resulting from the selective use of the PLC-γ isoforms had a functional consequence, we assessed the influence of the PLC-γ isoforms on NK cell-mediated killing. We infected an NK clone with either the control WR vaccinia, recombinant vaccinia-expressing PLC-γ1, or recombinant vaccinia-expressing PLC-γ2 (Fig. 3,A). These effectors were then tested for their ability to mediate either FcR-initiated or NKG2D-initiated killing. As shown in Fig. 3,A, FcR-initiated killing was amplified after the overexpression of either PLC-γ1 or PLC-γ2. This result suggests that both PLC-γ1 and PLC-γ2 are capable of coupling to the ITAM-containing FcR during the development of ADCC. In contrast, NKG2D-initiated cytotoxicity was amplified only after the overexpression of PLC-γ2, not PLC-γ1. To assess whether this result was broadly representative, we tested multiple NK clones in independent experiments (Fig. 3 B). Overexpression of PLC-γ2 enhanced both FcR-initiated killing (p < 0.002) and NKG2D-initiated killing (p < 0.0001), whereas PLC-γ1 overexpression only amplified FcR-initiated killing (p < 0.002) but not NKG2D-initiated killing (p = 0.54). These results suggest that the differences in calcium signaling based on PLC-γ isoform use noted before result in direct functional changes in the generation of NK cell-mediated cytotoxicity, and that whereas either PLC-γ isoform can couple to ITAM-containing receptors, only PLC-γ2 (but not PLC-γ1) can couple to DAP10.
Regulation of NKG2D-initiated killing in CTL
Although human CD8+ CTLs express NKG2D, they often are unable to mediate cellular cytotoxicity when stimulated through NKG2D alone (13, 14). Because CTLs express predominantly PLC-γ1, and because PLC-γ2 selectively couples to NKG2D-DAP10 in NK cells, we evaluated whether expression of PLC-γ2 in CTLs could render them competent to mediate NKG2D-initiated killing. We screened multiple human clonal CTL lines in a series of independent experiments. Six of 18 CTL lines were unable to perform significant levels of NKG2D-initiated killing, even after overexpressing PLC-γ1 or PLC-γ2. This suggests that in a portion of the CTLs, the inability to mediate NKG2D-initiated cytotoxicity is not limited to the absence of PLC-γ2. Interestingly, 12 of 18 CTL lines were able to mediate NKG2D-initiated killing after expressing PLC-γ2 (but not after overexpressing PLC-γ1). A representative example is shown in Fig. 4,A, and data from all of the experiments are shown in Fig. 4 B. This finding suggests that 1) subpopulations of lymphocytes may differ in their receptor-initiated functions based on alternative expression patterns of PLC-γ isoforms; and 2) if cellular activation of CTLs altered the expression of PLC-γ2, it could alter the capacity of the CTLs to mediate NKG2D-initiated killing.
ITAM-containing multisubunit receptors are expressed in several types of immune cells (e.g., TCR in T cells, BCR in B cells, and FcR in NK cells and mast cells). Some cells predominantly express PLC-γ1 (e.g., T cells), and others predominantly express PLC-γ2 (e.g., B cells and mast cells) (4). Because ITAM-containing receptors signal equally well in these cells regardless of the PLC-γ isoform present, the assumption has been that PLC-γ1 and PLC-γ2 are interchangeable and that all receptors can use either isoform. Knockout data have done little to clarify the issue. PLC-γ1 knockout mice are embryonic lethal (4). Although PLC-γ2 knockout mice have defective B cells, mast cells, and NK cells but normal T cells (4, 5, 6), these differences can be explained by the absence of the predominantly expressed PLC-γ isoform, neither supporting nor refuting the hypothesis of PLC-γ isoform redundancy.
Our data show that whereas both PLC-γ1 and PLC-γ2 can couple to some activating receptors, other triggering receptors selectively use one PLC-γ isoform. Both PLC-γ1 and PLC-γ2 enhanced ITAM-triggered functions. The ITAM-containing chimeric receptor F.CD4-zeta triggered equivalent Ca+2 release in a PLC-γ-deficient subline of Jurkat (J.gamma 1) that transiently expressed either PLC-γ1 or PLC-γ2. When stimulated through the ITAM-associated FcR, both PLC-γ1 and PLC-γ2 overexpression equivalently enhanced NK cytotoxicity. In contrast, only PLC-γ2 enhanced DAP10-triggered functions. J.gamma 1 cells could only trigger Ca+2 release with F.CD4-DAP10 when expressing PLC-γ2, and NKG2D-DAP10-stimulated cytotoxicity was only enhanced with PLC-γ2 overexpression.
The phosphorylation data presented in this study indicates that the selective use of PLC-γ2 by NKG2D-DAP10 occurs proximally in the signaling pathway, but additional studies will be needed to define the proximal regulatory events that enable this selective coupling. For example, because ITAM-containing receptors require the binding and activation of Syk family tyrosine kinases (15), perhaps the ability of NKG2D-DAP10 to recruit p85 and other potential molecules in a Syk-independent manner alters critical components in the signaling complex that are needed to bind and/or activate the different PLC-γ isoforms (3, 9, 16). In addition, whereas ITAM-containing receptors can drive lymphocyte activation in a Vav1-independent manner, the requirement for Vav1 in NKG2D-initiated signaling could influence the availability of Vav1 binding partners that differentially couple to the different PLC-γ isoforms (17, 18). Finally, whereas T cell adapters like linker of activated T cells are required for TCR-driven T cell activation, the linker of activated T cells-independent nature of NKG2D-initiated regulation could suggest the use of a separate adapter (9).
Although NKG2D is fully activating in NK cells, it acts only as a costimulatory receptor in normal CTLs (13, 14). We showed that in a fraction of normal CTLs, PLC-γ2 overexpression (but not PLC-γ1 overexpression) was sufficient to transform the NKG2D signal from a costimulatory to a fully cytotoxic response. Thus, PLC-γ isoform expression may provide lymphocytes an added level of control in determining how to respond to NKG2D stimuli. Alternatively, the expression of PLC-γ2 in CTLs may represent a step in the pathology of chronic inflammatory diseases. Recent studies have shown that chronically activated CTLs such as those found in patients with celiac disease can trigger cytotoxic responses when stimulated through NKG2D (14, 19). It will be interesting to see whether these abnormal CTLs have altered expression of the PLC-γ isoforms. Each of these signaling elements that differentiates one mode of lymphocyte activation from another represents a potential target for selective therapeutic manipulations.
We thank Andrew M. Scharenberg and Jean-Pierre Kinet for the recombinant vaccinia encoding PLC-γ1 and PLC-γ2; Robert T. Abraham for the PLC-deficient subline of Jurkat; and Theresa Lee for her assistance with the preparation of this manuscript.
The authors have no financial conflict of interest.
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 research was supported by the Mayo Foundation and by National Institutes of Health Grant CA47752. D.D.B. is a recipient of the Cancer Research Institute Investigator Award.
Abbreviations used in this paper: PLC, phospholipase C; ADCC, Ab-dependent cellular cytotoxicity.