NKp44 Triggers NK Cell Activation through DAP12 Association That Is Not Influenced by a Putative Cytoplasmic Inhibitory Sequence

Cell lines and Abs

NK-92 cells were grown in α-MEM supplemented with IL-2 as previously described (11). Phoenix-Amphotropic retroviral packaging cells were kindly provided by Garry Nolan (Stanford University, Stanford, CA) and grown as described (11). Jurkat cells were grown in RPMI 1640 medium supplemented with 10% FBS. DX9 (anti-KIR3DL1), 3.43.13 (anti-NKp44), and B159 (anti-CD56) mAbs were purified with protein G from hybridomas obtained from Drs. L. Lanier (University of California, San Francisco, CA), M. Colonna (Washington University, St. Louis), and B. Perussia (Thomas Jefferson University, Philadelphia, PA), respectively. PE-conjugated DX9 was purchased from BD PharMingen (San Diego, CA). The anti-FLAG mAb, M2, was purchased from Sigma-Aldrich (St. Louis, MO). Rabbit anti-DAP12 Ab was produced toward a recombinant form of the human DAP12 cytoplasmic domain and purified on protein A by Exalpha Biologicals (Boston, MA). PE-conjugated goat anti-mouse κ light chain Ab was obtained from Southern Biotechnology Associates (Birmingham, AL).

Cloning of cDNA constructs

cDNA constructs were prepared by PCR from a NK cell cDNA library using the synthetic oligonucleotides shown in Table I⇓ to generate wild-type, chimerized, or mutated NKp44 and KIR3DL1 receptor cDNAs with engineered endonuclease restriction sites, as marked in Fig. 2⇓A. Full-length or truncated NKp44 constructs were engineered with the NKp44 leader, the FLAG-epitope tag sequence, and the NKp44 protein sequence. Transition amino acid sequences between leader, FLAG, and amino terminus of full-length NKp44 is F_–6PGSQADYKDDDDKLQSKAQV₆ (FLAG sequence underlined, outside residues numbered (2), and italic leucine encoded by HindIII restriction site). The cDNA for 3DL1.p44 chimeric receptors consisted of the KIR3DL1 sequence until methionine-373 in the cytoplasmic domain, a BspHI site (as described in Ref. 11), and the cytoplasmic domain of NKp44 from leucine-226 (Fig. 2⇓A). Transition between KIR3DL1 and NKp44 is K₃₆₇KNAAVMNLRSL₂₀₈ (KIR3DL1 in bold, NKp44 underlined, asparagine encoded by BspHI site in italics, and representative amino acids in each protein numbered). Tyrosine-238 in the cytoplasmic domain of NKp44 was mutated to phenylalanine (Y238F constructs; numbering as in Ref. 2). The cDNA constructs were ligated into the pBMN-IRES-EGFP plasmid (from G. Nolan) or a modified version lacking the internal ribosome entry site (IRES) and enhanced green fluorescent protein (EGFP) sequences (pBMN-NoEGFP) as previously described (11). The KIR3DL1 construct was previously described (11). The cDNAs of DAP12/KARAP, DNAX-activation protein of 10 kDa (DAP10; also known known as PI-3 kinase-associated protein of 10 kDa (KAP10)), and FcεRIγ accessory proteins were cloned into pBMN-IRES-EGFP using the oligos listed in Table I⇓. DAP12/KARAP and FcεRIγ were cloned from a cDNA library, and DAP10/KAP10 was cloned from a His-tagged cDNA construct in the pJFE14 vector that was kindly provided by John Trowsdale and Julie Fairbanks (Cambridge University, Cambridge, U.K.). All PCR were performed using Platinum Pfx DNA polymerase (Life Technologies, Rockville, MD), and the integrity of all constructs was confirmed by automated sequencing in the Fox Chase Cancer Center (FCCC) DNA Sequencing Facility.

Retroviral transduction

Retrovirus was prepared as previously described (11). Briefly, the Phoenix packaging line was transfected with pBMN vectors, and culture supernatant was harvested 2 days later. NK-92 cells were incubated with viral supernatant for 8 h with Lipofectamine Plus (Life Technologies), and cells expressing EGFP and/or surface receptor were sorted 3 days later on a FACSVantage^SE (BD Biosciences, Mountain View, CA) in the FCCC Cell Sorting Facility. Surface expression of transduced receptors was stable for at least 2 months. To cotransduce NKp44 and accessory proteins into Jurkat cells, the Jurkat parent was transduced with retrovirus containing cDNA for either DAP12/KARAP, DAP10/KAP10, or FcεRIγ in pBMN-IRES-EGFP. Transduced cells were sorted for EGFP 5 days later and subsequently transduced with retrovirus containing the FLAG-NKp44 cDNA in the pBMN-NoEGFP vector. The entire cell population was analyzed 4 days later by staining for surface expression with anti-FLAG mAb.

Redirected cytotoxicity assay

Receptors transduced into NK-92 cells were tested for impacts on killing in a redirected cytotoxicity assay using the FcγR⁺ P815 murine mastocytoma target cell line as previously described (11). NK-92 cells were either resting (4 days after IL-2 addition) or stimulated (1 day after IL-2) to test activating or inhibitory properties of the full-length or chimeric receptors, respectively. Cytotoxicity was measured in triplicate determinations using a standard 4-h ⁵¹Cr release assay with or without Abs (5 μg/ml) to engage the KIR3DL1 extracellular domain (DX9), FLAG-tagged NKp44 (M2), or the control receptor CD56 (B159).

Cell surface activation marker expression assay

Six-well plates were coated with mAbs (10 μg/ml HBSS overnight at 4°C or 2 h at 37°C) to engage receptors, washed, and seeded with 2 × 10⁶ resting NK-92 cells (day 4 after IL-2) in 5 ml of complete medium. Cells were stimulated for 18 h before harvest by harsh pipetting, washed on ice, and labeled with PE-conjugated Abs to CD69, CD25, or CD95L (BD PharMingen) in the presence of azide. Cells were analyzed on a FACScan analyzer (BD Biosciences).

IFN-γ assay

Receptor-transduced NK-92 cells (2 × 10⁶) were stimulated with Abs coated on six-well plates (5 ml/well) as described above for 18 h. Supernatants were harvested and assayed for IFN-γ production using an ELISA kit (BD PharMingen).

Metabolic labeling, immunoprecipitation, and immunoblotting

For metabolic labeling, NK-92 cells (160 million) were washed twice in HBSS and cultured for 75 min in 5 ml of cysteine/methionine-free DMEM (Sigma-Aldrich) containing 5 mCi [³⁵S]cysteine/methionine (Promix; Amersham Pharmacia Biotech, Piscataway, NJ). Labeled cells were washed once with HBSS. Some cell samples were treated for 10 min with pervanadate (100 μM sodium orthovanadate plus 10 mM H₂O₂). Cells were lysed for 30 min on ice in 1 ml of buffer containing 1% of either Triton X-100, digitonin, Brij 58 (polyoxyethylene 20 cetyl ether) (all from Sigma-Aldrich), sodium n-dodecyl-β-d-maltoside, or n-octyl-β-d-glucopyranoside (octyl glucoside) (both from Calbiochem, La Jolla, CA) as previously described (11). Detergent-insoluble debris was removed from lysates by centrifugation at 20,800 × g for 15 min at 4°C. Lysates were then precleared twice for 30 min each at 4°C with protein G-coupled agarose (Upstate Biotechnology, Lake Placid, NY). FLAG-tagged NKp44 was immunoprecipitated for 90 min at 4°C with 30 μl of M2 mAb-Sepharose beads (1 mg mAb/ml cyanogen bromide-conjugated Sepharose 4 Fast Flow; Amersham Pharmacia Biotech), separated on SDS-PAGE, transferred to polyvinylidene difluoride (PVDF) membrane (Bio-Rad, Hercules, CA), and probed with rabbit anti-DAP12 Ab (0.25 μg/ml), rabbit anti-SHP-1 Ab (1 μg/ml; Upstate Biotechnology), rabbit anti-SHP-2 Ab (1 μg/ml; Santa Cruz Biotechnology, Santa Cruz, CA), or rabbit anti-SHIP (1 μg/ml; Santa Cruz Biotechnology) plus HRP-conjugated protein G (Calbiochem), anti-FLAG mAb (M2) plus HRP-conjugated donkey anti-mouse IgG (Jackson ImmunoResearch, West Grove, PA), or HRP-conjugated anti-phosphotyrosine mAb (4G10; Upstate Biotechnology). Immunoblotted proteins were visualized by chemiluminescence using the ECL detection reagents (Amersham, Arlington Heights, IL).

Yeast two-hybrid system

The yeast two-hybrid vectors pPC62 and pPC86 were obtained from P. Chevray and D. Nathans (Johns Hopkins University, Baltimore, MD) (12), and use of the yeast two-hybrid system was described previously (13, 14). Receptor cytoplasmic domain bait constructs for human KIR2DL1 (aa 265–348), human ICAM3 (CD50; aa 482–518), and human NKp44 (aa 220–276) were generated by PCR and ligated into a modified version of the pPC86 plasmid (named pPC87) in which the Gal4 DNA binding domain fusion cassette from pPC62 was introduced to replace the transcriptional activation cassette. In this way, pPC87 produced a Gal4 DNA binding fusion protein in a plasmid under TRP1 selection, which is compatible with the pACT vector encoding the Gal4 transcriptional activation domain fused with aa 3–436 of the μ2 (AP50) adaptor protein under LEU2 selection (construct kindly provided by Chander Raman, University of Alabama, Birmingham, AL (15)). A construct encoding the Gal4 DNA binding domain fused to the cytoplasmic domain of murine CD5 (aa 378–471) in pGBT9 plasmid (TRP1 selection; Clontech, Palo Alto, CA) was kindly provided by C. Raman (15). The reporter yeast strain was AH109 (BD Biosciences), which contained two integrated growth selection reporters (His3 and Ade2). Both reporters contain upstream LexA operators that are each believed to bind two LexA hybrid proteins. Yeast culture conditions, transfection protocol, and chemiluminescence assay for β-galactosidase were performed as previously described (14).

DAP12 association with NKp44 is required for receptor surface expression

The DAP12/KARAP accessory protein was previously reported to co-immunoprecipitate with NKp44 from digitonin lysates of NK cells (1, 2). We tested this association in NK-92 cells that had been metabolically labeled with [³⁵S]cysteine/methionine and lysed with digitonin (1, 2). The receptor complex was immunoprecipitated with a mouse anti-NKp44 mAb, 3.43.13 (kindly provided by Marco Colonna, Washington University, St. Louis). We confirmed that the mAb can immunoprecipitate NKp44 in association with DAP12, but our experiments showed only weak co-immunopreciptation of DAP12 from digitonin lysates as assessed by immunoblotting (Fig. 1⇓). The results prompted us to further examine the capacities of other mild nonionic detergents to optimally release NKp44 and retain the DAP12 association. When directly compared, dodecyl maltoside proved to be a significantly better detergent than digitonin for this purpose (Fig. 1⇓). In contrast, octyl glucoside only weakly solubilized the receptor and Brij 58 did not detectably release NKp44 from the plasma membrane (Fig. 1⇓). Therefore, we confirmed that DAP12 is associated with NKp44 in NK cells, and we conclude that dodecyl maltoside is a substantially better detergent for solublizing the intact receptor complex from NK cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Dodecyl maltoside detergent optimally releases DAP12-associated NKp44. NK-92 cells (45 million/sample) were metabolically labeled with [³⁵S]cysteine/methionine, and aliquots were lysed in buffers containing 1% of the indicated detergents. NKp44 was immunoprecipitated from cleared lysates, separated on 15% SDS-PAGE under reducing conditions, transferred to PVDF membrane, autoradiographed (top), and immunoblotted with anti-DAP12 Abs (bottom). DAP12 was not visible in the autoradiograph, presumably due to slow protein turnover.

We next addressed the molecular basis of DAP12 association and tested whether DAP12 association was required for surface expression of NKp44. Recombinant retrovirus engineered to co-express a FLAG-epitope tagged form of wild-type NKp44 (NKp44.wt; Fig. 2⇓A) and EGFP was used to transduce the DAP12-expressing NK-like cell line, NK-92, and the DAP12-deficient T cell line, Jurkat. EGFP-expressing transduced cells were sorted and stained with anti-FLAG mAb to detect surface expression of NKp44.wt. Although NKp44.wt was strongly expressed on transduced NK-92 cells, the DAP12-deficient Jurkat T cell line could not support surface expression of NKp44 (Fig. 3⇓A). Next, we tested whether DAP12 could rescue surface expression of NKp44 in Jurkat cells by co-expressing NKp44 in combination with either DAP12, DAP10, or FcεRIγ accessory proteins. As shown in Table II⇓, co-expression of DAP12 supported surface expression of NKp44, while DAP10 or FcεRIγ did not. Parental Jurkat cells endogenously express TCR-ζ (data not shown), which indicates that TCR-ζ is also incapable of promoting surface expression of NKp44. We further mutated the transmembrane lysine residue in NKp44 (K183) to alanine to test whether it is required for DAP12 association and receptor surface expression. As shown in Fig. 3⇓B, the K183A mutant was only marginally expressed on the surface of NK-92 cells, despite co-expression of DAP12. Furthermore, the NKp44.K183A mutant did not associate with DAP12 in dodecyl maltoside lysates, while the endogenous NKp44 (subsequently immunoprecipitated from the same lysates with the anti-NKp44 mAb, 3.43.13) was strongly associated with the accessory protein, as shown in Fig. 3⇓C. Therefore, DAP12 association is required for cell surface expression of the receptor, and association is dependent upon integrity of lysine-183 of NKp44.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

NKp44 constructs and their expression in the NK-92 cell line. A, Schematic diagram of protein products from cDNAs encoding FLAG-tagged and chimeric NKp44 receptors. Domains labeled include FLAG epitope tag, extracellular KIR Ig domains (D0, D1, D2), transmembrane domain (TM), and cytoplasmic domain (CY). The transmembrane lysine residue is marked (K or mutant A), as well as the tyrosine-based ITIM or ITIM-like sequences in the cytoplasmic domains. Numbered arrows represent locations of oligonucleotide primers (corresponding to those listed in Table I⇑) used to synthesize the cDNAs. B, The NK-92 cell line was transduced with retroviral preparations encoding the FLAG-tagged or chimeric NKp44 receptors, and receptor-expressing cells were sorted by flow cytometry. Surface expression (solid histograms) was determined with either anti-FLAG mAb (M2 + PE-anti-mouse κ) or anti-KIR3DL1 mAb (PE-DX9) compared with unstained or secondary alone (dashed histograms).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Surface expression of NKp44 requires association with DAP12 via the transmembrane lysine residue. A, NKp44 is not expressed on the surface of the DAP12-deficient T cell line, Jurkat. Jurkat T cells or NK-92 cells were transduced with FLAG-NKp44.wt, and EGFP-positive transduced cells were sorted and stained for surface expression with the anti-FLAG mAb (M2). B, Mutation of the transmembrane lysine on NKp44 to alanine significantly disrupts surface expression in NK-92 cells. NK-92 cells were transduced with a FLAG-tagged form of NKp44 in which the transmembrane lysine was mutated to alanine (K183A) or the empty vector (pBMN-IRES-EGFP), EGFP⁺ cells were sorted, and these were stained with anti-FLAG mAb. C, The K183A mutant NKp44 does not associate with DAP12 in NK-92 cells. Transduced NK-92 cells shown in B were lysed in dodecyl maltoside (1%) detergent, and lysates were sequentially immunoprecipitated with anti-CD56, anti-FLAG, and anti-NKp44 mAbs. Immunoprecipitates were separated on 15% SDS-PAGE, transferred to PVDF membrane, and sequentially immunoblotted with anti-FLAG and anti-DAP12 Abs. Note that some FLAG-tagged NKp44 was still isolated in the final immunoprecipitation with anti-NKp44 (lane 6), but the endogenous receptor in that immunoprecipitate was associated with DAP12.

The tyrosine-based sequence in NKp44 does not function as an ITIM

We next addressed whether the ITIM-like sequence in the NKp44 cytoplasmic domain influences activating function of the receptor complex. To test the inhibitory capacity of the cytoplasmic domain, we generated a chimeric receptor in which the cytoplasmic domain, including the ITIM-like sequence of NKp44, was directly fused to the extracellular and transmembrane domains of KIR3DL1 to create a chimeric receptor, which was designated 3DL1.p44.wt (schematized in Fig. 2⇑A). As a negative control, a 3DL1.p44 chimera was also produced in which the ITIM-like tyrosine of NKp44 (tyrosine-238) was mutated to phenylalanine (designated 3DL1.p44.Y238F). We have previously validated this chimeric killer cell Ig-like receptor (KIR) strategy as a powerful technique to demonstrate the strong inhibitory properties of the cytoplasmic domain of KIR2DL4 (11). The 3DL1.p44 chimeric receptors were expressed at similar levels on the surface of transduced NK-92 cells, as assessed by anti-KIR mAb staining (Fig. 2⇑B).

As in our previous studies (11, 16), the inhibitory capacities of the 3DL1.p44 chimeric receptors were tested on IL-2-stimulated NK-92 cells in “redirected” cytotoxicity assays toward the FcγR⁺ P815 target cell, using the DX9 mAb as a surrogate ligand to specifically engage the chimeric receptors. DX9-mediated engagement of the wild-type KIR3DL1 receptor in this assay induced potent inhibition of cytotoxicity (Fig. 4⇓A). In contrast, engagement of either 3DL1.p44 chimeric receptor did not influence the cytolytic response when compared with control cultures treated with anti-CD56 mAb or lacking added Ab (Fig. 4⇓A). Inhibition was not evident even when NK-92 cells were tested under resting conditions (4 days after low-dose IL-2; Fig. 4⇓A). These results indicated that the tyrosine-based motif on NKp44 is not functioning as an ITIM.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

The cytoplasmic tyrosine-based sequence of NKp44 does not affect the killing of P815 target cells or bind SH2-containing protein tyrosine phosphatases. A, The cytoplasmic domain of NKp44 does not inhibit redirected killing. KIR3DL1 or KIR3DL1.p44 chimeric receptors (wild-type or Y238F mutant) were expressed in NK-92 cells and engaged with anti-KIR3DL1 Ab (DX9; ▵) during cytotoxicity of the FcγR⁺ P815 target cell at various effector-to-target (E:T) ratios. Control assays were performed in the absence of mAb (○) or with anti-CD56 mAb (B159; •). NK-92 cells were stimulated with IL-2 on the day before assay to enhance their cytolytic capacity toward P815. The lower panel is an assay of KIR3DL1.p44.wt-transduced cells stimulated 4 days after low-dose IL-2 stimulation (half of normal amount). Percent of maximal cytotoxicity is indicated. B, The cytoplasmic tyrosine of NKp44 can be tyrosine-phosphorylated, but does not recruit SHP-1 or SHP-2 phosphatases. NK-92 cells transduced with KIR3DL1.p44.wt or KIR3DL1.p44.Y238F were either untreated (−) or pervanadate-treated (+) for 10 min before lysis in 1% Triton X-100 buffer and immunoprecipitation with anti-KIR3DL1 mAb (DX9). Immunoprecipitates were separated on 10% SDS-PAGE, transferred to PVDF membranes, and probed with anti-phosphotyrosine (4G10-HRP), anti-SHP-1, anti-SHP-2, or anti-SHIP Abs plus protein G-HRP secondary. SHIP was never observed to bind phosphorylated KIR3DL1. p44.wt (data not shown).

We next performed biochemical analysis to test if the ITIM-like motif on 3DL1.p44.wt can be tyrosine-phosphorylated and if it can recruit any of the inhibitory SH2-containing phosphatases. Pervanadate stimulation of NK-92 cells resulted in strong tyrosine phosphorylation of KIR3DL1.p44 at the predicted molecular mass of ∼62 kDa (Fig. 4⇑B). In contrast, the Y238F mutant chimeric receptor was not tyrosine-phosphorylated (Fig. 4⇑B). Together, these results demonstrate that the ITIM-like sequence of NKp44 can be tyrosine-phosphorylated in the context of the chimeric receptor. We could not, however, detect co-immunoprecipitation of SHP-1, SHP-2, or SHIP with tyrosine-phosphorylated 3DL1.p44.wt (Fig. 4⇑B; SHIP data not shown). On the contrary, our analyses have routinely shown that inhibitory KIR, such as 3DL1, strongly bind SHP-1 and/or SHP-2 under these conditions (11, 16). Therefore, the cytoplasmic tyrosine from NKp44 can serve as a protein tyrosine kinase substrate, but this phosphorylated tyrosine cannot serve as a docking site for SH2-containing phosphatases that are known effectors of negative signaling.

The tyrosine-based sequence in NKp44 does not affect activation properties of the receptor

Next, we tested whether the tyrosine-based sequence on NKp44 influences the activating properties of the full-length receptor. To study its influence, we mutated the ITIM-like tyrosine (Y238) to phenylalanine on the FLAG-epitope tagged NKp44 (to generate NKp44.Y238F) as shown in Fig. 2⇑A. NKp44.wt and NKp44.Y238F were expressed at similar levels in NK-92 cells by retroviral transduction (Fig. 2⇑B) and compared for functional activation capacities.

We first compared the abilities of these FLAG-tagged forms of NKp44 to stimulate the up-regulation of several cell surface activation markers in resting NK-92 cells in response to engagement with plate-bound anti-FLAG mAb. The activation markers CD25 (IL-2 receptor α-chain), CD69, and CD95L (FasL) were all increased to similar extents by engagement of either NKp44.wt or NKp44.Y238F (Fig. 5⇓A). In contrast, engagement of the 3DL1.p44.wt chimera did not increase activation marker expression (Fig. 5⇓A), which demonstrates that the tyrosine-based motif does not contribute direct activating capacity to receptor function. To confirm that the tyrosine-based sequence does not influence activation potential under weak stimulating conditions, we performed a dose-response comparison. As shown in Fig. 5⇓B, similar CD69 up-regulatory capacities were also observed when both full-length NKp44 receptors were engaged with low doses of anti-FLAG mAb. Thus, the tyrosine-based sequence in NKp44 did not exhibit intrinsic activating properties, nor did it inhibit or potentiate the capacity of the receptor to trigger up-regulation of activation markers on NK-92 cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Up-regulation of activation markers by NKp44 engagement is not affected by the cytoplasmic tyrosine. A, Receptor-transduced NK-92 cells were cultured for 18 h in six-well plates (2 million/well) that had been precoated with anti-FLAG mAb (M2 for FLAG-NKp44 receptors) or mouse anti-3DL1 mAb (DX9 for 3DL1.p44.wt). Cells were harvested and stained with PE-conjugated Abs specific for CD25 or CD69 or with biotinylated anti-CD95L plus PE-conjugated streptavidin. Dashed lines are staining of unstimulated cells, and solid lines are staining of stimulated cells. B, CD69 expression in response to engagement of NKp44 with increasing concentrations of anti-FLAG mAb. NK-92 cells transduced with NKp44 receptors were cultured for 18 h in six-well plates (2 million/well) that had been precoated with anti-FLAG mAb (M2) at indicated concentrations. Cells were then harvested and stained with PE-conjugated anti-CD69 mAb.

We next tested whether the presence of the tyrosine-based sequence could influence cytolytic or cytokine release responses through NKp44. NKp44.wt and NKp44.Y238F stimulated nearly identical levels of killing response by resting NK-92 cells when engaged with anti-FLAG mAb in a redirected cytotoxicity assay (Fig. 6⇓A). Similarly, plate-bound anti-FLAG mAb stimulated similar levels of IFN-γ release when NK-92 cells expressed either of the receptors (Fig. 6⇓B). In conclusion, our results indicate that the tyrosine-based sequence in the cytoplasmic domain of NKp44 does not influence activating function of the full-length receptor in assays of activation marker induction, cytotoxicity, or IFN-γ release.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

Cytotoxicity or IFN-γ production responses stimulated by NKp44 engagement are not influenced by the cytoplasmic tyrosine. A, Full-length FLAG-tagged NKp44 receptors (wt or Y238F) on transduced NK-92 cells were engaged with anti-FLAG mAb (M2; ▵), anti-CD56 mAb (B159; •), or without added mAb (○) during a 4-h redirected cytotoxicity assay toward P815 target cells. NK-92 cells were assayed under resting conditions (day 4 after IL-2). B, NK-92 cells expressing FLAG-NKp44 receptors were stimulated for 18 h with plate-bound anti-FLAG mAb (▪) or anti-CD56 mAb (□). Supernatants were collected and assayed for IFN-γ production by ELISA.

The cytoplasmic domain of NKp44 does not associate with the AP-2 clathrin adaptor complex

Because signaling function of the ITIM-like sequence was not evident, we reasoned that the sequence could alternatively contribute to the regulation of NKp44 surface expression through physical association with the AP-2 clathrin adaptor complex, which targets cell surface receptors for endocytosis through clathrin-coated vesicles (17). The μ2 subunit of AP-2 is known to associate with Yxxφ sequences (φ = large hydrophobic or aromatic residue), which matches the ITIM-like sequence of NKp44 (18, 19, 20). Therefore, we used the yeast two-hybrid system to test whether the ITIM-like sequence of NKp44 acts as a tyrosine-based internalization motif through association with the μ2 subunit (also known as AP50) of the AP-2 complex. As shown in Fig. 7⇓, we did not detect any μ2 association with the cytoplasmic domains of NKp44, KIR2DL1, or ICAM3, while interaction with the cytoplasmic domain of CD5 was readily detectable, as previously reported (15). In accordance with this result, we did not detect differential internalization rates of NKp44.wt, NKp44.Y238F, and NKp44.trunc receptors in response to engagement with anti-FLAG mAb and secondary cross-linking Abs in NK-92 cells (data not shown). The truncated receptor (NKp44.trunc) was prepared, because AP-2 is also known to interact with phenylalanine motifs (Fxxφ) (21). Nonetheless, even the receptor lacking the majority of its cytoplasmic domain did not exhibit altered rates of internalization, recycling, or functional response (cytotoxicity or IFN-γ production). Taken together, our experiments indicate that the ITIM-like motif of NKp44 does not contribute to receptor signaling or internalization functions of the NKp44/DAP12 receptor complex.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

The AP-2 clathrin adaptor component, AP50 (μ₂), does not interact with the NKp44 cytoplasmic domain in the yeast two-hybrid system. The cytoplasmic domains of NKp44, KIR2DL1, ICAM3, and CD5 were each fused to the Gal4 DNA binding domain (DB) and expressed in a reporter yeast strain (AH109) in combination with the Gal4 transcriptional activation domain (TA) either alone (empty) or directly fused with the μ2 subunit of AP-2, AP50. Transfected cells were streaked on agarose plates (lettered quadrants) under growth reporter selection conditions lacking histidine and adenine (-His, -Ade) or under nonrestrictive conditions in the presence of those amino acids (+His, +Ade). Growth under selecting conditions indicates a direct interaction between cytoplasmic domain and AP50, which was evident for the positive control, CD5.