not be used as targets to engage 3DL1, because background inhibition
resulted from HLA-B*2702 binding by the CD85 inhibitory receptor on
NK-92 cells (data not shown).
Conjugate assay
P815 cells were cultured with fresh medium 1 day before assay. Also 1 day
before assay, NK-92 cells were restimulated with fresh medium containing
human IL-2, and the cell numbers of cultured NK-92 cells were adjusted to
8
ϫ 10
4
cells/ml in a 75-cm
2
flask, because cell density and its activation
status influence the optimal conjugation. P815 and NK-92 cells were in-
cubated for 30 min at 2 million/ml with prewarmed (37°C) medium con-
taining either CellTracker Orange CMTMR (Molecular Probes, Eugene,
OR) at 2
M or CellTracker Green CMFDA (Molecular Probes) at 0.2
M, respectively. The stained cells were washed, resuspended, incubated
with prewarmed medium for 30 min at 2 million/ml, and washed again.
P815 cells were incubated at room temperature with anti-KIR3DL1 mAb
(DX9; 2
g/ml). Conjugate formation was performed, as previously de-
scribed (20), using a FACScan analyzer (BD Biosciences).
Pervanadate treatment and cell lysis
NK cells were washed three times in HBSS (Life Technologies), stimulated
with pervanadate (100
M), and lysed for 30 min on ice in 1 ml of 1%
Triton X-100 lysis buffer per sample, as previously described (31). Lysates
were cleared by centrifugation at 20,800
ϫ g for 15 min at 4°C.
Immunoprecipitation and immunoblotting
Lysates were precleared twice for 30 min each at 4°C with protein G-
coupled agarose (Upstate Biotechnology, Lake Placid, NY). CD56 and 2B4
or KIR were sequentially immunoprecipitated for 90 min at 4°C with
B159.5.2 and anti-2B4 or DX9 mAbs (5
g/sample precoupled to 30 l of
protein G-agarose), as performed previously (31), separated on SDS-
PAGE, transferred to polyvinylidine difluoride membrane (Bio-Rad, Her-
cules, CA), and probed either directly with HRP-coupled 4G10 mAb (anti-
phosphotyrosine, 1/10,000; Upstate Biotechnology) or initially with rabbit
polyclonal anti-SHP-1 (1
g/ml; Upstate Biotechnology) or anti-SHP-2 (1
g/ml; Santa Cruz Biotechnology, Santa Cruz, CA) Abs, and secondarily
with HRP-coupled protein G (1/10,000; Calbiochem, La Jolla, CA). SAP
was detected by rabbit polyclonal anti-SAP Abs (1/1,000), and secondarily
with HRP-coupled protein G.
Vaccinia virus infections
Recombinant vaccinia virus preparations generated with the plasmid
pSC65 containing cDNA of wild-type SHP-1 or DN-SHP-1 (C453S) (9)
were kindly provided by E. Long (National Institutes of Health, Rockville,
MD). Generation of the recombinant vaccinia virus encoding DN-SHP-2
(C459S) cDNA was previously described (20, 39). Aliquots of the recom-
binant virus preparations were dispersed by water-bath sonication and
stored at
Ϫ70°C before use. The titer of viral stocks was determined by
plaque assay, as described (39) (wild-type SHP-1, 1
ϫ 10
9
PFU/ml; C453S
SHP-1, 0.25
ϫ 10
9
PFU/ml; C459S SHP-2, 1
ϫ 10
9
PFU/ml). NK-92 cells
were infected with the recombinant vaccinia virus, as described previously
(31, 40). Vaccinia virus infections were monitored for SHP-1 or SHP-2
protein expression by immunoblot analysis of the infected cells. Viability
of the infected cells was monitored by trypan blue staining before mixing
with target cells (
Ͼ95% viable).
Results
Inhibitory function of the cytoplasmic domain of 2DL5
To examine the potential inhibitory function of the 2DL5 cyto-
plasmic domain, we generated a chimeric receptor construct
(3DL1/L5) in which the cytoplasmic domain of 2DL5, including
the N-terminal ITIM- and C-terminal ITSM-like sequences, was
fused to the extracellular and transmembrane domains of 3DL1
(Fig. 1, A and C). We previously used this chimeric receptor ap-
proach to establish the inhibitory potential of the 2DL4 cytoplas-
mic domain with the 3DL1/L4 fusion construct (Fig. 1, A and C)
(31). These chimeric receptors allowed us to directly compare in-
hibitory properties of the 2DL5 cytoplasmic domain with those of
3DL1 and 2DL4 in the context of a common extracellular domain
(3DL1). A truncated receptor, in which we deleted the 3DL1 cy-
toplasmic domain from just before the N-terminal ITIM (3DL1/
272P
⌬), was used as a negative control (Fig. 1A). All of these
receptors were expressed on the surface at similar levels in retro-
virus-transduced NK-92 cells (Fig. 1B).
We first tested the inhibitory capacity of the transduced KIR
using a redirected cytotoxicity assay against the murine mastocy-
toma cell line, P815. P815 expresses Fc
␥RII/III that can interact
with the Fc portion of IgG mAbs bound to the surface of NK cells,
thereby effectively making them surrogate ligands in redirected
cytotoxicity assays.
Ab engagement of the 3DL1/L5 chimera strongly inhibited tar-
get cell lysis to a degree that was nearly comparable to that of
3DL1 (Fig. 2A and Table I). The inhibition through 3DL1/L5 was
similar to that through engagement of 3DL1/L4, whereas cytotox-
icity was not affected by engagement of either CD56 or the control
3DL1/272P
⌬ receptor (Fig. 2A), demonstrating the fidelity of this
assay system. Therefore, despite the ITSM-like sequence, the cy-
toplasmic domain of 2DL5 can inhibit NK cell cytotoxicity to
almost the same extent as a classical KIR.
The initial stage of NK cell-mediated cytotoxicity is the forma-
tion of stable conjugates between target and effector cells, which
can be disrupted if inhibitory KIR engage with the target cell (10).
Therefore, we performed a comparative analysis of the impacts of
engaging 3DL1, 3DL1/L5, 3DL1/L4, and 3DL1/272P
⌬ on the in-
teraction of NK-92 cells with P815 cells using two-colored flow
cytometry. We found that the wild-type 3DL1 significantly re-
duced the conjugate formation between NK-92 and P815 cells
(Fig. 2B). However, the inhibitory capacities of the cytoplasmic
domains of 2DL5 and 2DL4 were weaker than that of 3DL1 (Fig.
2B). As we have previously observed (31), engagement of 3DL1/
272P
⌬ reproducibly enhanced the conjugate formation, which may
be due to the adhesion resulting from the multiple KIR/Ab/Fc re-
ceptor interactions (Fig. 2B, right panel). Thus, the cytoplasmic
domains of 2DL5 and 2DL4 can only weakly inhibit target cell
conjugate formation, despite the strong inhibition of target cell
cytotoxicity by all three KIR cytoplasmic domains. These results
indicate that differences in KIR ITIM configuration can more
readily affect the capacity to inhibit adhesion to target cells than it
can influence inhibition of cytotoxicity responses.
To determine the tyrosine sequence contributing to KIR2DL5
inhibition, the ITIM tyrosine in 3DL1/L5 was mutated to phenyl-
alanine to create the mutant chimeric receptor, named 3DL1/
L5.FY (VTYAQL to VTFAQL). The mutant chimera was ex-
pressed in the NK-92 cell line by retroviral transduction and tested
for inhibitory capacity in the redirected cytotoxicity assay. As
shown in Fig. 2C, mutation of the ITIM completely abolished in-
hibitory capacity of the 2DL5 cytoplasmic domain. This result
indicates that the ITSM-like sequence alone cannot elicit inhibi-
tory function. Our previous work has already demonstrated that
KIRs containing only the membrane-proximal ITIM (mutant
3DL1/YF or 3DL1/L4) can elicit strong inhibitory function
through recruitment of SHP-2 (20, 31). Therefore, we did not mu-
tate the ITSM-like tyrosine, because the remaining ITIM
(VTYAQL) would create a receptor that is virtually identical with
those previously tested (especially 3DL1/L4).
The KIR2DL5 cytoplasmic domain recruits SHP-1 and SHP-2
We next examined whether the 2DL5 cytoplasmic domain could
recruit SHP-1 and/or SHP-2 upon phosphorylation of the tyrosine
residues. Both SHP-1 and SHP-2 have been shown to be recruited
to classical KIRs to mediate inhibition (8, 9, 13–18, 20). In con-
trast, we have previously shown that the single ITIM of 2DL4
mediates strong inhibition of cytotoxicity, despite only binding
SHP-2, but not SHP-1 (31). Therefore, both 3DL1 and 3DL1/L5
were immunoprecipitated from pervanadate-treated NK-92 cells
and tested for SHP-1 and SHP-2 recruitment by immunoblotting.
7387
The Journal of Immunology
by guest on March 14, 2018
http://www.jimmunol.org/
Downloaded from