NK cells maintain self-tolerance through expression of inhibitory receptors that bind MHC class I (MHC-I) molecules. MHC-I can exist on the cell surface in several different forms, including “peptide-receptive” or PR-MHC-I that can bind exogenous peptide. PR-MHC-I molecules are short lived and, for H-2Kb, comprise ∼10% of total MHC-I. In the present study, we confirm that signaling through the mouse NK inhibitory receptor Ly49C requires the presence of PR-Kb and that this signaling is prevented when PR-Kb is ablated by pulsing with a peptide that can bind to it with high affinity. Although crystallographic data indicate that Ly49C can engage H-2Kb loaded with high-affinity peptide, our data suggest that this interaction does not generate an inhibitory signal. We also show that no signaling occurs when the PR-Kb complex has mouse β2-microglobulin (β2m) replaced with human β2m, although replacement with bovine β2m has no effect. Furthermore, we show that β2m exchange occurs preferentially in the PR-Kb component of total H-2Kb. These conclusions were reached in studies modulating the sensitivity to lysis of both NK-resistant syngeneic lymphoblasts and NK-sensitive RMA-S tumor cells. We also show, using an in vivo model of lymphocyte recirculation, that engrafted lymphocytes are unable to survive NK attack when otherwise syngeneic lymphocytes express human β2m. These findings suggest a qualitative extension of the “missing self” hypothesis to include NK inhibitory receptors that are restricted to the recognition of unstable forms of MHC-I, thus enabling NK cells to respond more quickly to events that decrease MHC-I synthesis.

The reactivity of a NK cell is controlled by the balance of signaling received through its activation and inhibitory receptors (1). Inhibitory receptor engagement of MHC class I (MHC-I)3 or MHC-I like homologues promotes functional quiescence such that a reduction of MHC surface expression can tilt the signaling balance toward NK activation, resulting in elimination of the target cell (2). In the mouse system, NK inhibitory receptors that bind to MHC-I molecules belong to the Ly49 family (3). There is some controversy as to the specific form of MHC-I that is engaged by various Ly49 family members. This study provides new evidence supporting a previous conclusion (4) that the mouse inhibitory receptor Ly49C, which recognizes H-2Kb as a ligand, can generate an inhibitory signal only following engagement of the unstable subset of H-2Kb that is capable of binding exogenous peptide.

MHC-I molecules found on the cell surface are heterogeneous in conformation. The most stable form is a heterotrimeric complex composed of H chain (H) folded with an appropriate short peptide (p) and stabilized by β2-microglobulin (β2m). Both peptide and β2m are bound noncovalently to the H chain and can dissociate and/or exchange with peptide or β2m from the surrounding medium (5, 6). Thus, in principle, MHC-I can exist on the cell surface in several forms, including lacking peptide or “empty” (H-β2m), lacking β2m (H-p), or lacking both (free H). It is difficult to quantify the relative numbers of these different forms. However, with the exception of H-2Ld (7), truly peptide-empty forms of murine MHC-I appear highly unstable and are likely very few in number. In this regard, radiochemical studies using the Tap-2-deficient cell line RMA-S incubated at 26°C have revealed that the “empty” H-2Kb molecules on the cell surface are actually occupied by low-affinity peptide (8). However, when the temperature is raised to 37°C, there is a rapid loss of this low-affinity peptide and loss of any detectable H chain expression (8). As well, molecular dynamics simulation studies measuring conformational flexibility of the H chain α12 domain indicate that the truly peptide-empty form of MHC-I rapidly assumes a partially unfolded state consistent with a molten globule-type structure (9, 10), which is no longer capable of binding peptide. Thus, it appears that the dominant form of the “empty” MHC-I molecule that exists on the surface is in fact occupied by low-affinity peptide that can readily be replaced with exogenous peptide. Following the precedent of Day et al. (11), we refer collectively to these molecules and to truly empty molecules as peptide-receptive (PR) MHC-I molecules based on their ability to bind exogenous peptide.

In studies of activated mouse T lymphoblasts, it was found that mouse MHC-I H-2Kb molecules on the cell surface can be classified into two populations of approximately the same size but very different mean half lives of ∼1 and 20 h, respectively (12). The more stable population contains peptide bound with medium to high affinity. The less stable, which includes PR-Kb as a major component, contains peptide bound with low affinity. Accordingly, PR-MHC-I represents a significant proportion (10–15%) of total MHC-I on the cell surface and thus provides a practical rationale for this species of MHC-I to function as a structural ligand for immune receptors.

B2m binds to MHC-I H chain and as well is present as a soluble factor in serum. B2m bound to MHC-I can exchange with β2m present in the serum. Both human (h-) and bovine β2m (b-β2m) free forms can bind murine MHC-I H chain with an affinity two to three times higher than the binding of mouse β2m (m-β2m) (13). Sequence comparisons indicate that h-β2m and b-β2m share 76 and 70% identity with m-β2m (13). The altered binding affinities and sequence differences of the different β2m moieties could well induce changes in MHC-I structure capable of influencing recognition by Ly49. Indeed, β2m has already been implicated in Ly49 recognition: MHC-I tetramer-analysis of the Ly49 ligand-binding repertoire indicates that certain Ly49 interactions, including H-2Kb engagement by Ly49C, do not occur when using tetramers complexed with h-β2m (14, 15, 16). Furthermore, functional studies using cultures containing bovine- and human-derived β2m suggest that the presence of xenogeneic β2m may interrupt recognition of H-2Dd by Ly49A (17) and that species-specific residues of m-β2m, namely Q29 and K58, when mutated to alanine or human-encoded residues, likewise interfere with Ly49A ligand binding (18, 19). Although it has been shown (20) that one of these β2m residues, the Q29A point mutation, has only a nominal affect in Ly49C recognition of H-2Kb as determined by surface plasmon resonance analysis, the cocrystal structure analyses of both of these MHC-I/Ly49 interactions underscore the crucial role of β2m in forming the inhibitory interface (20, 21).

There are currently two discordant hypotheses describing the molecular requirements for functional Ly49C/H-2Kb interaction. The first is based on an analysis of NK cell lysis of the H-2b thymoma, RMA-S. This cell line expresses minimal surface MHC-I and thus behaves as an excellent NK target. Incubating RMA-S cells overnight at a reduced temperature (26°C) in the presence of a H-2Kb-binding peptide partially protected RMA-S from Ly49C+ NK-mediated lysis to an extent that varied with the peptide used (22). The authors concluded that recognition of H-2Kb by Ly49C required the presence of a bound peptide. The second model system used syngeneic T lymphoblasts as targets. Lymphoblasts express relatively high levels of MHC-I (both total and PR-MHC-I) and are not lysed readily by NK cells. However, when briefly pulsed with a high-affinity MHC-I-binding peptide, so as to occupy all PR-Kb molecules, the cells became sensitized to syngeneic Ly49C+ NK cell lysis. This led to the conclusion that Ly49C engages the PR conformation of H-2Kb (4).

This last study used an in vitro system using FBS-supplemented medium (4). On account of the ability of serum-derived b-β2m present in the medium to bind to mouse MHC-I molecules, thereby potentially altering the conformation of the inhibitory ligand, interpretation of these findings has been questioned. In experiments reported here using both T lymphoblast and RMA-S targets, we confirm that PR-Kb is indeed the natural ligand for Ly49C and moreover show that functional recognition can occur when PR-Kb is complexed with either m- or b-β2m but not when complexed with h-β2m. Importantly, these data imply that the details of ligand conformation are extremely critical for determining the affinity of Ly49 recognition.

C57BL/6J, TapI−/−, and FcRγ−/− B6 mice (23) were purchased from The Jackson Laboratory and Taconic Farms. H-2Db−/− and H-2DbKb−/− B6 mice (24) were established from breeding pairs donated by Dr. F. Lemonnier (Pasteur Institute, Paris, France). C57BL/6 (B6) mice expressing h-β2m as a transgene either in the absence or in the presence of m-β2m were generated at the Hospital for Sick Children by breeding B6 h-β2m transgenic mice (25) with B6 murine β2m−/− mice (26). Mice were maintained in the specific pathogen-free vivarium at the Ontario Cancer Institute according to institutional guidelines. Typically, tissue was obtained from female mice, aged 6–10 wk and cultured in α-MEM (Invitrogen Life Technologies) supplemented with serum, 50 μM 2-ME, and 10 mM HEPES. RMA-S was obtained from Dr. P. Ohashi, (Ontario Cancer Institute, Toronto, Ontario, Canada).

Mouse serum (MS) was bought as pooled stock (Cedarlane Laboratories) or obtained from B6 mice via cardiac puncture. Human serum (HS) was obtained from healthy volunteers. FBS was purchased from Wisent. All sera were heat inactivated before use. For cytotoxicity experiments involving MS, the serum was fractionated at the 50,000 cutoff by size exclusion using an ultra-free 4-ml filter fitted with a Biomax 50K NMWL membrane (Millipore).

H-2Kb-binding peptides were pOVA (SIINFEKL), residues 257–264 derived from chicken OVA (27) and pVSV (RGYVYQGL), residues 52–59 from vesicular stomatitis virus nucleoprotein (28). H-2Kd-restricted control peptide was pFNP (TYQRTRALV). Peptides were purchased from the Alberta Peptide Institute and were >95% pure by HPLC. Lyophilized h-β2m was purchased from Sigma-Aldrich Canada (purity >90%). Peptides and h-β2m were dissolved in PBS−/− at 1 mg/ml and stored at −80°C. pOVA-H-2Kb tetramers were a kind gift from Dr. R. Tan (University of British Columbia, Vancouver, British Columbia, Canada).

The following mAbs were used: 5E6 (BD Biosciences), which reacts with Ly49C and Ly49I (29), Y-3 (American Type Culture Collection), which binds the α1α2 domain of H-2Kb irrespective of the β2m moiety present (30, 31, 32), AF6-88.5-88.5 (BD Biosciences), which recognizes H-2Kb when bound with m-β2m but not when bound with h-β2m or b-β2m (30, 33), KH95 (BD Biosciences), which binds H-2Db irrespective of the β2m moiety present (34, 35), W6/32 (American Type Culture Collection), which reacts with numerous HLA/MHC determinants, including b- or h-β2m-bound H-2Db but not m-β2m bound H-2Db (36), S19.8 (provided by Dr. U. Hammerling, Memorial Sloan-Kettering Cancer Center, New York, NY), which binds to m-β2m (37), and 25-D1.16 (provided by Dr. R. Germain, National Institute of Allergy and Infectious Diseases, Bethesda MD), which binds pOVA complexed with H-2Kb (38). Production of Abs from hybridomas involved isolation on protein G (Sigma-Aldrich Canada) or gammabind plus (Amersham Biosciences) Sepharose-based chromatography and FITC labeling using FITC-CELITE (Calbiochem).

For staining, cells were washed in BSA/serum-free PBS−/− buffer and resuspended at ∼2 × 105 cells in 100-μl volumes for 30 min on ice. Determination of PR-Kb surface expression was ascertained by pulsing ∼2 × 105 cells with 1 μg of pOVA or pVSV (control) on ice for 45 min, washing, and subsequently staining with 25-D1.16. In certain instances, cells were treated with brefeldin A (BFA) (5 μg/ml; Sigma-Aldrich Canada) to block intracellular transport. Samples were analyzed on a FACSCalibur (BD Biosciences) in logarithmic amplification of the fluorescent light signals and linear amplification of the forward/side scatter light signals following compensation to remove spectral overlap in the fluorescent channels. Data were analyzed on CellQuest software (BD Biosciences) using the gated, viable population as ascertained by linear scale forward/side scatter.

B6 Con A blasts were generated by incubating ∼107 splenocytes for 2 days in complete medium supplemented with 2 μg/ml Con A (ICN Pharmaceuticals Canada). Upon harvesting, viable cells were enriched on Lympholyte-M (Cedarlane Laboratories), washed, and incubated for 2 min in 200 mM methyl α-d-mannoside/methyl α-d-glucoside (Sigma-Aldrich Canada) to remove residual lectin.

LAK cells were generated as described previously (4). Briefly, splenocytes were red cell depleted using Lympholyte-M, and adherent cells were depleted using nylon wool columns. The nonadherent fraction was NK enriched using anti-CD4, CD8, and B220-conjugated magnetic beads (Dynal). Remaining cells were cultured for 6–11 days in 5.0 ml of α-MEM supplemented with 10% FBS, 50 μM 2-ME, and 10,000 U/ml human rIL-2 (Chiron). Ly49C+ and Ly49C LAK populations were stained using mAbs 5E6 and NK1.1 (PK136) and sorted using a BD FACSVantage. Cells were recultured for at least 48 h before use. Note that a previous study (4) had shown that Ly49I cannot recognize either H-2Kb or H-2Db.

RMA-S or Con A blasts were labeled for 90 min with 100 μCi Na51CrO4 (NEN Life Science) in 100 μl of serum (Con A blasts) or 1% FCS in complete medium (RMA-S). Targets were then washed. In certain instances, labeled targets were then incubated with 1 μg/ml pOVA or 10 μg/ml h-β2m for 45 min at 26°C. Anti-Ly49C (5E6) was directly added at 1–2 μg/well to the plated effectors, followed by mixing. Targets were plated at 20,000 viable cells/ml in 100-μl aliquots. Plates were incubated at 26°C (for RMA-S targets) or 37°C (for Con A targets) for 4 h, after which 100 μl of cell-free supernatant were harvested. Percent-specific lysis at each E:T ratio was determined by averaging the counts per minute from five replicate wells and calculated as previously (4). All experiments reported here were repeated a minimum of three times.

Splenocytes from h-β2m Tg, wild-type (Wt), and β2m−/− mice were depleted of red cells and labeled with 1 μM CFSE (BioCan Scientific) for 10 min at 37°C in serum-free PBS. Cells were washed twice in complete medium. Then 3.0 × 107 cells were injected into the lateral tail vein of recipient mice. In vivo NK activation was induced by i.p. administration of 100 μg of polyinosine-polycytidylic acid (poly I:C) (Sigma-Aldrich Canada) at 1 and 24 h postengraftment. At 48 h postengraftment, inguinal and axillary lymph nodes were harvested and pooled. Single-cell suspensions were analyzed by flow cytometry to determine the relative percentage of CFSE-labeled leukocytes. The experiment was repeated three times. Error refers to SEM generated by quadruplicate measurements from one experiment.

To test whether the source of serum used during the in vitro generation of mouse Con A blasts had an effect on sensitivity to lysis by syngeneic NK cells, B6 splenocytes were cultured for 2 days in complete medium containing 2 μg/ml Con A and either FBS, MS, or HS. The cells were then assessed for sensitivity to lysis by syngeneic lymphokine-activated NK cells (LAK). As demonstrated in Fig. 1.1.A and in accord with previous findings (4), Con A blasts that had been cultured in medium supplemented with 10% FBS and then pulsed with the H-2Kb high-affinity binding peptide, pOVA, before the killing assay, were 5.6-fold more sensitive to lysis (based on LU analysis of the titration data) than Con A blasts that had been left unpulsed or pulsed with a control peptide. Target cells grown in 3% (syngeneic) MS behaved similarly, becoming 6.8-fold more sensitive to lysis when pulsed with pOVA relative to unpulsed or control pulsed targets (Fig. 1.1.B). However, Con A blasts grown in 3% HS were not further sensitized to lysis following addition of peptide. Indeed, the background lysis of (unpulsed) targets grown in HS was much higher than for cells grown in FBS or MS and was comparable to that seen when FBS- or MS-cultivated Con A blasts were pulsed with pOVA (Fig. 1.1.C). Control growth kinetic experiments titrating the different sera indicated that 10% FBS, 3% MS, and 3% HS produced comparable growth of Con A blasts (data not shown).

FIGURE 1.

Cells grown using different sera compared for sensitivity to lysis by syngeneic LAK and for expression of different forms of MHC-I. 1.1) Day 2 B6 Con A blasts cultured in 10% FBS (A), 3% MS (B), or 3% HS (C) were assessed for sensitivity to lysis by syngeneic LAK (35% Ly49C+). Assay cultures contained 1% of the respective serum. Target cells were either untreated or pulsed with pFNP (pCTRL, does not bind) or with pOVA, as indicated. 1.2) Day 2 B6 Con A blasts cultured in 10% FBS were assessed for sensitivity to lysis by syngeneic LAK. The effects of pulsing with pOVA and blockading Ly49C are compared either performed separately or together. 1.3) Flow cytometric analysis of MHC-I and β2m expression levels on cells from Fig. 1.1.

FIGURE 1.

Cells grown using different sera compared for sensitivity to lysis by syngeneic LAK and for expression of different forms of MHC-I. 1.1) Day 2 B6 Con A blasts cultured in 10% FBS (A), 3% MS (B), or 3% HS (C) were assessed for sensitivity to lysis by syngeneic LAK (35% Ly49C+). Assay cultures contained 1% of the respective serum. Target cells were either untreated or pulsed with pFNP (pCTRL, does not bind) or with pOVA, as indicated. 1.2) Day 2 B6 Con A blasts cultured in 10% FBS were assessed for sensitivity to lysis by syngeneic LAK. The effects of pulsing with pOVA and blockading Ly49C are compared either performed separately or together. 1.3) Flow cytometric analysis of MHC-I and β2m expression levels on cells from Fig. 1.1.

Close modal

We next directly assessed the participation of Ly49C in the lysis of target cells cultured in the three different sera. LAK cells generated from FcRγ-deficient B6 mice were sorted into Ly49C+ and Ly49C populations for use as effectors against Wt Con A blast targets grown using the three different sera. Results are summarized in Table I. Killing of FBS- and MS-cultured Con A blasts by Ly49C+ LAK was enhanced when target cells were pretreated with pOVA or when anti-Ly49C was added to the killing assay as compared with lysis of untreated targets or killing mediated by Ly49C LAK. Our interpretation is that lysis occurs because either the receptor (Ly49C) is being blocked or its ligand (PR-Kb) is being ablated. This interpretation predicts that simultaneous addition of both pOVA and anti-Ly49C should produce no additional enhancement of lysis. The experiment of Fig. 1.2 shows that this is indeed the case.

Table I.

Sensitivity to lysis by Ly49C+ or Ly49C LAK of B6 lymphoblasts grown in medium containing FBS, HS, or MSa

Target B6 Con AEffector Phenotypeb
Wt Ly49C+ LAKWt Ly49C LAK
MockpOVAα-Ly49CMockpOVAα-Ly49C
10% FBS 5.2 ± 0.8 24.5 ± 1.0 16.7 ± 0.9 6.5 ± 1.2 5.9 ± 1.3 5.9 ± 1.7 
3% MS 3.5 ± 1.1 12.6 ± 1.2 8.2 ± 0.8 5.5 ± 1.5 6.0 ± 2.1 4.6 ± 1.2 
3% HS 11.5 ± 1.8 11.9 ± 1.2 11.9 ± 1.9 11.0 ± 1.0 10.6 ± 1.0 11.2 ± 1.5 
Target B6 Con AEffector Phenotypeb
Wt Ly49C+ LAKWt Ly49C LAK
MockpOVAα-Ly49CMockpOVAα-Ly49C
10% FBS 5.2 ± 0.8 24.5 ± 1.0 16.7 ± 0.9 6.5 ± 1.2 5.9 ± 1.3 5.9 ± 1.7 
3% MS 3.5 ± 1.1 12.6 ± 1.2 8.2 ± 0.8 5.5 ± 1.5 6.0 ± 2.1 4.6 ± 1.2 
3% HS 11.5 ± 1.8 11.9 ± 1.2 11.9 ± 1.9 11.0 ± 1.0 10.6 ± 1.0 11.2 ± 1.5 
a

Here and in Tables II and III, percent-specific lysis at E:T = 20:1 is shown with the error being the SEM calculated from five replicates of each point. Here and in Tables II and III, similar results were obtained at two other E:T ratios and in two repeats of the entire experiment.

b

Targets were prepulsed with pOVA or Ly49C recognition blockaded as indicated.

In the experiments summarized in Table I, killing of HS-cultured Con A targets by both Ly49C and Ly49C+ LAK was consistently elevated regardless of treatment protocol. We hypothesized that in the presence of HS, inhibitory signaling was suppressed, perhaps because ligand structure was altered through binding of h-β2m. We first determined how the serum present during culture might affect MHC-I epitopes (Fig. 1.3). Expression levels of total H-2Db, H-2Kb, and PR-Kb were roughly comparable for cells grown in the three different sera (Fig. 1.3.A–C). Note that measurement of PR-Kb was determined by pulsing the cells with pOVA and then staining with 25-D1.16, a mAb specific for pOVA-H-2Kb (see Materials and Methods). Cells cultured in FBS or HS could exchange endogenous β2m with β2m in the culture medium as xenogeneic β2m was detectable on the cell surface using the mAb W6/32 (Fig. 1.3.D), although the amount of m-β2m on the cell surface was only slightly diminished as a consequence of this exogenous h-β2m or b-β2m binding (Fig. 1.3.E).

We hypothesized that binding of h-β2m to H-2Kb interferes with the ability of Ly49C to engage its ligand. To assess directly whether the effect is due to h-β2m and not some other serum component, lymphoblast targets were cultured and assayed in medium containing reduced (5%) FBS with or without being pulsed with exogenous-purified h-β2m (10 μg/ml) immediately before being used as targets for Ly49C+ and Ly49C LAK in the presence of anti-Ly49C mAb, pOVA (1 μg/ml), or mock treatment. The results are summarized in Table II. Pulsing target cells with h-β2m resulted in enhanced Ly49C+ LAK-mediated killing that was comparable to the lysis achieved through either pOVA pulsing or Ly49C blockade. We conclude that inclusion of h-β2m is altering the PR-Kb structure such that it is no longer recognizable by Ly49C. However, h-β2m appeared to exert an additional effect: although killing by Ly49C LAK was unaffected by either peptide pulsing or Ly49C blockade, it was enhanced by addition of h-β2m to the same extent as the Ly49C+ LAK-mediated killing. Note that this same effect was seen previously for target cells cultured in HS (Table I). One explanation is that there is an inhibitory receptor (not Ly49C) that is specific for H-2Db but which cannot recognize H-2Db complexed with h-β2m. To test this, the experiment was repeated using H-2Db−/− B6 mice for the generation of both target cells and LAK (Table II). As before, inclusion of h-β2m in the assay made H-2Db−/− Con A blasts sensitive to Ly49C+ LAK-mediated killing. However, now inclusion of h-β2m did not augment Ly49C LAK-mediated killing, supporting the notion of a novel H-2Db-specific inhibitory receptor (see Discussion).

Table II.

Effect of exogenous h-β2m on B6 lymphoblast sensitivity to lysis by Ly49C+ or Ly49C LAKa

Target TreatmentEffector:Target Phenotype
Wt B6 Con A targetH-2Db−/− B6 Con A target
Ly49C+ Wt, LAKLy49C Wt LAKLy49C+ H-2Db−/− LAKLy49C H-2Db−/− LAK
Mock 3.9 ± 1.1 3.8 ± 0.8 13.6 ± 1.2 11.6 ± 0.9 
α-Ly49C 12.0 ± 2.0 4.7 ± 1.2 31.9 ± 1.9 15.2 ± 1.0 
pOVA 15.8 ± 1.3 4.2 ± 0.9 35.1 ± 1.2 13.1 ± 1.0 
h-β212.1 ± 1.7 12 ± 1.6 25.2 ± 1.3 15.4 ± 1.2 
Target TreatmentEffector:Target Phenotype
Wt B6 Con A targetH-2Db−/− B6 Con A target
Ly49C+ Wt, LAKLy49C Wt LAKLy49C+ H-2Db−/− LAKLy49C H-2Db−/− LAK
Mock 3.9 ± 1.1 3.8 ± 0.8 13.6 ± 1.2 11.6 ± 0.9 
α-Ly49C 12.0 ± 2.0 4.7 ± 1.2 31.9 ± 1.9 15.2 ± 1.0 
pOVA 15.8 ± 1.3 4.2 ± 0.9 35.1 ± 1.2 13.1 ± 1.0 
h-β212.1 ± 1.7 12 ± 1.6 25.2 ± 1.3 15.4 ± 1.2 
a

Target cells were from Wt B6 (left column) or H-2Db−/− B6 (right column). Targets were prepulsed with pOVA or h-β2m or Ly49C recognition blockaded as indicated.

To determine whether variation in serum β2m concentrations might account for differences in serum-induced target cell susceptibility, HS and FBS β2m concentrations were estimated using the mAb W6/32 (see Materials and Methods). Splenocytes were pulsed with varying concentrations of h-β2m, and staining was measured by flow cytometry to establish a relationship between mean fluorescent intensity (MFI) and h-β2m concentration (data not shown). Assuming comparable affinities of W6/32 for h- and b-β2m-defined epitopes, xeno-β2m concentrations were estimated as ∼1.7 μg/ml for FBS and ∼10 μg/ml for HS. As xeno-β2m concentrations present in cultures using 10% FBS or 3% HS were comparable, 0.17 and ∼0.3 μg/ml, respectively, it is unlikely that target cell susceptibility was influenced by differences in β2m concentrations.

Previous work (12, 39) has shown that PR-MHC-I on the cell surface has a short half-life but is being continuously replaced from the cell interior. Cells made sensitive to lysis by being pulsed briefly with peptide (thus ablating PR-MHC-I) regain resistance to lysis after an incubation of ∼90 min in the absence of peptide through replacement of the inhibitory ligand by de novo export of PR-MHC-I from the cell interior. However, cells remained sensitive to lysis when export of protein from the interior was blocked by addition of BFA, a blocker of protein export. The converse of this is that cells resistant to lysis should become sensitive after being incubated with BFA for a time sufficiently long (at least 3 h) for the number of PR-MHC-I on the cell surface to decay below the critical number required for the generation of an inhibitory signal. The experiment of Fig. 2 demonstrates this to be the case. Normal B6 Con A blasts either pulsed with pOVA or pulsed with h-β2m or incubated with BFA for 3.5 h all became approximately four times more sensitive to lysis (LU analysis) than control untreated cells.

FIGURE 2.

Effect on sensitivity to lysis of long incubation in BFA vs short pulse with pOVA or h-β2m. Day 2 B6 Con A blasts were pulsed with pOVA or h-β2m as in Table II or were cultured for 4 h in BFA before being compared for sensitivity to lysis by syngeneic LAK.

FIGURE 2.

Effect on sensitivity to lysis of long incubation in BFA vs short pulse with pOVA or h-β2m. Day 2 B6 Con A blasts were pulsed with pOVA or h-β2m as in Table II or were cultured for 4 h in BFA before being compared for sensitivity to lysis by syngeneic LAK.

Close modal

When Con A blasts were incubated in vitro with HS, h-β2m appeared to bind to only a small fraction of total H-2Kb molecules (compare Fig. 1.3, D and E). Similarly, when Con A blasts are pulsed with pOVA, peptide binds to only a small fraction (∼10%) of H-2Kb molecules (see e.g., Ref.12). In either case, this low level of binding is sufficient to alter H-2Kb such that it was no longer functionally recognizable by Ly49C. As addition of either h-β2m or pOVA produced equivalent enhancement of lysis by Ly49C+ NK cells (e.g., Tables I and II), we assessed whether h-β2m might have a preferential ability to complex with PR-Kb molecules. B6 Wt Con A blasts were either untreated or pretreated with pOVA (45 min at 37°C) to ablate PR-Kb molecules and then assessed by flow cytometry for the ability of H-2Kb molecules to bind exogenous h-β2m. Because nascent PR-Kb molecules are being continuously exported to the cell surface, the experiments were done in the presence of BFA. Binding of h-β2m was then assessed by comparing flow cytometric MFI values for mAb Y-3 (recognizes H-2Kb bound by either h-β2m or m-β2m; see Fig. 3) with MFI values for mAb AF6-88.5 (recognizes H-2Kb bound by m-β2m but not h-β2m; again see Fig. 3) for cells exposed to h-β2m that had or had not been pretreated with pOVA. Results are shown in Fig. 4.A and are normalized to the MFI for cells treated with neither pOVA nor h-β2m (Fig. 4.A, line 1). For cells not exposed to pOVA, incubation with h-β2m (45 min at 37°C) reduced m-β2m specific AF6-88.5 binding by ∼13%, while slightly increasing Y-3 staining (Fig. 4.A, lines 2 vs 1), suggesting that h-β2m was binding to ∼13% of total H-2Kb. When cells were pretreated with pOVA before addition of h-β2m, total H-2Kb expression was elevated ∼12–15% (Fig. 4.A, compare lines 1 and 3), but there was little difference in AF6-88.5 staining (line 3), which is consistent with there being no sites available for h-β2m to bind following pOVA addition. Thus, binding of pOVA to PR-Kb appears to prevent β2m exchange. The elevated H-2Kb expression seen in Fig. 4.A, line 3, was likely due to the stabilizing effect of peptide binding to PR-Kb molecules that otherwise would have denatured. Consistent with this reasoning, pulsing with pOVA alone increased the staining of both mAbs (Fig. 4.A, line 4). We conclude that ablation of PR-Kb greatly reduces the ability of H-2Kb to bind exogenous h-β2m.

FIGURE 3.

Flow cytometric analysis of the binding of the anti-H2-Kb mAbs AF6-88.5 (A) and Y-3 (B) to spleen cells from either Wt or transgenic h-β2m mice.

FIGURE 3.

Flow cytometric analysis of the binding of the anti-H2-Kb mAbs AF6-88.5 (A) and Y-3 (B) to spleen cells from either Wt or transgenic h-β2m mice.

Close modal
FIGURE 4.

Alterations in H-2Kb epitopes produced by addition of peptide and/or h-β2m. Con A blasts cultured in 5% FBS were treated with BFA (5 μg/ml) and then given two consecutive 45-min incubations with nothing or pOVA (1.0 μg/ml) in the first and nothing or h-β2m (10 μg/ml) in the second (A) or nothing, h-β2m (1.0 μg/ml), or pVSV (1.0 μg/ml) in the first and nothing or pOVA (1.0 μg/ml) in the second (B). Cells in each group in A were stained with either mAb Y-3 (recognizes H-2Kb containing h- or m-β2m) or with mAb AF6-88.5 (recognizes only H-2Kb containing m-β2m). Cells in each group in B were stained with mAb 25D-1.16 (recognizes PR-Kb).

FIGURE 4.

Alterations in H-2Kb epitopes produced by addition of peptide and/or h-β2m. Con A blasts cultured in 5% FBS were treated with BFA (5 μg/ml) and then given two consecutive 45-min incubations with nothing or pOVA (1.0 μg/ml) in the first and nothing or h-β2m (10 μg/ml) in the second (A) or nothing, h-β2m (1.0 μg/ml), or pVSV (1.0 μg/ml) in the first and nothing or pOVA (1.0 μg/ml) in the second (B). Cells in each group in A were stained with either mAb Y-3 (recognizes H-2Kb containing h- or m-β2m) or with mAb AF6-88.5 (recognizes only H-2Kb containing m-β2m). Cells in each group in B were stained with mAb 25D-1.16 (recognizes PR-Kb).

Close modal

Reciprocally, to determine whether H-2Kb molecules bound by h-β2m could still function as PR-Kb molecules, Con A blasts (in the presence of BFA) were either untreated or pretreated with h-β2m and then pulsed with pOVA (Fig. 4.B). Subsequent binding of pOVA was detected using the mAb 25D-1.16 (see Fig. 1.2.C and Ref.38). Pretreatment of cells with h-β2m did not ablate PR-Kb because subsequent binding of pOVA was not altered (Fig. 4.B, lines 1–3). As a control, prepulsing with pVSV (which also binds PR-Kb) greatly reduced the ability of pOVA to bind H-2Kb (Fig. 4.B, lines 2 vs 5). Thus, H-2Kb molecules containing h-β2m can exist as PR-Kb molecules, although it seems this altered PR architecture is not recognizable by Ly49C.

In all the above experiments, β2m exchange has been limited to a small subset of all MHC-I molecules. We sought to confirm our findings using cells in which all MHC-I molecules contained h-β2m. To this end, B6 m-β2m−/− mice were bred with h-β2m transgenic B6 mice (see the Materials and Methods). Con A blasts derived from Wt (only m-β2m), heterocompound (both h-β2m and m-β2m) and h-β2m transgenic (only h-β2m) B6 mice were tested as targets for Wt B6 LAK (Fig. 5.1.A–C). Both heterozygous and h-β2m transgenic targets demonstrated elevated sensitivity to Wt LAK. Peptide pulsing increased the lysis of Wt control targets but caused no additional increase in lysis of targets containing h-β2m.

FIGURE 5.

Effect of replacing m-β2m with h-β2m on sensitivity to LAK-mediated lysis and on expression of MHC-I. 5.1) Cells from Wt B6 (A), heterozygous B6 (both h- and m-β2m; B), and transgenic B6 (only h-β2m; C) were cultured in medium containing FBS and tested for sensitivity to lysis by Wt B6 LAK. Target cells were either untreated or pOVA pulsed before assay. 5.2) Flow cytometry analysis of H-2Db, H-2Kb, PR-Kb, h-β2m, and m-β2m on Wt, heterozygous, and transgenic Con A target cells.

FIGURE 5.

Effect of replacing m-β2m with h-β2m on sensitivity to LAK-mediated lysis and on expression of MHC-I. 5.1) Cells from Wt B6 (A), heterozygous B6 (both h- and m-β2m; B), and transgenic B6 (only h-β2m; C) were cultured in medium containing FBS and tested for sensitivity to lysis by Wt B6 LAK. Target cells were either untreated or pOVA pulsed before assay. 5.2) Flow cytometry analysis of H-2Db, H-2Kb, PR-Kb, h-β2m, and m-β2m on Wt, heterozygous, and transgenic Con A target cells.

Close modal

We next tested the ability of Wt Ly49C+ and Ly 49C LAK to lyse m-β2m+ or h-β2m+ targets. Con A blasts generated from either Wt or h-β2m transgenic B6 mice were cultivated in medium supplemented with 3% MS or 3% HS (respectively) and then assessed for sensitivity to Wt Ly49C+ and Ly 49C LAK. As summarized in Table III, Wt Con A blast targets were selectively sensitive to Ly49C+ LAK-mediated lysis upon inclusion of peptide, anti-Ly49C or h-β2m, while killing by Ly49C LAK was enhanced only by h-β2m. In contrast, h-β2m transgenic tissue demonstrated enhanced sensitivity to lysis for both the Ly49C+ and Ly 49C effector populations and this sensitivity was not increased with either pOVA pulsing or Ly49C blockade. The results recapitulate the pattern demonstrated in Tables I and II where only a small fraction of all MHC-I molecules contained h-β2m.

Table III.

Effect of endogenous h-β2m on B6 lymphoblast sensitivity to Wt B6 Ly49C+ or Ly49C LAKa

Target Con ATarget TreatmentEffector Phenotype
Ly49C+ Wt LAKLy49C Wt LAK
Wt B6 3% MS 3.8 ± 0.7 4.4 ± 1.3 
Wt B6 3% MS; pOVA 15.0 ± 1.5 5.6 ± 0.9 
Wt B6 3% MS; α-Ly49C 11.6 ± 0.7 3.7 ± 0.7 
Wt B6 h-β211.8 ± 0.9 11.5 ± 1.5 
h-β2m B6 3% HS 19.9 ± 1.8 19.3 ± 1.5 
h-β2m B6 3% HuS; pOVA 19.3 ± 0.7 18.6 ± 0.9 
h-β23% HuS; α-Ly49C 18.6 ± 1.5 17.5 ± 0.8 
Target Con ATarget TreatmentEffector Phenotype
Ly49C+ Wt LAKLy49C Wt LAK
Wt B6 3% MS 3.8 ± 0.7 4.4 ± 1.3 
Wt B6 3% MS; pOVA 15.0 ± 1.5 5.6 ± 0.9 
Wt B6 3% MS; α-Ly49C 11.6 ± 0.7 3.7 ± 0.7 
Wt B6 h-β211.8 ± 0.9 11.5 ± 1.5 
h-β2m B6 3% HS 19.9 ± 1.8 19.3 ± 1.5 
h-β2m B6 3% HuS; pOVA 19.3 ± 0.7 18.6 ± 0.9 
h-β23% HuS; α-Ly49C 18.6 ± 1.5 17.5 ± 0.8 
a

Cells used as targets were from Wt B6 or h-β2m transgenic B6 mice and were cultured in MS or HS. Targets were prepulsed with pOVA or h-β2m or Ly49C recognition blockaded as indicated.

Comparison of MHC-I ligand structures present in the context of the h-β2m transgene was assessed via FACS analysis (Fig. 5.2). H-2Db expression was elevated ∼2-fold in the transgenic mouse as compared with Wt tissue (Fig. 5.2.A), whereas H-2Kb expression was only slightly increased (Fig. 5.2.B). Relative PR-Kb expression in the transgenic mouse was elevated ∼3-fold (Fig. 5.2.C). In the heterozygous mouse tissue expressing both h-β2m and m-β2m, most MHC-I molecules carried h-β2m (Fig. 5.2, D and E), probably because h-β2m binds MHC-I with higher affinity than m-β2m (13). This is the likely explanation for why target cells derived from the heterozygous mouse containing both forms of β2m were killed almost as well as target cells from the homozygous h-β2m mouse (compare Fig. 5.1, B and C). Importantly, the number of H-2Kb molecules per cell in the transgenic mouse carrying h-β2m was >10-fold higher than for Wt cells cultured in HS (compare Figs. 5.2 and 1.3, particularly Fig. 1.3, B, D, and E, with Fig. 4.2, B, D, and E), and the fraction of h-β2m-complexed H-2Kb that was peptide receptive was >30-fold higher (Fig. 5.2.C and comparison of total H-2Kb expression). The observation that target cells expressing H-2Kb complexed exclusively with h-β2m were efficiently killed by NK cells expressing Ly49C provides direct evidence that Ly49C cannot recognize H-2Kb molecules containing h-β2m.

MHC-I incompatible lymphocytes i.v. injected into a mouse are rapidly eliminated by host NK cells, whereas syngeneic cells survive and enter the recirculating lymphocyte pool (40). Survival can be easily monitored by fluorescence labeling of the grafted cells before injection. We applied this model to assess whether injected leukocytes, syngeneic to the host except for replacement of m-β2m with h-β2m, would persist in vivo. In this regard, splenocytes obtained from Wt (syngeneic MHC-I, should survive in vivo), β2m−/− mice (insignificant MHC-I expression, should be rejected in vivo), or h-β2m transgenic mice were labeled with CFSE and injected (3.0 × 107/recipient) into Wt B6 mice. In addition, a portion of recipients were treated with poly I:C via i.p. injections, as this is known to induce activation of NK cells in vivo (41). At 48 h posttransplantation, rejection of both h-β2m+ cells and β2m-deficient cells was moderate in control animals but near total in animals that were treated with poly I:C (Fig. 6). As expected, the Wt graft was retained, regardless of treatment. As H-2Kb and H-2Db expression levels were similar for both the Wt and h-β2m transgenic tissue (Fig. 5.2, A and B), rejection of h-β2m transgenic leukocytes coincided with altered MHC-I conformation and not altered MHC-I frequency.

FIGURE 6.

Effect of replacing m-β2m with h-β2m on otherwise syngeneic leukocyte engraftment: CFSE-labeled splenocytes (3.0 × 107/recipient) from Wt, h-β2m-transgenic, and β2m-deficient B6 mice were infused into Wt B6 mice. Recipients were either untreated (□) or treated with poly I:C (▦). Inguinal and axillary lymph nodes were harvested at 48 h, and pooled leukocytes were assessed for percentage of CFSE+ cells via flow cytometry. Error refers to SEM generated by quadruplicate measurements of the same experiment. Results are representative of three experiments.

FIGURE 6.

Effect of replacing m-β2m with h-β2m on otherwise syngeneic leukocyte engraftment: CFSE-labeled splenocytes (3.0 × 107/recipient) from Wt, h-β2m-transgenic, and β2m-deficient B6 mice were infused into Wt B6 mice. Recipients were either untreated (□) or treated with poly I:C (▦). Inguinal and axillary lymph nodes were harvested at 48 h, and pooled leukocytes were assessed for percentage of CFSE+ cells via flow cytometry. Error refers to SEM generated by quadruplicate measurements of the same experiment. Results are representative of three experiments.

Close modal

RMA-S tumor cells have greatly reduced MHC-I expression. This makes the cells very sensitive NK targets. Expression of MHC-I can be partially rescued by incubating the cells overnight at 26°C (42, 43). RMA-S cells were incubated overnight at 26°C and then also incubated for an additional 1.5 h (still at 26°C) with nothing, with pOVA, or with h-β2m. Cells were then immediately tested for sensitivity to lysis by Ly49C+ LAK effectors at 26°C (Fig. 7.A). Both the h-β2m-pulsed and pOVA-pulsed targets were substantially more sensitive to lysis than unpulsed targets. Based on the results presented above using Con A blasts, we hypothesize that there were appreciable numbers of PR-Kb molecules on the RMA-S cell surface after overnight incubation at 26°C. The h-β2m pulsing rendered these PR-Kb molecules unrecognizable by Ly49C, and the pOVA pulsing converted them to equally unrecognizable stable H2-Kb molecules. As a first test of this, MHC-I levels were also measured for the three conditions (Table IV). H-2Kb expression was increased >5-fold by pulsing with pOVA. The implication is that this increase in total H-2Kb expression was due to binding of pOVA to unstable PR-Kb molecules present during the pOVA pulse. Indeed, staining with the mAb against the pOVA-Kb complex (indicated in the pOVA-Kb column) was up ∼20-fold over background. In the present study, this is a measure of the amount of PR-Kb present at the time of pulsing. It is clear that the pOVA pulsing produced not only an increased level of H-2Kb expression but also an increase in lysis, clearly incompatible with peptide-loaded H-2Kb being the ligand for Ly49C but in accord with the PR-MHC-I model of Ly49C/H-2Kb recognition. Note that Table IV contains a quantitative analysis of target cell sensitivity for the different treatments and effector populations (E:T ratio required for 15% lysis), as well as the MHC-I expression level data.

FIGURE 7.

Lysis of a NK-sensitive tumor target. RMA-S was incubated overnight at 26°C with or without pOVA (10 μg/ml). Target cells were labeled and then some pulsed with h-β2m or pOVA at 26°C as indicated and assessed for sensitivity to lysis, also at 26°C, by Ly49C+ (A) or Ly49C (B) LAK effectors. C, RMA-S cells treated with hβ2m, pVSV, or pOVA as indicated were assessed for sensitivity to lysis at 37°C by unfractionated B6 LAK.

FIGURE 7.

Lysis of a NK-sensitive tumor target. RMA-S was incubated overnight at 26°C with or without pOVA (10 μg/ml). Target cells were labeled and then some pulsed with h-β2m or pOVA at 26°C as indicated and assessed for sensitivity to lysis, also at 26°C, by Ly49C+ (A) or Ly49C (B) LAK effectors. C, RMA-S cells treated with hβ2m, pVSV, or pOVA as indicated were assessed for sensitivity to lysis at 37°C by unfractionated B6 LAK.

Close modal
Table IV.

Effect of exogenous peptide and h-β2m on RMA-S sensitivity to Wt Ly49C+ LAK and on MHC-I expressiona

Target Phenotype
Target Sensitivity to Lysis: E:T Ratio Required for 15% Lysis ByMHC-I MFI staining
Target RMA-SLy49C+ Wt LAKLy49C Wt LAKKbDbpOVA-Kb
Mock 4.5 3.8 16.6 22.9 3.9 
h-β2m (10 μg/ml) 0.63 0.57 23.3 23.5 3.6 
pOVA 1.5 h (1 μg/ml) 0.75 115 82.8 27.1 82.1 
pOVA O/N (1 μg/ml) 655.0 187 90.6 26.4 103.7 
pVSV O/N + pOVA 1.5 h 1.0 138 109.4 26.9 41.8 
Target Phenotype
Target Sensitivity to Lysis: E:T Ratio Required for 15% Lysis ByMHC-I MFI staining
Target RMA-SLy49C+ Wt LAKLy49C Wt LAKKbDbpOVA-Kb
Mock 4.5 3.8 16.6 22.9 3.9 
h-β2m (10 μg/ml) 0.63 0.57 23.3 23.5 3.6 
pOVA 1.5 h (1 μg/ml) 0.75 115 82.8 27.1 82.1 
pOVA O/N (1 μg/ml) 655.0 187 90.6 26.4 103.7 
pVSV O/N + pOVA 1.5 h 1.0 138 109.4 26.9 41.8 
a

RMA-S cells cultured overnight at 26°C were incubated either with or without peptide as indicated. After 16 h, cells were washed and left at 26°C for 90 min and then pulsed with pOVA, h-β2m, or nothing for another 90 min. Target cells were either used in cytotoxicity assays as in Fig. 6 or stained to measure influence of treatment on H-2Kb, H-2Db, and PR-Kb. Relative killing values, derived from the E:T titration curve, are given as the number of effector cells per target cell required to achieve 15% lysis. MFI staining for PE and FITC isotype controls were 3.9 and 3.4, respectively.

To test more directly the model of Franksson et al. (22), we examined the effect of including pOVA in the overnight incubation at 26°C (Fig. 7.A). Groups were also included in which cells incubated overnight with pOVA were then incubated an additional 3 h (at 26°C) without peptide present, the 3-h time being chosen as representative of the time range required to prepare target cells for the cytotoxicity assay. Target cells were then further incubated for 1.5 h (at 26°C) with nothing or with pOVA and then used immediately in cytotoxicity assays performed at 26°C. Cells incubated overnight with peptide were protected from lysis by Ly49C+ effectors relative to control overnight cultures (Fig. 7.A, compare ♦ with ▪). However, when pulsed with pOVA immediately before assay, thus ablating PR-Kb, the targets became much more sensitive to lysis and were just as sensitive as pOVA-pulsed cells from control overnight cultures (Fig. 7.A, compare ♦ with crosses and ▵). MHC-I expression after the different treatments was also measured. Incubating cells overnight in pOVA produced more than a 25-fold increase in pOVA-Kb expression level as measured with mAb 25D1.16 (Table IV). PR-Kb at the time of assay cannot be measured in our standard assay by using its ability to bind pOVA because most/all of the stable Kb already present has been stabilized by binding pOVA during the overnight incubation. To overcome this limitation, a group was included in which cells were incubated overnight with the H-2Kb-binding peptide pVSV and then pulsed with pOVA as above for 3 h. In this context, 25D1.16 mAb staining is a true indication of the PR-Kb present immediately before the pOVA pulse. In this scenario, killing was up ∼10-fold over background (control) lysis (Table IV). Once again, cells that were pulsed with pOVA immediately before the assay became more sensitive NK targets (Table IV), as predicted by our PR-MHC-I model.

Lysis of RMA-S cells by Ly49C NK cells was similarly tested (Table IV, Fig. 7.B). Exposure to pOVA overnight produced a >10-fold decrease in sensitivity to lysis (LU analysis) as compared with cells unexposed to pOVA. In contrast, RMA-S cells exposed to pOVA overnight and then prepulsed with pOVA again immediately before assay produced a very similar increase in susceptibility to lysis (>5-fold) as only prepulsing before assay with peptide. It is known that pOVA can bind to H-2Db, although it does so with very low affinity (44), consistent with the small increase in H-2Db expression seen in Table IV. The results are consistent with a novel inhibitory receptor that is specific for peptide-loaded H-2Db but one that cannot recognize H-2Db that contains h-β2m. It is most likely the same receptor as implicated above (Tables I–III) in the lysis of B6 Con A blasts by Ly49C LAK following pulsing with h-β2m (see Discussion).

The lysis values indicated in Fig. 7, A and B, are low as the cytotoxicity assay was done at 26°C. To verify that the same conclusions apply at 37°C and also to test directly the effect of overnight pulsing with a peptide of lower affinity, we did the experiment of Fig. 7.C in which we also tested pVSV, a peptide that binds H-2Kb but with much less affinity than pOVA. Thus, the KD, defined here as the concentration of peptide present during pulsing required to sensitize B6 Con A blasts to half-maximum values of NK lysability, was 1000-fold higher for pVSV than for pOVA (∼100 vs ∼0.1 pM) (4). Cells incubated overnight with either pOVA or pVSV were protected compared with the no peptide control, but the protection conferred by pVSV was substantially less (Fig. 7.C, compare ♦ with ▪ and □). Cells incubated overnight without peptide and then pulsed immediately with either pVSV or pOVA before assessment were also tested. Lysis of both pVSV- and pOVA-pulsed targets increased relative to unpulsed control, with pOVA-pulsed target cells demonstrating greater sensitization (Fig. 7.C, compare ♦ with ▴ and ▵). Thus, although the lysis levels are greater at 37°C, the results are qualitatively similar at the two temperatures. Furthermore, pVSV, which binds with lower affinity, conferred less effective protection, thus suggesting that peptides with different MHC-I-binding affinities can selectively affect ligand recognition by Ly49C+ LAK.

As a final test of the PR-MHC-I concept, we assessed the sensitivity to LAK lysis of two untransformed cell sources that have either very low or no MHC-I expression and which are thus very sensitive to NK lysis. Specifically, target cells were Con A blasts derived from B6 Tap-1−/− and B6 KbDb−/− mice. Tap-1-deficient tissue, as with RMA-S, is capable of exporting and presenting PR-Kb molecules on the cell surface. High-affinity peptide-filled MHC-I does not occur in these cells due to the absence of an intact peptide transporter complex. In contrast, KbDb−/− mice completely lack PR-Kb as the genes for both H-2Kb and H-2Db are deleted in these animals. Results of measuring their respective sensitivity to lysis by Wt B6 LAK are shown in Fig. 8.A. Both knockout targets were more sensitive to lysis than Wt controls, but KbDb−/− target cells were ∼5-fold more sensitive (LU analysis) than the Tap-1−/− Con A blasts. Pulsing with either pOVA or h-β2m did not change the sensitivity to lysis of the KbDb−/− targets but increased the lysis of the Tap-1−/− ∼5-fold to a level comparable to that seen for the KbDb−/− targets. This is consistent with the fact that Tap-1−/− but not the KbDb−/− targets bear PR-Kb on the cell surface and are thus partially spared from B6 LAK-mediated cytolysis. To functionally map this observation, Ly 49C+ and Ly 49C B6 LAK were tested separately for their ability to kill the Tap-1−/− and KbDb−/− B6 targets (Fig. 8, B and C). Ly49C effectors killed both types of targets equally well, but Ly49C+ were significantly less effective at killing Tap-1−/− than KbDb−/− target cells. This finding is consistent with the model that Ly49C recognizes PR-Kb, which is present on Tap-1−/− targets but not on KbDb−/− targets.

FIGURE 8.

Lysis of two NK-sensitive nontumor targets. Con A blasts were generated from B6 Tap-1−/− or B6 DbKb−/− mice and tested as target cells for B6 LAK effector cells, either unfractionated (A) or sorted into Ly49C+ and Ly49C fractions (B and C). A, Targets were tested with or without added pOVA as in Fig. 1.

FIGURE 8.

Lysis of two NK-sensitive nontumor targets. Con A blasts were generated from B6 Tap-1−/− or B6 DbKb−/− mice and tested as target cells for B6 LAK effector cells, either unfractionated (A) or sorted into Ly49C+ and Ly49C fractions (B and C). A, Targets were tested with or without added pOVA as in Fig. 1.

Close modal

There is general agreement that H-2Kb is a ligand for the NK inhibitory receptor Ly49C, although there remains debate as to the form of H-2Kb that is being recognized. We have presented here data obtained measuring NK function under different conditions capable of altering the relative numbers of different forms of H-2Kb on the cell surface. These data support the conclusion that the peptide-receptive form of H-2Kb (PR-Kb) and not H-2Kb loaded with high-affinity peptide is the relevant ligand for cells expressing normal levels of MHC-I (Con A blasts from normal mice), for cells expressing greatly reduced MHC-I (RMA-S tumor cells and Con A blasts from Tap-1−/− mice), and for cells expressing no MHC-I (KbDb−/− Con A blasts). Our data do not rule out that Ly49C may bind to other forms of H-2Kb, including H-2Kb loaded with high-affinity peptide, only that such an interaction does not lead to the generation of an inhibitory signal. We emphasize that in all our data we have inferred Ly49C recognition through its ability to generate an inhibitory signal.

We found that B6 Con A blasts pulsed briefly with a peptide (pOVA) that binds to H-2Kb with high affinity became sensitive to lysis by syngeneic NK cells. When the NK cells were fractionated on the basis of Ly49C expression, only the Ly49C+ NK cells lysed pOVA-pulsed syngeneic blasts. Blockading effector cell Ly49C with anti-Ly49C mAb instead of pulsing the target cell with pOVA also enabled lysis (Fig. 1.1, Tables I–III, confirming Ref.4). Simultaneously, pulsing with peptide and blockading Ly49C produced no additional lysis (Fig. 1.2). Our interpretation of these data is that Ly49C binds to PR-Kb with sufficient affinity to generate an inhibitory signal and that binding of pOVA to PR-Kb is altering the configuration of the PR-Kb molecule such that, even if weak Ly49C binding can still occur, it can no longer generate an inhibitory signal.

It is important to note that the mAb 5E6, used here to separate Ly49C+ and Ly49C NK cells, can also recognizes Ly49I (29, 45). In a previous study (4), we concluded that Ly49I was not involved in the phenomena described here. 5E6+ cells were sorted into Ly49C+ and Ly49I+ fractions. Neither fraction lysed normal B6 Con A blasts, whereas B6 Con A blasts pulsed with Kb-binding peptide were lysed by the Ly49C+ fraction but not by the Ly49I+ fraction. It was further shown that H-2Db−/− Con A blasts (but not H-2Kb−/− Con A blasts) could bind to COS-7 cells transiently expressing Ly49C and that this binding did not occur when pOVA was added to the medium. COS-7 cells transiently expressing Ly49I could bind neither B6 H-2Db−/− nor B6 H-2Kb−/− Con A blasts in either the presence or the absence of pOVA. Therefore, it seems unlikely that Ly49I is contributing to the effects mapped out in these experiments.

We found that the above results were equally valid for Con A blasts generated in medium containing either FBS or MS. Although when cells were cultured with FBS, exchange of b-β2m for m-β2m could occur on a small fraction of MHC-I molecules (Fig. 1.3, D and E); this appeared to have no effect on the recognition processes (Fig. 1.1, A and B). However, the situation was different for cells cultured in HS. Again, β2m exchange occurred for only a small fraction of all MHC-I molecules. For H-2Kb, exchange occurred predominantly for the ∼10% of all molecules that were peptide receptive (Fig. 4.A). These PR-Kb molecules remained peptide receptive even after binding h-β2m (Fig. 4.B). However, binding of h-β2m made the B6 Con A blast targets equally as sensitive to lysis by Ly49C+ NK cells as either the ablation of PR-Kb by pulsing with pOVA or the blockade of Ly49C with anti-Ly49C mAb (Table II). Our interpretation of these data is that exchange of h-β2m for m-β2m on PR-Kb is altering the conformation of the H-2Kb molecule such that it is no longer recognizable by Ly49C.

As indicated in the introduction, MHC-I molecules on the surface of a B6 Con A lymphoblast can be evenly partitioned on the basis of stability into a long-lived group containing peptide bound with medium to high affinity and a short-lived population, to which PR-MHC-I belongs (12). To maintain the 50:50 steady-state ratio between the two populations, it appears that ∼20 short-lived molecules must be exported to the surface for each long-lived molecule. Biochemical evidence indicates that a newly synthesized MHC-I molecule can be exported from the endoplasmic reticulum (ER) only after the H chain has been properly folded and has associated with both β2m and peptide (46). Only peptides with a correct binding motif can bind MHC-I with high affinity. In general, high-affinity peptides are supplied in the ER via Tap, a pore complex required for the export of peptides from the cytoplasm into the ER. The best evidence for this model comes from studies using RMA-S, which, as a result of a lesion in the gene encoding for Tap-2, lacks functional Tap complexes and therefore exports few stable MHC-I molecules to the cell surface (47). Day et al. (11) have made a detailed comparison of the properties of PR-Kb on RMA-S and its parent line, RMA, which expresses customary levels of MHC-I on the surface. They found that parental RMA expresses three to five times more PR-Kb on the cell surface than RMA-S but that the total rate of production and transport of H-2Kb was approximately the same in the two cell lines. These seemingly paradoxical results can be explained as follows. Both RMA and RMA-S cell lines are assembling MHC-I in the ER at the same rate (11), and these molecules require a peptide to complete assembly. Assuming RMA are similar to Con A blasts (12), only 1 in 20 H-2Kb molecules in RMA will bind a peptide with high affinity (this peptide being provided via Tap). The remainder will bind peptides with lower affinity, provided either via Tap or via a Tap-independent pathway. PR-MHC-I molecules denature rapidly upon reaching the cell surface but some will transiently function as PR-Kb molecules before doing so. RMA-S cells only have access to peptides that reach the ER via a Tap-independent pathway. Given that the total export rate is the same as for RMA, the availability of peptides that can bind is therefore not limiting. However, the average binding affinity of these imperfectly derived peptides is likely even lower than the low-affinity peptides loaded by RMA and, accordingly, the lifetime of these PR-MHC-I on the cell surface even shorter than for RMA. This would explain why Day et al. (11) detected three to five times more PR-Kb on the surface of RMA than on RMA-S.

The effects of peptide pulsing on the lysis of RMA-S (Table IV, Fig. 7) have the same pattern as observed for peptide pulsing of Con A blasts. RMA-S cells were used after overnight culture at 26°C to allow for maximal expression of MHC-I on the cell surface. In accordance with Day et al. (11), we hypothesize that a major portion of this MHC-I will be in the peptide-receptive state. Despite the overnight incubation, these cells were still strongly lysed by NK cells. However, they became ∼5-fold more sensitive to lysis by Ly49C+ NK cells when pulsed with either pOVA or h-β2m immediately before the cytotoxicity assay (Fig. 7.A). Our interpretation is that peptide pulsing is ablating the PR-Kb ligand by converting it into stable H-2Kb bound with high-affinity peptide and that addition of h-β2m is making the PR-Kb ligand unrecognizable by Ly 49C. In support of this, the pOVA pulsing increased the total amount of detectable H-2Kb by ∼5-fold, whereas the h-β2m pulsing produced only a slight increase (Table IV, compare lines 1–3). Note that if Ly49C were recognizing H-2Kb loaded with peptide, then, because by flow cytometric measurement the level of H-2Kb expression went up substantially, the peptide pulsing should have decreased not increased sensitivity to lysis.

RMA-S cells were also tested after overnight culture at 26°C in the presence of pOVA peptide. These cells were significantly less sensitive to lysis by Ly49C+ NK cells than cells incubated overnight without peptide (Fig. 7.A). The incubation produced large increases in both H-2Kb and PR-Kb expression (Table IV). We propose that it is the increased expression of PR-Kb that is responsible for protection because, when these cells were pulsed with pOVA immediately before the cytotoxicity assay, their sensitivity to lysis increased substantially and was now nearly identical to the lysis values of cells not incubated with pOVA overnight but pulsed with pOVA immediately before the cytotoxicity assay (Fig. 7.A). The observations are fully compatible with the PR-Kb model but cannot be explained with the peptide-loaded H-2Kb model. We offer the following possible explanation as to why overnight exposure to pOVA favors generation of PR-Kb. From the data of Day et al. (11), PR-Kb on RMA-S probably has a shorter mean lifetime than on RMA. In the overnight cultures, it is clearly possible that much of the added pOVA will degrade into pieces that will bind with only low affinity but perhaps with an affinity higher than the very poor peptides that can be found in the ER in the absence of Tap. These might well increase the size of the PR population by increasing its mean lifetime.

The above arguments provide an explanation for how cells lacking functional Tap can still express PR-MHC-I. However, cells completely lacking the genes for MHC-I should also completely lack expression of PR-MHC-I. We compared cells from Tap-1−/− and KbDb−/− mice and found that this was indeed true (Fig. 8).

In the B6 Con A lymphoblast system, we found (Table II) that Wt cells pulsed with h-β2m became targets for Ly49C NK cells but that cells from a H-2Db knockout mouse did not. Even though pOVA can bind (but with very low affinity (44)) to H-2Db, pulsing with pOVA had little effect. We conclude that a NK inhibitory receptor that recognizes H-2Db loaded with peptide is likely involved. The same receptor appears to be relevant in the RMS-S system. In the present study, either overnight incubation or brief pulsing with pOVA greatly reduced sensitivity to lysis (Table IV, Fig. 7.B), perhaps by stabilizing enough peptide-loaded H-2Db to serve as ligand for this receptor. In both the RMA-S and Con A blast systems, it appears that pulsing with h-β2m converted H-2Db into an unrecognizable conformation as lysis went up markedly following h-β2m pulsing.

MHC-I tetramer technology has been applied to the visualization of the Ly49-binding repertoire (14, 15, 16). In particular, the H-2Kb-pOVA tetramer has been shown to bind to Ly49C if the β2m component is of mouse origin but not if it is of human origin (16). To test whether such tetramers can contain PR-MHC, we obtained H-2Kb tetramers containing pOVA and h-β2m. We then assessed whether these tetramers contained PR-Kb subcomponents by raising bulk CTL lines specific for H-2Kb-pOVA and H-2Kb-pVSV. As expected, H-2Kb-pOVA tetramer stained the H-2Kb-pOVA-specific CTL but not the H-2Kb-pVSV CTL. However, when an aliquot of the tetramer was incubated with pVSV, it acquired specificity for the H-2Kb-pVSV CTL line. The degree of staining was almost undetectable for a short pulse but increased linearly with time for longer incubation periods (unpublished data). These results imply that the monomer components of the tetramer can form short-lived PR-MHC molecules that can be rescued from decay by binding another peptide such as pVSV. This pVSV is most likely binding to a truly “empty” H-2Kb molecule. The fact that there was almost no binding of pVSV in an initial pulse implies that the truly empty H-2Kb molecule has an extremely short half-life. When these semi-PR tetramers were used to stain Ly49C+ LAK, we found, in agreement with others (16), that they did not bind Ly49C (unpublished data). This observation is consistent with our earlier observation that, although h-β2m can contribute to the formation of PR-MHC-I, the structure of the PR heterodimer is altered such that is it unable to be recognized by Ly49C.

Complexes of Ly49C and H-2Kb containing high-affinity peptide (the same pOVA peptide as we have used here) have been crystallized and characterized (20). This observation is not necessarily incompatible with our concept of PR-Kb being the ligand that triggers Ly49C signaling. Our argument hinges on the possibility that Ly49C can bind to a number of different ligands with a range of different affinities but with only binding events of higher affinity being able to trigger a signaling event. In the same study (20), surface plasmon resonance analysis was used to compare the binding of Ly49C to H-2Kb loaded with pOVA or pVSV, the pVSV also being the same peptide as we have used here. It was found that Ly49C binds much more strongly to pOVA-Kb complexes than to pVSV-Kb complexes. Yet, the interaction of pOVA-Kb with Ly49C was measured to have a KD of only 80 μM. In an interaction between cells, the total Ly49C-Kb interaction strength may be markedly strengthened by the interaction of the Ly49C homodimer on the NK cell with two peptide-loaded H-2Kb molecules on the target cell (20). However, this particular configuration may still prove to have too low an affinity to trigger an inhibitory signal.

It is known that Ly49C possesses a remarkably promiscuous H-2 ligand-binding repertoire, including H-2b, d, f, k, q, r, s, v elements (14, 15, 29, 48), although a functional consequence of Ly49C engagement has only been demonstrated for H-2Kb (4, 22, 48, 49). It is possible that this functional consequence only occurs after binding to PR-Kb. However, “PR-Kb” is an operational definition that likely includes two very different structures, these being H-2Kb complexed with a peptide bound with low affinity and a transitory true “empty” that has lost the low-affinity peptide and will rapidly denature unless stabilized by recapture of another peptide. As argued above, this truly empty form is probably present in extremely low levels such that almost all PR-Kb is likely to be H-2Kb loaded with low-affinity peptide. It is possible that some of these molecules, which may only be slightly different from H-2Kb loaded with high-affinity peptide, will bind to Ly49C with much higher affinity.

PR-Kb is likely to be very heterogeneous. Binding of peptide to H-2Kb with high affinity requires that the peptide be the right length (8 or 9 aa) and have two critical anchor residues (50). Loss of either anchor and/or an improper length could lead to a low-affinity binding event, and the details of how this peptide is bound could affect overall structure. By our measurements, only a subset (∼20%) of all unstable H-2Kb complexes can bind to Ly49C with sufficient affinity to trigger a signaling event (4, 12). Because these MHC-I structures are all unstable, they are unlikely to be caught in a crystal structure.

Mouse NK cells appear to have several inhibitory receptors with specificity for PR-MHC-I. Con A blasts from either B6 or BALB/c become sensitive to lysis after pulsing with a peptide that can bind to H-2Kb, H-2Db, H-2Kd, H-2Dd, or H-2Ld (51). However, to date, Ly49C is the only characterized inhibitory receptor that has been identified as having PR-MHC-I as ligand (4). In contrast, Ly49A, which is expressed on BALB/c NK cells, has been shown to engage H-2Dd complexed with high-affinity peptide (39), although this same study provided evidence that BALB/c NK cells have inhibitory receptors with specificity for both PR-Kd and PR-Dd and that these receptors are present on both Ly49A+ and Ly49A NK cells. Humans may also have NK inhibitory receptors that respond to PR-MHC-I (52). Moreover, Mandelboim et al. (53) showed that RMA-S cells transfected with different alleles of HLA-C were partly protected from lysis by human NK clones that could recognize the particular allele transfected. These authors emphasized that these HLA-C constructs were being recognized in the absence of high-affinity peptide binding.

The PR-MHC model of Ly49C recognition suggests a mechanism whereby NK cells may be more effective in host defense. PR-MHC is present on the cell surface under normal physiological conditions, and moreover, its expression is highly responsive to cellular fitness (54). Accordingly, following disturbances in normal cell physiology, such as viral infection or neoplastic transformation, that lead to a reduction in protein transport, PR-MHC-I will decrease much more rapidly than total MHC-I. A NK inhibitory receptor that is specific for this ligand rather than a more stable form of MHC-I loaded with high-affinity peptide will lose inhibitory ligand much more rapidly, thus leading to reactivity against the affected cell. Therefore, the PR-MHC-I model represents an extension of the “missing self” hypothesis in which it is a particular conformation of MHC-I and not just total MHC-I that becomes the self-ligand of importance.

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.

1

This work was supported by funding to R.G.M from the National Cancer Institute of Canada and to L.A.B. from the Ontario Graduate Scholarship Program.

3

Abbreviations used in this paper: MHC-I, MHC class I; β2m, β2-microglobulin; PR, peptide receptive; h-β2m, human β2m; b-β2m, bovine β2m; m-β2m, mouse β2m; MS, mouse serum; HS, human serum; BFA, brefeldin A; LAK, lymphokine-activated killer; Wt, wild type; poly I:C, polyinosine-polycytidylic acid; MFI, mean fluorescent intensity; ER, endoplasmic reticulum.

1
Tomasello, E., M. Blery, E. Vely, E. Vivier.
2000
. Signaling pathways engaged by NK cell receptors: double concerto for activating receptors, inhibitory receptors and NK cells.
Semin. Immunol.
12
:
139
.-147.
2
Kärre, K., H. G. Ljunggren, G. Piontek, R. Kiessling.
1986
. Selective rejection of H-2-deficient lymphoma variants suggests alternative immune defense strategy.
Nature
319
:
675
.-678.
3
Natarajan, K., N. Dimasi, J. Wang, R. A. Mariuzza, D. H. Margulies.
2002
. Structure and function of natural killer cell receptors: multiple molecular solutions to self, nonself discrimination.
Annu. Rev. Immunol.
20
:
853
.-885.
4
Su, R.-C., S. K. Kung, E. T. Silver, S. Lemieux, K. P. Kane, R. G. Miller.
1999
. Ly-49CB6 NK inhibitory receptor recognizes peptide-receptive H-2Kb.
J. Immunol.
163
:
5319
.-5330.
5
Cook, J. R., N. B. Myers, T. H. Hansen.
1996
. The mechanisms of peptide exchange and β2-microglobulin exchange on cell surface Ld and Kb molecules are noncooperative.
J. Immunol.
1157
:
2256
.-2261.
6
Bernabeu, C., M. van de Rijn, P. G. Lerch, C. P. Terhorst.
1984
. β2-Microglobulin from serum associates with MHC class I antigens on the surface of cultured cells.
Nature
308
:
642
.-645.
7
Yu, Y. Y., N. B. Myers, C. M. Hilbert, M. R. Harris, G. K. Balendiran, T. H. Hansen.
1999
. Definition and transfer of a serological epitope specific for peptide-empty forms of MHC class I.
Int. Immunol.
11
:
1897
.-1906.
8
De Silva, A. D., A. Boesteanu, R. Song, N. Nagy, E. Harhaj, C. V. Harding, S. Joyce.
1999
. Thermolabile H-2Kb molecules expressed by transporter associated with antigen processing-deficient RMA-S cells are occupied by low-affinity peptides.
J. Immunol.
163
:
4413
.-4420.
9
Bouvier, M., D. C. Wiley.
1998
. Structural characterization of a soluble and partially folded class I major histocompatibility heavy chain/β2m heterodimer.
Nat. Struct. Biol.
5
:
377
.-384.
10
Zacharias, M., S. Springer.
2004
. Conformational flexibility of the MHC class I α1-α2 domain in peptide bound and free states: a molecular dynamics simulation study.
Biophys. J.
87
:
2203
.-2214.
11
Day, P. M., F. Esquivel, J. Lukszo, J. R. Bennink, J. W. Yewdell.
1995
. Effect of TAP on the generation and intracellular trafficking of peptide-receptive major histocompatibility complex class I molecules.
Immunity
2
:
137
.-147.
12
Su, R.-C., R. G. Miller.
2001
. Stability of surface H-2Kb, H-2Db, and peptide-receptive H-2Kb on splenocytes.
J. Immunol.
167
:
4869
.-4877.
13
Shields, M., L. Moffat, R. Ribaudo.
1998
. Functional comparison of bovine, murine and human β2m: interactions with murine MHC-I molecules.
Mol. Immunol.
35
:
919
.-928.
14
Hanke, T., H. Takizawa, C. W. McMahon, D. H. Busch, E. G. Pamer, J. D. Miller, J. D. Altman, Y. Liu, D. Cado, F. A. Lemonnier, P. J. Bjorkman, D. H. Raulet.
1999
. Direct assessment of MHC class I binding by seven Ly49 inhibitory NK cell receptors.
Immunity
11
:
67
.-77.
15
Makrigiannis, A. P., A. T. Pau, A. Saleh, R. Winkler-Pickett, J. R. Ortaldo, S. K. Anderson.
2001
. Class I MHC-binding characteristics of the 129/J Ly49 repertoire.
J. Immunol.
166
:
5034
.-5043.
16
Michaëlsson, J., A. Achour, A. Rolle, K. Kärre.
2001
. MHC class I recognition by NK receptors in the Ly49 family is strongly influenced by the β2-microglobulin subunit.
J. Immunol.
166
:
7327
.-7334.
17
Matsumoto, N., M. Mitsuki, K. Tajima, W. M. Yokoyama, K. Yamamoto.
2001
. The functional binding site for the C-type lectin-like natural killer cell receptor Ly49A spans three domains of its major histocompatibility complex class I ligand.
J. Exp. Med.
193
:
147
.-158.
18
Wang, J., M. C. Whitman, K. Natarajan, J. Tormo, R. A. Mariuzza, D. H. Margulies.
2002
. Binding of the natural killer cell inhibitory receptor Ly49A to its major histocompatibility complex class I ligand: crucial contacts include both H-2Dd and β2-microglobulin.
J. Biol. Chem.
277
:
1433
.-1442.
19
Mitsuki, M., N. Matsumoto, K. Yamamoto.
2004
. A species-specific determinant on β2-microglobulin required for Ly49A recognition of its MHC class I ligand.
Int. Immunol.
16
:
197
.-204.
20
Dam, J., R. Guan, K. Natarajan, N. Dimasi, L. K. Chlewicki, D. M. Kranz, P. Schuck, D. H. Margulies, R. A. Mariuzza.
2003
. Variable MHC class I engagement by Ly49 natural killer cell receptors demonstrated by the crystal structure of Ly49C bound to H-2Kb.
Nat. Immunol.
4
:
1213
.-1222.
21
Tormo, J., K. Natarajan, D. H. Margulies, R. A. Mariuzza.
1999
. Crystal structure of a lectin-like natural killer cell receptor bound to its MHC class I ligand.
Nature
402
:
623
.-631.
22
Franksson, L., J. Sundback, A. Achour, J. Bernlind, R. Glas, K. Kärre.
1999
. Peptide dependency and selectivity of the NK cell inhibitory receptor Ly-49C.
Eur. J. Immunol.
29
:
2748
.-2758.
23
Takai, T., M. Li, D. Sylvestre, R. Clynes, J. Ravetch.
1994
. FcR γ chain deletion results in pleiotrophic effector cell defects.
Cell
76
:
519
.-529.
24
Tanchot, C., F. A. Lemonnier, B. Perarnau, A. A. Freitas, B. Rocha.
1997
. Differential requirements for survival and proliferation of CD8 naive or memory T cells.
Science
276
:
2057
.-2062.
25
Chamberlain, J. W., J. A. Nolan, P. J. Conrad, H. A. Vasavada, H. H. Vasavada, H. L. Ploegh, S. Ganguly, C. A. Janeway, Jr, S. M. Weissman.
1988
. Tissue-specific and cell surface expression of human major histocompatibility complex class I heavy (HLA-B7) and light (β2-microglobulin) chain genes in transgenic mice.
Proc. Natl. Acad. Sci. USA
85
:
7690
.-7694.
26
Zijlstra, M., M. Bix, N. E. Simister, J. M. Loring, D. H. Raulet, R. Jaenisch.
1990
. β2-Microglobulin deficient mice lack CD4–8+ cytolytic T cells.
Nature
344
:
742
.-746.
27
Rötzschke, O., K. Falk, S. Stevanovie, G. Jung, P. Walden, H. G. Rammensee.
1991
. Exact prediction of a natural T cell epitope.
Eur. J. Immunol.
21
:
2891
.-2894.
28
Van Bleek, G. M., S. G. Nathenson.
1990
. Isolation of an endogenously processed immunodominant viral peptide from the class I H-2Kb molecule.
Nature
348
:
213
.-216.
29
Stoneman, E. R., M. Bennett, J. An, K. A. Chesnut, E. K. Wakeland, J. B. Scheerer, M. J. Siciliano, V. Kumar, P. A. Mathew.
1995
. Cloning and characterization of 5E6 (Ly49C), a receptor molecule expressed on a subset of murine natural killer cells. [Published erratum appears in 1996 J. Exp. Med. 183: 2705.].
J. Exp. Med.
182
:
305
.-313.
30
Kuhns, S. T., L. R. Pease.
1998
. A region of conformational variability outside the peptide-binding site of a class I MHC molecule.
J. Immunol.
161
:
6745
.-6750.
31
Jones, B., C. A. Janeway, Jr.
1981
. Cooperative interaction of B lymphocytes with antigen-specific helper T lymphocytes is MHC restricted.
Nature
292
:
547
.-549.
32
Bluestone, J. A., S. Jameson, S. Miller, R. Dick 2nd.
1992
. Peptide-induced conformational changes in class I heavy chains alter major histocompatibility complex recognition.
J. Exp. Med.
176
:
1757
.-1761.
33
Loken, M. R., A. M. Stall.
1982
. Flow cytometry as an analytical and preparative tool in immunology.
J. Immunol. Methods
50
:
R85
.-R112.
34
Hasenkrug, K. J., J. M. Cory, J. H. Stimpfling.
1987
. Monoclonal antibodies defining mouse tissue antigens encoded by the H-2 region.
Immunogenetics
25
:
136
.-139.
35
Lilic, M., Z. Popmihajlov, J. J. Monaco, S. Vukmanovic.
2004
. Association of β2-microglobulin with the α3 domain of H-2Db heavy chain.
Immunogenetics
55
:
740
.-747.
36
Jefferies, W. A., G. G. MacPherson.
1987
. Expression of the W6/32 HLA epitope by cells of rat, mouse, human and other species: critical dependence on the interaction of specific MHC heavy chains with human or bovine β2-microglobulin.
Eur. J. Immunol.
17
:
1257
.-1263.
37
Tada, N., S. Kimura, A. Hatzfeld, U. Hammerling.
1980
. Ly-m11: the H-3 region of mouse chromosome 2 controls a new surface alloantigen.
Immunogenetics
11
:
441
.-449.
38
Porgador, A., J. Yewdell, Y. Deng, J. Bennink, R. Germain.
1997
. Localization, quantitation and in situ detection of specific peptide-MHC Class I complexes using a monoclonal antibody.
Immunity
6
:
715
.-726.
39
Su, R.-C., S. K. Kung, J. Gariepy, B. H. Barber, R. G. Miller.
1998
. NK cells can recognize different forms of class I MHC.
J. Immunol.
161
:
755
.-766.
40
Sheng-Tanner, X., R. G. Miller.
1992
. Correlation between lymphocyte-induced donor-specific tolerance and donor cell recirculation.
J. Exp. Med.
176
:
407
.-413.
41
Djeu, J. Y., J. A. Heinbaugh, H. T. Holden, R. B. Herberman.
1979
. Augmentation of mouse natural killer cell activity by interferon and interferon inducers.
J. Immunol.
122
:
175
.-181.
42
Ljunggren, H. G., L. Van Kaer, H. L. Ploegh, S. Tonegawa.
1994
. Altered natural killer cell repertoire in Tap-1 mutant mice.
Proc. Natl. Acad. Sci. USA
91
:
6520
.-6524.
43
Ljunggren, H. G., N. J. Stam, C. Öhlén, J. J. Neefjes, P. Höglund, M. T. Heemels, J. Bastin, T. N. Schumacher, A. Townsend, K. Kärre.
1990
. Empty MHC class I molecules come out in the cold.
Nature
346
:
476
.-480.
44
Deres, K., W. Beck, S. Faath, G. Jung, H. G. Rammensee.
1993
. MHC/peptide binding studies indicate hierarchy of anchor residues.
Cell. Immunol.
151
:
158
.-167.
45
Brennan, J., S. Lemieux, J. D. Freeman, D. L. Mager, F. Takei.
1996
. Heterogeneity among Ly-49C natural killer (NK) cells: characterization of highly related receptors with differing functions and expression patterns.
J. Exp. Med.
184
:
2085
.-2090.
46
Rock, K. L., C. Gramm, L. Rothstein, K. Clark, R. Stein, L. Dick, D. Hwang, A. L. Goldberg.
1994
. Inhibitors of the proteasome block the degradation of most cell proteins and the generation of peptides presented on MHC class I molecules.
Cell
78
:
761
.-771.
47
Powis, S. J., A. R. M. Townsend, E. V. Deverson, J. Bastin, G. W. Butcher, J. C. Howard.
1991
. Restoration of antigen presentation to the mutant cell line RMA-S by an MHC-linked transporter.
Nature
354
:
528
.-531.
48
Brennan, J., G. Mahon, D. L. Mager, W. A. Jefferies, F. Takei.
1996
. Recognition of class I major histocompatibility complex molecules by Ly-49: specificities and domain interactions.
J. Exp. Med.
183
:
1553
.-1559.
49
Yu, Y. Y., T. George, J. R. Dorfman, J. Roland, V. Kumar, M. Bennett.
1996
. The role of Ly49A and 5E6(Ly49C) molecules in hybrid resistance mediated by murine natural killer cells against normal T cell blasts.
Immunity
4
:
67
.-76.
50
Rammensee, H. G..
1995
. Chemistry of peptides associated with MHC class I and class II molecules.
Curr. Opin. Immunol.
7
:
85
.-96.
51
Chadwick, B. S., R. G. Miller.
1992
. Hybrid resistance in vitro: possible role of both class I MHC and self peptides in determining the level of target cell sensitivity.
J. Immunol.
148
:
2307
.-2313.
52
Storkus, W. J., R. D. Salter, P. Cresswell, J. R. Dawson.
1992
. Peptide-induced modulation of target cell sensitivity to natural killing.
J. Immunol.
149
:
1185
.-1190.
53
Mandelboim, O., H. T. Reyburn, M. Vales-Gomez, L. Pazmany, M. Colonna, G. Borsellino, J. L. Strominger.
1996
. Protection from lysis by natural killer cells of group 1 and 2 specificity is mediated by residue 80 in human histocompatibility leukocyte antigen C alleles and also occurs with empty major histocompatibility complex molecules.
J. Exp. Med.
184
:
913
.-922.
54
Ortiz-Navarrete, V., G. J. Hammerling.
1991
. Surface appearance and instability of empty H-2 class I molecules under physiological conditions.
Proc. Natl. Acad. Sci. USA
88
:
3594
.-3597.