Killer Cell Immunoglobulin Receptors and T Cell Receptors Bind ...

THE JOURNAL OF BIOLOGICAL CHEMISTRY © 1999 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 274, No. 40, Issue of October 1, pp. 28329 –28334, 1999 Printed in U.S.A.

Killer Cell Immunoglobulin Receptors and T Cell Receptors Bind Peptide-Major Histocompatibility Complex Class I with Distinct Thermodynamic and Kinetic Properties* (Received for publication, June 16, 1999)

Katsumi Maenaka‡§¶, Takeo Juji§, Takahiro Nakayama§, Jessica R. Wyeri, George F. Gaoi, Taeko Maenaka‡, Nathan R. Zaccai‡, Akiko Kikuchi§, Toshio Yabe§, Katsushi Tokunaga§, Kenji Tadokoro§, David I. Stuart‡**‡‡, E. Yvonne Jones‡**§§, and P. Anton van der Merwe‡‡¶¶ From the ‡Structural Biology, Wellcome Trust Centre for Human Genetics, Roosevelt Drive, Headington, Oxford OX3 7BN, United Kingdom, §Japanese Red Cross Central Blood Center, 4-3-1 Hiro-o, Shibuya-ku, Tokyo 120, Japan, iMRC Human Immunology Unit, Institute of Molecular Medicine, University of Oxford, John Radcliffe Hospital, Oxford OX3 9DS, the **Oxford Center for Molecular Sciences, New Chemistry Building, South Parks Road, Oxford OX1 3QT, and the ¶¶Sir William Dunn School of Pathology, University of Oxford, Oxford OX1 3RE, United Kingdom

Human natural killer cells and a subset of T cells express a repertoire of killer cell immunoglobulin receptors (KIRs) that recognize major histocompatibility complex (MHC) class I molecules. KIRs and T cell receptors (TCRs) bind in a peptide-dependent manner to overlapping regions of peptide-MHC class I complexes. KIRs with two immunoglobulin domains (KIR2Ds) recognize distinct subsets of HLA-C alleles. Here we use surface plasmon resonance to study the binding of soluble forms of KIR2DL1 and KIR2DL3 to several peptideHLA-Cw7 complexes. KIR2DL3 bound to the HLA-Cw7 allele presenting the peptide RYRPGTVAL with a 1:1 stoichiometry and an affinity (Kd ;7 mM at 25 °C) within the range of values measured for other cell-cell recognition molecules, including the TCR. Although KIR2DL1 is reported not to recognize the HLA-Cw7 allele in functional assays, it bound RYRPGTVAL/HLA-Cw7, albeit with a 10 –20-fold lower affinity. TCR/peptide-MHC interactions are characterized by comparatively slow kinetics and unfavorable entropic changes (Willcox, B. E., Gao, G. F., Wyer, J. R., Ladbury, J. E., Bell, J. I., Jakobsen, B. K., and van der Merwe, P. A. (1999) Immunity 10, 357–365), suggesting that binding is accompanied by conformational adjustments. In contrast, we show that KIR2DL3 binds RYRPGTVAL/HLA-Cw7 with fast kinetics and a favorable binding entropy, consistent with rigid body association. These results indicate that KIR/ peptide-MHC class I interactions have properties typical of other cell-cell recognition molecules, and they highlight the unusual nature of TCR/peptide-MHC recognition.

* This work was supported in part by a Grant-in-aid for Scientific Research on Priority Areas 2782 from the Ministry of Education, Science, and Culture of Japan. 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. ¶ Supported in part by a JSPS Research Fellowship for Young Scientists and by a Human Frontier Science Program long term fellowship. To whom correspondence should be addressed: Structural Biology, Wellcome Trust Centre for Human Genetics, Roosevelt Dr., Headington, Oxford OX3 7BN, UK. Tel.: 44-1865-287550; Fax: 144-1865287547; E-mail: [email protected]. ‡‡ Supported by the MRC. §§ Supported by the Royal Society. To whom correspondence should be addressed: Structural Biology, Wellcome Trust Centre for Human Genetics, Roosevelt Dr., Headington, Oxford OX3 7BN, UK. Tel.: 441865-287559; Fax: 44-1865-287547; E-mail: [email protected]. This paper is available on line at http://www.jbc.org

NK1 cells and CD8 T lymphocytes have complementary roles in the cellular immune response. Whereas CD8 lymphocytes kill cells presenting non-self peptides on MHC class I molecules, NK cells kill cells deficient in MHC class I molecules (1–3). In so doing they make it difficult for intracellular pathogens to evade the immune response by interfering with the expression of MHC class I molecules. It is thought that NK cells are stimulated to kill somatic cells by ligation of one or more activatory receptors but are held in check if inhibitory receptors are able to bind MHC class I molecules on these cells. In humans a family of killer cell Ig receptors (KIRs) has been identified on NK cells, and a subset of T cells, that bind to MHC class I molecules (4). The KIR genes are clustered on chromosome 19q13.4 (5, 6). The precise number (;10) appears to vary between individuals (7). These characteristics, and the absence of homologous genes in rodents, suggest that the KIR gene family evolved fairly recently, perhaps driven by the rapid evolution of MHC class I molecules (8). KIRs have either two (KIR2D) or three (KIR3D) Ig domains in the extracellular region (9). They can be further grouped according to whether they have a short (e.g. KIR2DS) or long (e.g. KIR2DL) cytoplasmic tail (9). The long cytoplasmic tails contain immunoreceptor tyrosine-based inhibitory motifs and transduce inhibitory signals (10). In contrast, the short cytoplasmic tails mediate association with DAP12 (11), which contains immunoreceptor tyrosine-based activation motifs and transduces activatory signals. The KIR2D receptors bind HLA-C alleles, whereas KIR3D receptors bind HLA-A and -B alleles (12, 13). Like the TCR, KIRs bind to the peptide-presenting platform on MHC class I molecules (14, 15). The KIR2D-binding site on HLA-C includes residues in the carboxyl-terminal half of the a1-helix (residues 73, 76 and 80) and the adjacent b-sheet (residue 90). KIR2DL1 binds preferentially to group 1 HLA-C alleles (Cw2, Cw4, Cw5, Cw6, Cw15), which have Lys at position 80, whereas KIR2DL2 and KIR2DL3 bind preferentially to group 2 HLA-C alleles (Cw1, Cw3, Cw7, Cw8), which have Asn at position 80. However, there appears to be some cross-reactivity between these groups (16 –18). KIR recognition of peptide-MHC class I has been shown to depend to some extent on the peptide presented on the MHC molecule (17, 19 –22). These observations suggest 1 The abbreviations used are: NK, natural killer; KIR, killer cell immunoglobulin receptor; KIR2D, KIR with 2 Ig domains; MHC, major histocompatibility complex; TCR, T cell receptor; Ni-NTA, nickel charged nitrilotriacetic acid.

28329

28330

Kinetics and Thermodynamics of KIR/Peptide-MHC Interactions

that a single KIR will bind different peptide-MHC complexes with different affinities, raising the question as to what the affinity and/or kinetic threshold is for functional recognition? Recently Vale´s-Go´mez et al. (23, 24) used surface plasmon resonance to measure the affinity of KIR2DL1 and KIR2DL3 binding HLA-Cw6-peptide and HLA-Cw7-peptide complexes, respectively. The dissociation rate constant (koff) was also estimated, but the association rate constant (kon) could not be measured directly (23, 24). We extend this study by providing more precise kinetic measurements of the KIR2DL3/peptideHLA-Cw7 interaction, obtaining thermodynamic data, and determining the stoichiometry of the interaction. We also use affinity measurements to quantitate the effects of the peptide on the binding affinity and the degree of cross-reactivity between KIR2DL1 and HLA-Cw7. We find that KIR2DL3 binds peptide-HLA-Cw7 with a 1:1 stoichiometry and with thermodynamic and kinetic properties very similar to other cell-cell recognition molecules. These properties differ from those reported for TCR/peptide-MHC interactions, underlining the unusual nature of TCR recognition. EXPERIMENTAL PROCEDURES

Production of Soluble Forms of KIR2DL1 and KIR2DL3—DNA encoding the extracellular portions (residues 1–224) of KIR2DL1 and KIR2DL3 were amplified from cDNA, obtained as described previously (6), using 59-G GCC ATG GCA CAT GAG GGA GTC CAC-39 as forward primer and 59-C AGC GGC CGC GTG CAG GTG TCG GGG GTT ACC-39 as reverse primer. The resultant fragments were digested with the restriction enzymes NcoI and NotI and ligated into a derivative of pGEM2 (Promega). The final construct encodes, in tandem, the ShineDalgano and signal peptide sequence of pelB, the extracellular portion of KIR2DL1 or KIR2DL3, a c-myc epitope and an oligohistidine tag (HHHHHH), all between the HindIII and EcoRI restriction sites of pGEM2 and under the control of a T7 promoter. The resulting expression plasmid was designated pKMATHNK1 or -2. Escherichia coli strain BL21(DE3)pLysS cells (Novagen) harboring pKMATHNK1 or -2 were grown at 37 °C in 23 YT medium containing ampicillin (100 mg/ml) and chloramphenicol (34 mg/ml), induced with 0.1 mM isopropylb-D-thiogalactopyranoside in logarithmic phase (A280 0.6 – 0.8), and then incubated for 17 h at 27 °C. Recombinant protein secreted into the periplasmic space and media (yield approximately 0.2 mg/l) were concentrated by ammonium sulfate precipitation and purified by metalchelate affinity chromatography (Ni-NTA superflow, Qiagen) followed by ion exchange chromatography (MonoQ, Amersham Pharmacia Biotech). A second expression system was established to increase the yield of KIR2Ds. The genes of the extracellular portions (residues 1–224) of KIR2DL1 and KIR2DL3 were amplified using 59-GGAACATATGCACGAGGGAG TCCACAG-39 as forward primer and 59-CGGAGGCTTACTAATGC AGGTGTCTGGGGTTAC-39 as reverse primer. The resultant fragments were digested with the restriction enzymes NdeI and HindIII and ligated into pGMT7 (25), creating the plasmids pGMNK1 and pGMNK2 encoding KIR2DL1 and KIR2DL3, respectively. The plasmids were expressed in the E. coli strain BL21(DE3) pLysS. Recombinant proteins, which accumulate as insoluble aggregates in inclusion bodies, were refolded and purified by the method of Reid et al. (25). In some cases proteins were further purified by ion exchange chromatography (ResourceQ, Amersham Pharmacia Biotech). Production of Soluble HLA Molecules—A DNA fragment encoding the extracellular portion of HLA-Cw0702 heavy chain (residues 1–276) was amplified from cDNA kindly provided by Dr. H. Wang, using 59-CCCACACATATGGGATCCCACTCCATG AGGTATTTCGAC-39 as forward primer and 59-CCCACAAAGCTTCTATCATGGCTCCCAGCTCAGGGTGAGGGG-39 as reverse primer. The resultant fragment was digested with the restriction enzymes NdeI and HindIII and ligated into pGMT7 (25). HLA-Cw7 was recovered from inclusion bodies, refolded with peptide and b2-microglobulin, and purified using the method of Reid et al. (25). One of the following peptides was used: KYFDEHYEY (DS-10), RYRPGTVAL (DS11), or NKADVILKY (DS12). Biotinylated HLA-Cw7 was prepared by refolding in the presence of biotinylated b2-microglobulin (26, 27). HLA-A2 (with peptide ILKEPVHGV), HLAB35 (TPEGIIPTL), HLA-E (VMAPRTVLL), and HLA-G1 (RIIPRHLQL) were produced in the same way. Surface Plasmon Resonance—Surface plasmon resonance experi-

ments were performed using a BIAcoreTM 2000 (BIAcore AB, St. Albans, UK). All the experiments were performed at 25 °C unless otherwise indicated using HBS-EP (10 mM Hepes (pH 7.4), 150 mM NaCl, 3.4 mM EDTA, and 0.005% Surfactant P20) as running buffer when using CM5 sensor chips and HBS-P (HBS-EP without EDTA) when using Ni-NTA sensor chips (BIAcore AB). Biotinylated proteins were immobilized via streptavidin, which was covalently coupled to CM5 research grade sensor chips as described previously (28). Proteins with oligohistidine tags (sKIR2DL3H and sKIR2DL1H) were immobilized onto Ni-NTA sensor chips (BIAcore AB). Ni-NTA surfaces were regenerated by injection of 0.35 M EDTA (pH 8.3) for 1 min to elute bound protein followed by 0.5 M NiCl2 for 1 min to recharge the NTA with nickel. Kinetic constants were derived using the curve-fitting facility of the BIAevaluation program (version 3.0, BIAcore) that deploys the Marquardt-Levenberg algorithm. Rate equations were derived from the simple 1:1 Langmuir binding model (A 1 B 7 AB). Other curve fitting was performed in Origin version 3 (MicroCal). Affinity constants were derived by Scatchard analysis or by non-linear curve fitting of the standard Langmuir binding isotherm. Thermodynamic data were obtained by fitting to the data in Fig. 4A the non-linear form of the van’t Hoff Equation (29),

DG0 5 DHTo 2 T z DS0To 1 DCp~T 2 To! 1 T z DCp z ln

SD T To

(Eq. 1)

where T is the temperature in Kelvin (K); To is an arbitrary reference temperature (e.g. 298.15 K); DG0 is the standard free energy of binding at T (kcalzmol21) and is calculated from the Kd; DHTo is the enthalpy change upon binding at To (kcalzmol21); DS0To is the standard state entropy change upon binding at To (kcalzmol21), and DCp is the specific heat capacity (kcalzmol21zK21), and is assumed to be temperature-independent. Gel Filtration—sKIR2DL3 (60 ml at 1 mM) and HLA-Cw7-DS11 (40 ml at 1 mM) were mixed and incubated for 1 h at room temperature before separation by fast protein liquid chromatography on a Superdex 200 (Amersham Pharmacia Biotech) in 20 mM Tris-Cl (pH 8) at a flow rate of 0.4 ml/min. RESULTS

Stoichiometry and Peptide Dependence of sKIR2DL3 Binding to HLA-Cw7—Soluble MHC class I heavy chains were expressed in bacteria and refolded in vitro together with peptide and b2-microglobulin (see “Experimental Procedures”). Biotinylated peptide-MHC class I complexes were produced by refolding with chemically biotinylated b2-microglobulin (28). Soluble forms of KIR2D molecules were also expressed in bacteria, either with (sKIR2DL1H and sKIR2DL3H) or without (sKIR2DL1 and sKIR2DL3) carboxyl-terminal c-myc and oligohistidine tags (see “Experimental Procedures”). sKIR2DL3H was used successfully for an x-ray crystallographic structure determination (30), indicating that it is correctly folded. Purified sKIRs migrated as monomers on size exclusion chromatography (Fig. 1B and data not shown). Binding of KIR constructs was analyzed by surface plasmon resonance, which measures the changes in refractive index near a sensor surface (31). sKIR2DL3 was injected through 4 flow cells containing sensor surfaces to which different peptideMHC class I complexes had been immobilized using their biotin tag (Fig. 1A, solid bar). A “background” response (measured in response units) is seen in the negative control HLA-A2 flow cell, a consequence of the high concentration, and therefore high refractive index, of injected sKIR2DL3 sample. However, a greater response is seen with injection over HLA-Cw7 complexed with the DS11 peptide, indicating binding (Fig. 1A). sKIR2DL3 bound at a much lower level to HLA-Cw7 complexed with the DS12 peptide and did not bind at all to the HLA-Cw7DS10 complex (Fig. 1A). In order to assess the stoichiometry of binding, HLA-Cw7DS11 and sKIR2DL3 were mixed together, with a molar excess of sKIR2DL3 (1:1.5), and then fractionated by size exclusion chromatography (Fig. 1B). The elution position of the complex


28331

FIG. 2. Affinity of KIR molecules binding to immobilized HLACw7-DS11. A, sKIR2DL3 was injected for 30 s at the indicated concentrations through flow cells with HLA-Cw7. B, a plot of the equilibrium binding response of sKIR2DL3 (●, derived from A) and sKIR2DL1 (Œ, raw data not shown) versus concentration. The solid lines represent direct non-linear fits of the 1:1 Langmuir binding isotherm to the data, which yielded the indicated Kd values. Inset, Scatchard plots of the same data are shown. The solid lines are linear fits, which yielded Kd estimates of 8 and 84 mM for sKIR2DL3 and sKIR2DL1, respectively. RU, response units. TABLE I Summary of affinity constants Injected, Kd at 25 °C (mM)

Immobilized

FIG. 1. Peptide dependence and stoichiometry of sKIR2DL3 binding to HLA-Cw7. A, sKIR2DL3 (106 mM) was injected at a flow rate of 10 ml/min (solid bar) through flow cells with the indicated peptide-MHC class I complex immobilized to the sensor surface. HLACw7 was complexed with DS10, DS11, or DS12 peptides (see “Experimental Procedures”). B, size exclusion chromatography of sKIR2DL3/ HLA-Cw7-DS11 complex. A 100-ml mixture of sKIR2DL3 (0.6 mM) and HLA-Cw7-DS11 (0.4 mM) was fractionated on a Sephadex 200 column. The elution positions are shown of protein molecular mass standards (in kDa). A calibration curve based on these standards was used to estimate the Mr of sKIR2DL3 and the sKIR2DL3-HLA-Cw7-DS11 complex. Also shown is the expected elution position of free HLA-Cw7DS11, which was determined on a separate run (data not shown). RU, response units.

(;74 kDa, calculated Mr 69,304) is consistent with the presence of one sKIR2DL3 molecule and one HLA-Cw7-DS11 molecule. Furthermore, whereas free sKIR2DL3 was present, no free HLA-Cw7-DS11 was detected, indicating that there was an excess of the sKIR2DL3 (Fig. 1B). Taken together, these data indicate that sKIR2DL3 binds HLA-Cw7-DS11 with a 1:1 stoichiometry. Affinity of sKIR2DL3 and sKIR2DL1 Binding to HLA-Cw7 Peptide—The affinity of sKIR2Ds binding to peptide-MHC molecules was measured by equilibrium binding analysis on the BIAcore. A range of concentrations of sKIR2DL3 (Fig. 2A) was injected through flow cells with HLA-Cw7-DS11 or a control peptide-MHC class I complex immobilized. The binding response at each concentration was calculated by subtracting the equilibrium response measured in the control flow cell from

HLA-Cw7-DS10a HLA-Cw7-DS11a HLA-Cw7-DS12a HLA-A2c HLA-B35c HLA-Ec HLA-G1c

sKIR2DL3H

sKIR2DL1

sKIR2DL3

NBb 117 6 12 (n 5 2) NB NB NB NB NB

NB 6.9 6 2.4 (n 5 4) 115 (n 5 1) NB NB NB NB

HLA-Cw7-DS10

HLA-Cw7-DS1

HLA-Cw7-DS12

NB

4 (n 5 1)

115 (n 5 1)

* Mean 6 S.D. (n . 2) or mean 6 range (n 5 2). DS10, KYFDEHYEY; DS11, RYRPGTVAL; DS12, NKADVILKY. NB, no binding detected with indicated molecule injected at $3 mM. c See “Experimental Procedures” for peptide sequence. a b

response in the HLA-Cw7-DS11 flow cell. Conventional (Fig. 2B) and Scatchard (Fig. 2B, inset) plots of these binding data indicate that the interaction conforms to a simple 1:1 (Langmuir) binding model with a Kd of ;9 mM. The results of several experiments are summarized in Table I. Other soluble recombinant forms of sKIR2DL3 bound immobilized HLA-Cw7-DS11 with a similar affinity (data not shown). These included sKIR2DL3H and a truncated version of sKIR2DL3 (comprising amino acids 1–200) which lacked the membrane-proximal stalk region. A similar affinity was measured in the reverse orientation, with sKIR2DL3H immobilized to a Ni21-NTA sensor chip via its oligohistidine tag, and HLA-Cw7-DS11 in solution (Table I). The weak interaction between sKIR2DL3 and HLA-Cw7-

28332


DS12 peptide (Fig. 1A) was confirmed by affinity analysis (data not shown). sKIR2DL3 bound to HLA-Cw7-DS12 with an affinity (Kd ;108 mM) ;12-fold lower than its affinity for HLACw7-DS11 (Fig. 2B). This very low affinity was confirmed in the reverse orientation, with sKIR2DL3H immobilized and HLA-Cw7/DS12 in solution. sKIR2DL1 also bound to HLA-Cw7-DS11, albeit with a 10 – 15-fold lower affinity than sKIR2DL3 (Fig. 2B and Table I). This lower affinity was largely a consequence of a faster koff (data not shown), indicating that it is not a consequence of low sKIR2DL1 activity. No binding was detected when high concentrations (up to 3 mM) of sKIR2DL1 and sKIR2DL3 were injected over several other classical (HLA-A2 and -B35) and non-classical (HLA-E and -G1) MHC class I molecules (Table I). Binding Kinetics—Although binding and dissociation were

very fast, it was possible to analyze both the association and dissociation phases of binding (Fig. 3). Global fitting with mono-exponential rate equations derived from the simple 1:1 Langmuir binding model produced reasonable fits, yielding a kon of 1.6 3 105 M21 s21 and a koff of 1.2 s21 (Fig. 3A). The rate constants did not change significantly when the level of immobilized HLA-Cw7-DS11 varied 2-fold (Table II), indicating that binding was not substantially affected by mass transport or rebinding artifacts. The excellent agreement between calculated Kd (Table II) and the Kd determined by equilibrium binding (Table I) further supports the notion that these kinetic constants are correct. Because recent studies have shown that the kinetics of TCR binding to peptide-MHC are strongly temperature-dependent (32), we analyzed the temperature dependence of the KIR/ peptide-MHC interaction (Fig. 3B). The koff of a TCR/peptideMHC interaction increased ;40-fold as the temperature was raised from 5 to 25 °C (32). In contrast, the koff of the sKIR2DL3/HLA-Cw7-DS11 interaction increased a modest ;4fold over the same temperature range (Fig. 3B). Arrhenius plots yield an activation energy of 13 kcalzmol21 for sKIR2DL3/ HLA-Cw7-DS11 dissociation (Fig. 3B, inset), far lower than the ;30 kcalzmol measured for TCR/peptide-MHC dissociation (32). Thermodynamic Analysis—The enthalpy change (DH) that accompanies KIR binding to peptide-MHC was estimated by van’t Hoff analysis, which involves measuring the dependence of affinity on temperature (Fig. 4A). Because the enthalpy and entropy vary with temperature, the non-linear form of the van’t Hoff equation was used (see “Experimental Procedures”). At 25 °C favorable enthalpic (DHvH ; 24.1 kcalzM21) and entropic (2TDS0 ; 23.1 kcalzM21) changes contribute in approximately equal measure to the binding energy (DG0 ; 27.2 kcalzM21). The heat capacity derived from this fit (DCp ; 2100 calzM21), which is a measure of the dependence of the binding enthalpy and entropy change on temperature, is well within the range determined for other protein/protein interactions (33). Similar values for DHvH and 2TDS0 were obtained when using the linear form of van’t Hoff equation (data not shown), which assumes that the enthalpy is temperature-independent. DISCUSSION

FIG. 3. Kinetic analysis of sKIR2DL3 binding to HLA-Cw7DS11. A, sKIR2DL3 was injected (solid bar) at the indicated concentrations at high flow rate (50 mlzmin21) over HLA-C7-DS11 (1500 response units (RU)). The traces shown have had their corresponding background responses (obtained with injection over the HLA-A2 surface) subtracted. Rate equations derived from the 1:1 Langmuir binding model (A 1 B 7 AB) were fitted by numerical integration simultaneously to the association and dissociation phases of all three injections (“global fitting”). Residual errors from the fits are shown in the bottom panel. B, temperature dependence of binding kinetics. sKIR2DL3 (10 mM) was injected (solid bar) at 50 mlzmin21 over HLA-Cw7-DS11 at the indicated temperatures. The responses observed with injection of the same samples over HLA-A2-peptide have been subtracted. To aid comparison the traces have been normalized so that maximum binding at each temperature equals 100%. The koff values, determined by fitting mono-exponential decay curves (solid lines), were 0.2, 0.34, and 0.8 s21 at 5, 15, and 25 °C, respectively. Inset, Arrhenius plot of koff data (KIR). Also shown is an Arrhenius plot of the koff of the JM22z TCR dissociating from HLA-A2-Flu peptide, taken from Ref. 32.

Peptide Dependence—Although several studies have demonstrated that NK cell recognition is dependent on the peptide as well as the MHC on target cells, no studies have measured directly the effect of peptide on the affinity of a KIR for a peptide-MHC class I complex. Consistent with functional (21) and binding (18) assays, we found that the affinity of sKIR2DL3 binding to HLA-Cw7-peptide was dramatically affected by the nature of the peptide (Fig. 1A and Table I). sKIR2DL3 bound to HLA-Cw7-DS11, -DS12, and -DS10 with a Kd ;7 mM, ;115 mM, and .3 mM, respectively (Table I). In agreement with this, Vale´s-Go´mez et al. (18) found that sKIR2DL3 bound to HLA-Cw7-DS11 but not to HLA-Cw7DS10. Interestingly, Mandelboim et al. (21) showed that killing by several NK clones specific for group 2 HLA-C alleles was inhibited when target cells expressed HLA-Cw7 loaded with the DS12 peptide. This suggests that affinities as low as Kd

TABLE II Summary of kinetic data Immobilized

Soluble

Immobilization level RU

HLA-Cw7-DS11

a

M

kona

koffa

21

s21

mM

21

s

Kd(calc)

sKIR2DL3

;1500

210,000 6 23,000

1.1 6 0.13

5.2

sKIR2DL3

;750

190,000 6 36,000

1.2 6 0.15

6.4

Values are means 6 S.D. of three determinations.

Kinetics and Thermodynamics of KIR/Peptide-MHC Interactions ;115 mM are sufficient to mediate inhibition. Paradoxically, however, HLA-Cw7-loaded with DS10 peptide (KYFDEHYEY), which we and others (18) show does not bind to sKIR3DL3, was a better inhibitor of most of these NK clones than HLA-Cw7DS12 (21). In only one of the clones studied (dp10.7) did the functional inhibition data (21) correlate with the direct binding data. A likely explanation for these results is that these NK clones express multiple KIRs, highlighting the importance of using purified KIRs to analyze the binding specificity. Stoichiometry—Based on the structural similarity between KIR2DL1 and hematopoietic receptors (e.g. growth hormone

FIG. 4. A, thermodynamic analysis of sKIR2DL3 binding to HLACw7-DS11. Measurement of enthalpy by van’t Hoff analysis (DHvH). Affinity constants for the sKIR2DL3/HLA-Cw7-DS11 interaction were measured at several temperatures (5–30 °C) and converted into the standard free energy of binding (DG0). Values for the enthalpic (DHvH) and standard entropic (2TDS0) changes (at 25 °C) and the specific heat capacity (DCp) were derived by fitting the non-linear form of the van’t Hoff equation to these data (see “Experimental Procedures”). B, comparison of thermodynamic properties (at 25 °C) of several macromolecular interactions. The values for protein/protein interactions (excluding antibody/protein interactions) are the mean and S.E. of 30 calorimetric determinations taken from Ref. 33. Data for TCR/peptide-MHC interactions are the mean and range of two determinations, one of which was by calorimetry (32).

28333

receptor, which binds growth hormone with a 2:1 stoichiometry), Fan et al. (34) proposed that a single peptide-MHC class I complex might bind to two KIR molecules, providing a possible mechanism for signaling. However, our results suggest that sKIR2DL3 binds to soluble HLA-Cw7-DS11 in a 1:1 complex. First, free sKIR2DL3 but no free HLA-Cw7-DS11 is detected when sKIR2DL3 and HLA-Cw7-DS11 are mixed in a ratio 1.5:1. Second, the sKIR2DL3/HLA-Cw7-DS11 complex migrated on gel filtration at the position expected for a 1:1 complex. Third, SDS-polyacrylamide gel electrophoresis of the peak complex indicated that there were equimolar amounts of sKIR2DL3 and HLA-Cw7 heavy chain (data not shown). Finally, the standard Langmuir 1:1 binding model fits very well to the equilibrium binding and kinetic data. We cannot exclude a second sKIR2DL3 site on HLA-Cw7 with a much lower affinity (e.g. Kd .200 mM). However, such a second site would need to achieve a much higher “physiological” affinity at the cell/cell interface in order to contribute to KIR2DL3 binding (35). The 1:1 binding stoichiometry suggests that, despite some structural similarities between KIRs and hematopoietic receptors, they do not bind their ligands in the equivalent manner. Indeed, comparison of the recently determined KIR2DL3 crystal structure with Ig superfamily and fibronectin type III domains indicates that KIRs bear a closer resemblance to Ig domains than to the fibronectin type III domains of hematopoietic receptors (30). Affinity, Kinetics, and Thermodynamics—The affinity measured here between soluble forms of KIR2DL3 and the HLACw7-DS11 peptide-MHC complex agrees well with the affinity measured independently by Vale´s-Go´mez et al. (24) (Kd ;9 mM) for sKIR2DL3 binding HLA-Cw7-DS11. Vale´s-Go´mez et al. (24) measured a similar affinity (Kd ;10 mM) between soluble forms of sKIR2DL1 and HLA-Cw6-peptide. This consistency between completely independent studies suggests that these affinity measurements are likely to be correct. These affinities are well within the range of affinities measured for many other cell-cell recognition molecules, including TCR/peptide-MHC interactions (Table III). However TCR/peptide-MHC interactions differ from other cell-cell molecule interactions in that low affinity of TCR binding is a consequence of a relatively slow kon rather than a fast koff (32, 36). Unlike Vale´s-Go´mez et al. (23, 24), we were able to obtain precise estimates of the binding kinetics of the sKIR2DL3/HLA-Cw7DS11 interaction. This revealed that, in contrast to TCR/peptide-MHC interactions, the low affinity is a consequence of a faster koff, whereas the kon is unremarkable, being typical of other cell surface protein/protein interactions (Table III). Furthermore, the KIR binding kinetics did not show the strong temperature dependence observed with TCR binding. Further differences between KIR/peptide-MHC and TCR/ peptide-MHC interactions were evident in thermodynamic studies. Unlike TCR binding, which is characterized by large, unfavorable entropic changes compensated for by even larger

TABLE III Summary of affinity and kinetic data for lymphocyte cell-cell recognition molecules Interaction

kon M

TCR/peptide-MHC KIR2DL3/HLA-Cw7-DS11 KIR2DL1/HLA-Cw6 CD8aa/MHC class I CD2/CD58 Mouse CD48/CD2 Mouse CD48/2B4 CD28/CD80 CTLA-4/CD80

21

s21

900–20,000 ;200,000 .200,000 .100,000 .400,000 .120,000 .200,000 .660,000 .940,000

koff

Kd

s21

mM

0.01–0.1 ;1.2 .2 .18 .4 .11 .3 .1.6 .0.4

1–90 ;7 ;10 ;200 ;10 ;90 ;15 ;4 0.46

Refs.

32, 36 This study and Ref. 24 24 28 38 39 40 41 41

28334


favorable enthalpic changes, KIR binding is driven by favorable entropic and enthalpic changes at 25 °C (Fig. 4B). The latter thermodynamic characteristics are typical of protein/protein interactions (Fig. 4B) including low affinity interactions between cell-cell recognition molecules, such as the CD2/CD48 interaction.2 A key finding in this study is that TCRs and KIRs, although recognizing overlapping portions of peptide-MHC class I molecules, bind with very different kinetic and thermodynamic properties. The KIR/peptide-MHC interaction has binding properties consistent with rigid body association, whereas TCR binding has been shown to require conformational adjustments (37) and is likely to be accompanied by a reduction in conformational flexibility (32). The difference in the binding properties of KIRs and TCRs for very similar ligands supports the suggestion the TCR and not the peptide-MHC are the primary source of conformational flexibility. We have proposed that these binding properties arise from the structure of TCR antigen-binding sites and the unique manner in which they are generated (32). First, these antigen-binding sites are formed from peptide loops exclusively, 3– 4 from each of the two V domains. Second, the loops have highly variable primary structures and are combined in a semi-random manner. Third, because TCR antigen-binding sites are not germline-encoded and cannot be inherited, they are denied the opportunity, available to other ligand-binding proteins, to acquire a more stable tertiary structure during the course of evolution. Cross-reactivity of KIR2DL1 with HLA-Cw7—Although KIR2DL molecules are grouped according to whether they bind to group 1 (KIR2DL1) or group 2 (KIR2DL2 and -L3) HLA-C alleles, recent data suggest that there is some crossreactivity. KIR2DL2 and -L3 have been shown to bind to, and mediate inhibition by, group 1 HLA-C alleles. We show here that KIR2DL1 can also cross-react with a group 2 HLA-C allele. The affinity of the latter interaction is ;10-fold lower than the affinity of sKIR2DL1 and sKIR2DL2 or 3 molecules for group 1 and group 2 HLA-C alleles, respectively. Interestingly, Vale´s-Go´mes et al. (18) also found some evidence for weak binding of sKIR2DL1 and HLA-Cw7-peptide. Whether such cross-reactivity is functionally significant remains to be demonstrated. Conclusion—We show here that a KIR binds peptide-MHC with a 1:1 stoichiometry and that the thermodynamic and kinetic features of this interaction are typical of cell/cell recognition molecules and consistent with rigid body association. This contrasts with TCR/peptide-MHC interactions, which have thermodynamic and kinetic properties more consistent with conformational adjustments at the binding interface. These differences draw attention to some unusual features of antigen recognition by TCRs and suggest that there are fundamental differences in the signal transduction mechanisms that follow KIR and TCR ligation of peptide-MHC. Acknowledgments—We thank I. Kumagai, M. Matsushima, and K. Tsumoto for their advice and helpful discussion; B. Willcox, C. O’Callaghan, B. Jakobsen, A. McMichael, and J. Bell for MHC class I molecules; and H. Wang for providing a cDNA of HLA-Cw0702. 2

J. Ladbury, P. A. van der Merwe, and S. J. Davis, unpublished data.

We are grateful to our colleagues in the Japanese Red Cross Blood Center for their encouragement. The Oxford Center for Molecular Sciences is supported by the BBSRC, MRC, and EPSRC. REFERENCES 1. Karre, K., Ljunggren, H. G., Piontek, G., and Kiessling, R. (1986) Nature 319, 675– 678 2. Ljunggren, H. G., and Karre, K. (1990) Immunol. Today 11, 237–244 3. Liao, N. S., Bix, M., Zijlstra, M., Jaenisch, R., and Raulet, D. (1991) Science 253, 199 –202 4. Lanier, L. L. (1997) Immunity 6, 371–378 5. Baker, E., D’Andrea, A., Phillips, J. H., Sutherland, G. R., and Lanier, L. L. (1995) Chromosome Res. 3, 511–512 6. Suto, Y., Maenaka, K., Yabe, T., Hirai, M., Tokunaga, K., Tadokoro, K., and Juji, T. (1996) Genomics 35, 270 –272 7. Suto, Y., Ishikawa, Y., Kasahara, M., Kasai, F., Yabe, T., Akaza, T., and Juji, T. (1998) Immunogenetics 48, 235–241 8. Gumperz, J. E., and Parham, P. (1995) Nature 378, 245–248 9. Selvakumar, A., Steffens, U., Palanisamy, N., Chaganti, R. S., and Dupont, B. (1997) Tissue Antigens 49, 564 –573 10. Lanier, L. L. (1998) Annu. Rev. Immunol. 16, 359 –393 11. Lanier, L. L., Corliss, B. C., Wu, J., Leong, C., and Phillips, J. H. (1998) Nature 391, 703–707 12. Colonna, M. (1996) Curr. Opin. Immunol. 8, 101–107 13. Lanier, L. L., and Phillips, J. H. (1996) Immunol. Today 17, 86 –91 14. Biassoni, R., Falco, M., Cambiaggi, A., Costa, P., Verdiani, S., Pende, D., Conte, R., Di Donato, C., Parham, P., and Moretta, L. (1995) J. Exp. Med. 182, 605– 609 15. Mandelboim, O., Reyburn, H. T., Sheu, E. G., Vale´s-Go´mez, M., Davis, D. M., Pazmany, L., and Strominger, J. L. (1997) Immunity 6, 341–350 16. Winter, C. C., Gumperz, J. E., Parham, P., Long, E. O., and Wagtmann, N. (1998) J. Immunol. 161, 571–577 17. Kim, J., Chwae, Y. J., Kim, M. Y., Choi, I. H., Park, J. H., and Kim, S. J. (1997) J. Immunol. 159, 3875–3882 18. Vale´s-Go´mez, M., Reyburn, H. T., Erskine, R. A., and Strominger, J. L. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 14326 –14331 19. Malnati, M. S., Peruzzi, M., Parker, K. C., Biddison, W. E., Ciccone, E., Moretta, A., and Long, E. O. (1995) Science 267, 1016 –1018 20. Rajagopalan, S., and Long, E. O. (1997) J. Exp. Med. 185, 1523–1528 21. Mandelboim, O., Wilson, S. B., Vale´s-Go´mez, M., Reyburn, H. T., and Strominger, J. L. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 4604 – 4609 22. Zappacosta, F., Borrego, F., Brooks, A. G., Parker, K. C., and Coligan, J. E. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 6313– 6318 23. Vale´s-Go´mez, M., Reyburn, H. T., Mandelboim, M., and Strominger, J. L. (1998) Immunity 9, 337–344 24. Vale´s-Go´mez, M., Reyburn, H. T., Mandelboim, M., and Strominger, J. L. (1998) Immunity 9, 892 25. Reid, S. W., Smith, K. J., Jakobsen, B. K., O’Callaghan, C. A., Reyburn, H., Harlos, K., Stuart, D. I., McMichael, A. J., Bell, J. I., and Jones, E. Y. (1996) FEBS Lett. 383, 119 –123 26. Garboczi, D. N., Hung, D. T., and Wiley, D. C. (1992) Proc. Natl. Acad. Sci. U. S. A. 89, 3429 –3433 27. Vessey, S. J., Barouch, D. H., McAdam, S. N., Tussey, L. G., Davenport, M. A., O’Callaghan, C. A., Bell, J. I., McMichael, A. J., and Jakobsen, B. K. (1997) Eur. J. Immunol. 27, 879 – 885 28. Wyer, J. R., Willcox, B. E., Gao, G. F., Gerth, U. C., Davis, S. J., Bell, J. I., van der Merwe, P. A., and Jakobsen, B. K. (1999) Immunity 10, 219 –225 29. Yoo, S. H., and Lewis, M. S. (1995) Biochemistry 34, 632– 638 30. Maenaka, K., Juji, T., Stuart, D. I., and Jones, E. Y. (1999) Structure 7, 391–398 31. Karlsson, R., and Falt, A. (1997) J. Immunol. Methods 200, 121–133 32. Willcox, B. E., Gao, G. F., Wyer, J. R., Ladbury, J. E., Bell, J. I., Jakobsen, B. K., and van der Merwe, P. A. (1999) Immunity 10, 357–365 33. Stites, W. E. (1997) Chem. Rev. 97, 1233–1250 34. Fan, Q. R., Mosyak, L., Winter, C. C., Wagtmann, N., Long, E. O., and Wiley, D. C. (1997) Nature 389, 96 –100 35. Dustin, M. L., Golan, D. E., Zhu, D. M., Miller, J. M., Meier, W., Davies, E. A., and van der Merwe, P. A. (1997) J. Biol. Chem. 272, 30889 –30898 36. Davis, M. M., Boniface, J. J., Reich, Z., Lyons, D., Hampl, J., Arden, B., and Chien, Y. (1998) Annu. Rev. Immunol. 16, 523–544 37. Garcia, K. C., Degano, M., Pease, L. R., Huang, M., Peterson, P. A., Teyton, L., and Wilson, I. A. (1998) Science 279, 1166 –1172 38. van der Merwe, P. A., Barclay, A. N., Mason, D. W., Davies, E. A., Morgan, B. P., Tone, M., Krishnam, A. K. C., Ianelli, C., and Davis, S. J. (1994) Biochemistry 33, 10149 –10160 39. Davis, S. J., Ikemizu, S., Wild, M. K., and van der Merwe, P. A. (1998) Immunol. Rev. 163, 217–236 40. Brown, M. H., Boles, K., van der Merwe, P. A., Kumar, V., Mathew, P. A., and Barclay, A. N. (1998) J. Exp. Med. 188, 2083–2090 41. van der Merwe, P. A., Bodian, D. L., Daenke, S., Linsley, P., and Davis, S. J. (1997) J. Exp. Med. 185, 393– 403