HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 275, Issue 31, 23700-23706, August 4, 2000

This Article Abstract Full Text (PDF) All Versions of this Article: 275/31/23700    most recent M003318200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Fan, Q. R. Articles by Wiley, D. C. Search for Related Content PubMed PubMed Citation Articles by Fan, Q. R. Articles by Wiley, D. C. Social Bookmarking                      What's this?

Cobalt-mediated Dimerization of the Human Natural Killer Cell Inhibitory Receptor^*

Qing R. Fan^§, Eric O. Long^¶, and Don C. Wiley

From the  Department of Molecular and Cellular Biology and Howard Hughes Medical Institute, Harvard University, Cambridge, Massachusetts 02138 and the ^¶ Laboratory of Immunogenetics, NIAID, National Institutes of Health, Rockville, Maryland 20852

Received for publication, April 17, 2000, and in revised form, May 15, 2000

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Upon engagement of specific class I major histocompatibility complex (MHC) molecules on target cells, inhibitory receptors on natural killer (NK) cells deliver a negative signal that prevents the target cell lysis by NK cells. In humans, killer cell immunoglobulin-related receptors (KIR) with two immunoglobulin-like domains (KIR2D) modulate the lysis of target cells bearing specific HLA-C alleles (Moretta, A., Vitale, M., Bottino, C., Orengo, A. M., Morelli, L., Augugliaro, R., Barbaresi, M., Ciccone, E., and Moretta, L. (1993) J. Exp. Med. 178, 597-604). The transduction of inhibitory signals by KIR2D molecules is impaired by the zinc chelator, 1,10-phenanthroline, and mutation of a putative zinc-binding site (Rajagopalan, S., and Long, E. O. (1998) J. Immunol. 161, 1299-1305), but the mechanism by which zinc may affect the function of KIR remains unknown. In this study, the inhibitory NK receptor KIR2DL1 was discovered to dimerize in the presence of Co²⁺ as observed on native gel electrophoresis and by gel filtration column chromatography. Furthermore, Co²⁺-mediated KIR2DL1 dimer binds to HLA-Cw4 with higher affinity than the wild type KIR2DL1 monomer. Replacement of the amino-terminal His residue by Ala abolishes the ability of KIR2DL1 to bind Co²⁺, indicating that Co²⁺-mediated KIR2DL1 dimerization involves pairing of the D1 domain. Although not observed on native gels, the inhibitory receptor KIR2DL1 can be chemically cross-linked into dimers in the presence of Zn²⁺ and its related divalent metal ions, suggesting that Co²⁺-mediated dimerization of KIR2DL1 may mimic a weaker interaction between KIR2DL1 and zinc in vivo.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Natural killer (NK)¹ cells express an array of inhibitory and activating receptors on the cell surface that control their cytotoxicity (reviewed in Refs. 1 and 2). In humans, the killer cell immunoglobulin-related receptors (KIR) consist of either two (KIR2D) or three (KIR3D) Ig-like domains and recognize polymorphic HLA-C and HLA-B molecules, respectively (reviewed in Refs. 1 and 2). Inhibitory KIR molecules contain immunoreceptor tyrosine-based inhibition motifs in their cytoplasmic tails, which are involved in negative signaling. The activating KIR molecules have similar extracellular domain structures as the inhibitory receptors, but lack the immunoreceptor tyrosine-based inhibition motifs in their cytoplasmic domains and contain a positively charged Lys residue in their transmembrane segments (reviewed in Refs. 1 and 2).

The inhibitory KIR2D molecules (KIR2D with a long cytoplasmic tail and referred to as KIR2DL or p58) specifically interact with HLA-C alleles (3, 4). For example, KIR2DL1 recognizes HLA-Cw4 and related HLA-C molecules that contain Asn⁷⁷ and Lys⁸⁰; KIR2DL2 recognizes HLA-Cw3 and related molecules that have Ser⁷⁷ and Asn⁸⁰ (3-5). KIR2DL molecules contain a putative zinc-binding motif (HEXXH) sequence (6), HEGVH (residues 1-5), in their extracellular domains, suggesting that zinc plays a role in the structure and function of KIR2D. Because the extracellular domain of KIR2DL1 can be refolded in the presence of EDTA and the zinc chelator, 1,10-phenanthroline, and the refolded KIR2DL1 binds to HLA-Cw4, zinc is not required for the proper folding of KIR2D or the specific interaction between KIR2D and HLA-C (7). However, functional assays have established that KIR2D-mediated inhibition of target cell lysis can be reduced by the addition of 1,10-phenanthroline and polyHis (8, 9). The clustering of HLA-C induced by inhibitory KIR molecules at the NK cell/target cell synapse is also abrogated by 1,10-phenanthroline (10). Furthermore, mutation of the His residues in the amino-terminal zinc motif (HEGVH) impairs the ability of a KIR2D molecule to deliver inhibitory signals but not its ability to bind HLA-C (8). It has been proposed that zinc mediates a protein-protein interaction that contributes to the signaling by KIR after ligand binding (8).

In this study, we demonstrate that the extracellular domain of the inhibitory NK receptor KIR2DL1 dimerizes in the presence of Co²⁺. Both soluble 2DL1H (amino acids H1-H224) and 2DL1T (amino acids H1-T200) can be dimerized by Co²⁺, indicating that dimerization of KIR2DL1 does not require the carboxyl-terminal stem region, residues 201-224. Co²⁺-mediated KIR2DL1 dimer also binds to HLA-Cw4 with higher affinity or avidity than KIR2DL1 monomer. Mutation of the amino-terminal His residue alone abolishes the ability of KIR2DL1 to form dimers through Co²⁺. The activating receptor KIR2DS4 does not have the amino-terminal His, and does not bind to Co²⁺. The formation of KIR2DL1 dimers can also be observed in the presence of Zn²⁺ and its related divalent metal ions (Co²⁺, Cu²⁺, Ni²⁺, and Cd²⁺) when the dimers are stabilized by chemical cross-linking. Other divalent metal ions (Mn²⁺, Mg²⁺, Ca²⁺, and Fe²⁺) do not cause dimerization of KIR2DL1. Zn²⁺, like Co²⁺, might have the ability to cause aggregation of KIR2D/HLA-C complexes.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Construction of Expression Plasmids-- The expression plasmids for 2DL1H (amino acids 1-224; previously named sol-cl42H), 2DL1T (amino acids 1-200; previously named sol-cl42T), HLA-Cw4 heavy chain, and ₂m were constructed as described (7). The protein coding region for the extracellular domain of KIR2DS4 (amino acids Gln¹-His²²⁴, named 2DS4H) was amplified by polymerase chain reaction from the plasmid pSPORT I p58-cl39 (11) using the 5'-primer TAGGCGAATTCTAAGGAGGATATTAAAATGCAAGAAGGAGTACACCGTAAACCAAGTTTCCTGGCCCTC and 3'-primer GTTTCAAAGCTTTAATGCAGGTGACGCGGGTTACCGGTTTT. The polymerase chain reaction fragments were digested with EcoRI and HindIII and ligated into the pLM-1 vector. The plasmids were transformed into the Escherichia coli strain BL21 (DE3) pLysS for expression.

Construction of Mutant Plasmids-- The mutant constructs H1A, H5A, H1AH5A, H13A, H182A, and E2A were created by site-directed mutagenesis, using the expression plasmid of 2DL1T as the template and the QuickChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). For the H1A mutant, the His residue at position 1 was replaced by Ala using the 5'-primer AAGGAGGATATTAAAATGGCTGAAGGAGTACACCGT and a 3'-primer that contains the complementary sequence. The H5A mutant involves changing His⁵ to Ala, and was constructed with the 5'-primer ATGCACGAAGGAGTAGCTCGTAAACCTTCTCTCCTG and its complement. The double mutant H1AH5A was engineered by mutating both His¹ and His⁵ residues into alanines with the 5'-primer AAGGAGGATATTAAAATGGCTGAAGGAGTAGCTCGTAAACCTTCTCTCCTG and its complement. The H13A mutant was generated by replacing His¹³ with Ala, using the 5'-primer CCTTCTCTCCTGGCCGCTCCAGGTCCCCTGGTG and its complement. The H182A mutant was constructed by mutating His¹⁸² to Ala with the 5'-primer TGCTTCGGCTCTTTCGCTGACTCTCCATACGAG and its complement. The E2A mutant contains the mutation from Glu residue at position 2 to Ala, and was created using the 5'-primer GATATTAAAATGCACGCTGGAGTACACCGTAAACCT and a 3'-primer that encodes the complementary sequence. The plasmids for all of the mutant constructs were transformed into BL21 (DE3) pLysS for expression.

Protein Refolding and Purification-- The inclusion bodies for the 2DS4H and all of mutant 2DL1T constructs were prepared using the same protocol as for 2DL1H and 2DL1T (7). Refolding of 2DS4H and the mutants were carried out following the same procedure as described for 2DL1H and 2DL1T, and involved injections of 2.5-3.0 mg/ml inclusion body solutions to a refolding buffer containing 100 mM Tris, pH 8.3/0.5 M L-arginine/2 mM EDTA/6.4 mM cysteamine/3.6 mM cystamine/0.1 mM phenylmethylsulfonyl fluoride (7). Similar to 2DL1H and 2DL1T, refolded 2DS4H and the 2DL1T mutants were concentrated by a weak anion-exchange (diethylaminoethyl cellulose, DE-52) column and then purified by Superdex-200 gel filtration chromatography (Amersham Pharmacia Biotech) in 20 mM HEPES, pH 8.0/150 mM NaCl. The class I major histocompatibility complex molecule HLA-Cw4 was refolded in the presence of ₂m and the peptide QYDDAVYKL (7, 12).

Divalent Metal Ion Binding Assays-- Binding assays between soluble KIR2DL1 molecules (2DL1H and 2DL1T) and divalent metal ions were carried out at 4 °C for 30 min in a reaction buffer containing 50 mM Na₂HPO₄/NaH₂PO₄, pH 7.5. The divalent metal ion solutions include Co²⁺ (CoCl₂·6H₂O), Zn²⁺ (ZnCl₂), Ni²⁺ (NiCl₂·6H₂O), Mn²⁺ (MnCl₂ ·4H₂O), Mg²⁺ (MgCl₂·6H₂O), Ca²⁺ (CaCl₂·2H₂O), Cd²⁺ (3CdSO₄·8H₂O), Cu²⁺ (CuSO₄·5H₂O), and Fe²⁺ (Fe(NH₄)₂(SO₄)₂·6H₂O).

The reaction mixtures were analyzed on 15% native acrylamide gels using either the Tris/glycine or morpholine/bicine buffer system. The Tris/glycine gel mix consisted of 15% acrylamide in 300 mM Tris, pH 8.5, and the gel running buffer contained 24.8 mM Tris/192 mM glycine. Morpholine/bicine native gels contained 15% acrylamide in 60 mM morpholine/27 mM bicine, pH 9.0. Morpholine/bicine gel electrophoresis was also carried out in 60 mM morpholine/27 mM bicine, pH 9.0.

Chemical Cross-linking-- Soluble 2DL1H or 2DL1T (0.4 mg/ml in phosphate-buffered saline, pH 7.4) was cross-linked with BS³ (bis(sulfosuccinimidyl) suberate; Pierce) at room temperature for 15 min and then quenched with 50 mM Tris-HCl, pH 6.8. The BS³ concentration used was 1.5 mM. Chemical cross-linking was carried out in the absence or presence of various divalent metal ions (Co²⁺, Cu²⁺, Zn²⁺, Ni²⁺, Mn²⁺, Mg²⁺, Ca²⁺, Cd²⁺, and Fe²⁺) at 0.2 mM concentration.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

KIR2DL1 Dimerizes in the Presence of Co²⁺-- The human inhibitory NK receptor KIR2DL1 dimerizes in the presence of Co²⁺ as demonstrated by the appearance of a KIR2DL1 dimer peak upon gel filtration chromatography for both the soluble constructs 2DL1H and 2DL1T. Soluble 2DL1H (previously named sol-cl42H) contains the entire extracellular domain of KIR2DL1, whereas 2DL1T (previously named sol-cl42T) lacks the carboxyl-terminal 24 amino acids (7). As shown in Fig. 1a, soluble 2DL1T formed a mixture of dimers (47 kDa) and monomers (21 kDa) after incubation with 0.05 mM Co²⁺ (expected molecular mass of the monomer is 22.2 kDa). Similarly, 2DL1H was eluted from a gel filtration column as a mixture of dimers (57 kDa) and monomers (26 kDa) in the presence of Co²⁺ (expected molecular mass of the monomer is 24.7 kDa) (data not shown). The fact that both 2DL1H and 2DL1T formed dimers in the presence of Co²⁺ indicates that the carboxyl-terminal stem region is not required for dimerization. Both 2DL1H and 2DL1T exist as monomers in solution in the absence of Co²⁺ (7).

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Dimerization of KIR2DL1 in the presence of Co2+. a, a gel filtration chromatography (Superdex-200, Amersham Pharmacia Biotech) profile of 2DL1T preincubated with 0.05 mM Co2+. The running buffer contained 20 mM HEPES, pH 8.0, and 150 mM NaCl. The apparent molecular masses of the peaks were determined by comparing with the profile of a mixture of standard proteins (Bio-Rad). b, Tris/glycine native gel electrophoresis (15%) of the inhibitory receptor 2DL1H in the presence of divalent metal ions. The leftmost lane, marked P220C dimer, contained a disulfide-linked 2DL1H dimer that was engineered by introducing a free cysteine at position 220 in the stem region of 2DL1H (13). The control lane contained the monomeric 2DL1H. 2DL1H was incubated with 0.1, 0.5, and 1.0 mM of Co2+ in lanes 1, 2, and 3, respectively. In the two lanes marked EDTA, 5 mM EDTA was added to the 2DL1H control as well as 2DL1H that had been incubated with 0.5 mM Co2+. The soluble 2DL1H was also incubated with 1 mM Zn2+, Ni2+, Mn2+, Mg2+, Cu2+, or Fe2+. The amount of 2DL1H protein loaded in each lane was 4 µg. c, SDS-PAGE (15%) of 2DL1H in the absence (control lane) or presence (Co2+ lane) of 0.5 mM Co2+. The amount of protein loaded in each lane was 3 µg. The SDS-PAGE was carried out under nonreducing conditions to avoid reduction of the cobalt ion by dithiothreitol.

The Co²⁺-initiated dimerization of KIR2DL1 can be detected by native gel electrophoresis. As shown in Fig. 1b, in the presence of 0.1-10 mM Co²⁺, 2DL1H was shifted into an upper band similar in position to a disulfide-linked 2DL1H dimer, which we had constructed previously (13). A dimer band with an apparent molecular mass of around 44 kDa (expected molecular mass is 49 kDa) was also observed on SDS-polyacrylamide gel electrophoresis (PAGE) under nonreducing conditions when soluble 2DL1H was incubated with Co²⁺ (Fig. 1c). Similar to 2DL1H, 2DL1T also formed dimers on native gels in the presence of Co²⁺ (shown later as a control in Fig. 5a). Neither 2DL1H nor 2DL1T can be completely shifted into the dimer band even in the presence of excess amount of Co²⁺ (shown in Fig. 1b for 2DL1H), probably because an equilibrium exists between the Co²⁺-bound dimers and Co²⁺-unbound monomers.

The Co²⁺-bound KIR2DL1 dimers are stable in the presence of EDTA. Once formed, the Co²⁺-bound 2DL1H or 2DL1T dimers could not be dissociated by the addition of 10-fold excess EDTA (shown in Fig. 1b for 2DL1H). Neither soluble 2DL1H nor 2DL1T formed stable dimers in the presence of Zn²⁺, Ni²⁺, Mn²⁺, Mg²⁺, Cu²⁺, or Fe²⁺.

Co²⁺-mediated KIR2DL1 Dimer Binds to HLA-Cw4 with Higher Affinity or Avidity Than KIR2DL1 Monomer-- Both the Co²⁺-bound 2DL1H and 2DL1T dimers bind to HLA-Cw4 with higher affinity or avidity than their corresponding monomers as demonstrated by native gel shift assays. As shown in Fig. 2, a and b (lanes marked Co²⁺ (1)), a mixture of the KIR2DL1 monomer and Co²⁺-bound dimer was incubated with HLA-Cw4 to allow complex formation, and the complex band formed by the Co²⁺-bound KIR2DL1 dimer and HLA-Cw4 (band a) differed in its mobility on native gels from the complex between the KIR2DL1 monomer and HLA-Cw4 (band c). When excess monomer and Co²⁺-bound dimer were incubated with HLA-Cw4 at the same time, only the dimer/HLA-Cw4 complex was formed while the monomer remained unshifted (Fig. 2, a and b, lanes marked Co²⁺ (2)). The fact that the formation of Co²⁺-bound KIR2DL1 dimer/HLA-Cw4 complex always dominates over the formation of KIR2DL1 monomer/HLA-Cw4 complex suggests that Co²⁺-bound KIR2DL1 dimers have higher affinity or avidity for HLA-Cw4 than their corresponding monomers.

View larger version (49K): [in this window] [in a new window]   Fig. 2.   Co2+- mediated KIR2DL1 dimer binds to HLA-Cw4 with higher affinity or avidity than KIR2DL1 monomer. a, Tris/glycine native gel electrophoresis (15%) of the binding between 2DL1H and HLA-Cw4 in the presence of Co2+. In the control binding assay, soluble 2DL1H (25 µM) was incubated with HLA-Cw4 (12.5 µM) in the absence of Co2+. In the binding assays marked Co2+ (1) and Co2+ (2), 2DL1H (25 µM overall) and HLA-Cw4 were incubated in the presence of 0.5 mM Co2+. The concentration of HLA-Cw4 was 20 µM for Co2+ (1), and 12.5 µM for Co2+ (2). For each set of binding assays, lane 1 is 2DL1H, lane 2 is HLA-Cw4, and lane 3 shows the binding between 2DL1H and HLA-Cw4. Band a corresponds to the complex formed between Co2+-bound 2DL1H dimer and HLA-Cw4, and band b is the Co2+-bound 2DL1H dimer. Band c is the complex formed between monomeric 2DL1H and HLA-Cw4 in the absence of Co2+; bands d and e correspond to free 2DL1H and HLA-Cw4, respectively. b, Tris/glycine native gel electrophoresis (15%) of the binding between 2DL1T and HLA-Cw4 in the presence of Co2+. In the control binding assay, soluble 2DL1T (25 µM) was incubated with HLA-Cw4 (12.5 µM) in the absence of Co2+. In the binding assays marked Co2+ (1) and Co2+ (2), 2DL1T (25 µM overall) and HLA-Cw4 were incubated in the presence of 0.5 mM Co2+. The concentration of HLA-Cw4 was 20 µM for Co2+ (1) and 12.5 µM for Co2+ (2). For each set of binding assays, lane 1 is 2DL1T, lane 2 is HLA-Cw4, and lane 3 shows the binding between 2DL1T and HLA-Cw4. Band a corresponds to the complex formed between Co2+-bound 2DL1T dimer and HLA-Cw4, and band b is the Co2+-bound 2DL1T dimer. Band c is the complex formed between monomeric 2DL1T and HLA-Cw4 in the absence of Co2+; bands d and e correspond to free 2DL1T and HLA-Cw4, respectively.

KIR2DL1 Binds Cu²⁺ and Remains Monomeric-- The metal binding properties of KIR2DL1 have also been assayed on morpholine/bicine native gels. The morpholine/bicine buffer system was chosen to avoid the metal-chelating activities of Tris and glycine buffers (14) that are normally used for native gels. Native gel electrophoresis using morpholine/bicine buffers has confirmed the Co²⁺-mediated dimerization of KIR2DL1 (shown in Fig. 3a for 2DL1H). Furthermore, it reveals that KIR2DL1 also binds to Cu²⁺. As shown in Fig. 3a, soluble 2DL1H migrates more slowly on native gels in the presence of Cu²⁺. This is consistent with the fact that binding of 2DL1H to a cation would reduce the overall negative charge of the protein, thereby reducing its mobility on native gels. The band corresponding to Cu²⁺-bound 2DL1H has a much faster mobility on native gels than Co²⁺-bound 2DL1H, possibly because 2DL1H does not form dimers in the presence of Cu²⁺. Similar band shifts have been observed for 2DL1T in the presence of Cu²⁺ (shown later as a control in Fig. 5b). Neither 2DL1H nor 2DL1T was gel-shifted in the presence of Zn²⁺, Ni²⁺, Mn²⁺, Mg²⁺, Ca²⁺, Cd²⁺, or Fe²⁺ (shown in Fig. 3a for 2DL1H).

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Soluble KIR2DL1 binds to Cu2+ but remains monomeric upon binding. a, morpholine/bicine native gel electrophoresis (15%) (see "Experimental Procedures") of soluble 2DL1H in the presence of divalent metal ions. The leftmost lane (control) contained the soluble monomeric 2DL1H. In the following lanes, 2DL1H was incubated with 0.5 mM Co2+, or 1 mM Zn2+, Ni2+, Mn2+, Mg2+, Ca2+, Cd2+, Cu2+, or Fe2+. The amount of 2DL1H protein loaded in each lane was 4 µg. b, a gel filtration chromatography (Superdex-200, Amersham Pharmacia Biotech) profile of 2DL1H preincubated with 0.1 mM Cu2+. The running buffer contained 20 mM HEPES, pH 8.0, and 150 mM NaCl. The apparent molecular masses of the peaks were determined by comparing with the profile of a mixture of standard proteins (Bio-Rad).

KIR2DL1 remained monomeric after Cu²⁺ binding as judged by gel filtration chromatography. Fig. 3b shows the gel filtration chromatography profile of 2DL1H that had been incubated with 0.05 mM Cu²⁺. Only a monomeric peak of 2DL1H corresponding to 24 kDa was observed. Similarly, 2DL1T incubated with Cu²⁺ exhibited a retention time corresponding to a 22-kDa protein on a gel filtration column (data not shown). No dimer peak was observed for either 2DL1H or 2DL1T that would suggest a monomer/dimer equilibrium in the presence of Cu²⁺.

Co²⁺- and Cu²⁺-binding Sites on KIR2DL1 Overlap-- The Co²⁺- and Cu²⁺-binding sites on KIR2DL1 overlap, because the binding of KIR2DL1 to one of the metal ions prevents its interaction with the other. Fig. 4 shows a morpholine/bicine native gel that compares the Co²⁺- and Cu²⁺-binding sites on 2DL1H. When 2DL1H was first incubated with Co²⁺, and then with Cu²⁺ (Co²⁺ + Cu²⁺ lane), the amount of the Co²⁺-bound 2DL1H dimer formed was the same as when 2DL1H was incubated with Co²⁺ alone (Co²⁺ lane). The faint lower band in the Co²⁺ + Cu²⁺ lane corresponds to Cu²⁺-bound 2DL1H and results from the binding of Co²⁺-unbound 2DL1H monomer to Cu²⁺. In the reverse situation, when 2DL1H was incubated with Cu²⁺, followed by Co²⁺, only a band shift corresponding to Cu²⁺ binding was observed (Cu²⁺ + Co²⁺ lane). Similarly, binding of 2DL1T to Co²⁺ inhibited its binding to Cu²⁺ and vice versa (data not shown). The Co²⁺- and Cu²⁺-binding sites on KIR2DL1 may be identical or at least share some common residues.

View larger version (86K): [in this window] [in a new window]   Fig. 4.   The Co2+- and Cu2+-binding sites on KIR2DL1 overlap. Morpholine/bicine native gel electrophoresis (15%) of 2DL1H in the presence of Co2+ and Cu2+. The control lanes contained soluble 2DL1H in the absence of any divalent metal ions. In the lane marked Co2+, 2DL1H was incubated with 0.5 mM Co2+. In the lane marked Cu2+, 2DL1H was incubated with 1 mM Cu2+. In the lane marked Co2+ + Cu2+, 2DL1H was first incubated with 0.5 mM Co2+ and then with 1 mM Cu2+. In the lane marked Cu2+ + Co2+, 2DL1H was first incubated with 1 mM Cu2+ and then with 0.5 mM Co2+. All incubation was carried out at 4 °C for 30 min. The amount of 2DL1H protein loaded in each lane was 4 µg.

Amino-terminal His of KIR2DL1 Is Involved in Co²⁺ or Cu²⁺ Binding-- To compare the metal binding properties of KIR2DL1 with the activating NK receptor KIR2DS4, we expressed the extracellular domain of KIR2DS4 (amino acids 1-224; 2DS4H) in E. coli and refolded it from inclusion bodies. The soluble activating receptor 2DS4H exists as a monomer in solution, because it was eluted as a 30-kDa protein upon gel filtration chromatography (expected molecular mass of a 2DS4H monomer is 24.7 kDa) (data not shown). 2DS4H did not form dimers in the presence of Co²⁺ or any of the divalent metal ions tested (Zn²⁺, Ni²⁺, Mn²⁺, Mg²⁺, Ca²⁺, Cd²⁺, or Fe²⁺), as demonstrated on both Tris/glycine and morpholine/bicine native gels (data not shown). Soluble 2DS4H did not bind to Cu²⁺ either (data not shown).

We exploited the fact that the inhibitory receptor KIR2DL1 and the activating receptor KIR2DS4, which have very similar sequences, differ in their ability to bind Co²⁺ and Cu²⁺ to probe the location of the Co²⁺- and Cu²⁺-binding sites on KIR2DL1 by mutagenesis. A survey of several known structures indicated that His and Cys are the most common ligands for Zn²⁺ (15). Because Co²⁺, Cu²⁺, and Zn²⁺ have similar chemical properties, we suspect that His and Cys residues may be used for the coordination of Co²⁺ and Cu²⁺ in KIR2DL1. KIR2DL1 and KIR2DS4 contain Cys residues at identical positions. A sequence comparison of the extracellular domains of KIR2DL1 and KIR2DS4 indicates that the His residues that are present in KIR2DL1 but missing in KIR2DS4 include His¹ at the amino terminus, His¹³ in the first -strand of the D1 domain, and His¹⁸² at the domain elbow (16). Another His residue that may play a role in Co²⁺ binding is His at position 5 of the amino-terminal zinc-binding motif (HEGVH, residues 1-5). Because mutation of His¹ and His⁵ in the amino-terminal zinc-binding motif reduced the ability of KIR2DL1 to deliver inhibitory signals (8), His⁵ together with His¹ in the amino-terminal region may participate in the coordination of Co²⁺ or Cu²⁺.

We have engineered five different histidine mutants of KIR2DL1, replacing one or more of the His residues (positions 1, 5, 13, and 182) with Ala for the 2DL1T construct: H1A, H5A, H1AH5A, H13A, and H182A (see "Experimental Procedures"). All of the mutants were refolded to their soluble forms using the same procedure as for wild type 2DL1T. All the mutants bind to HLA-Cw4 as demonstrated by native gel shift assays (data not shown). A mutated KIR2DL1-Ig fusion protein having six His residues (positions 1, 5, 36, 40, 55, and 56) replaced by Ala was shown previously to bind to HLA-Cw4 (8). The Co²⁺ and Cu²⁺ binding properties of the mutants were assayed on Tris/glycine and morpholine/bicine native gels, respectively. As shown in Fig. 5a, mutation of the amino-terminal His¹ alone (H1A) abolished the Co²⁺-binding activity of 2DL1T. Mutating His⁵ (H5A), His¹³ (H13A), or His¹⁸² (H182A) to Ala did not affect the dimerization of 2DL1T by Co²⁺. The double mutant H1AH5A no longer bound to Co²⁺, because it contained the mutation from His¹ to Ala. As expected from the fact that the Co²⁺- and Cu²⁺-binding sites on KIR2DL1 overlap, the H1A and H1AH5A mutants did not bind to Cu²⁺ either (Fig. 5b). We have therefore identified the amino-terminal His residue as part of the Co²⁺- and Cu²⁺-binding sites on KIR2DL1.

View larger version (18K): [in this window] [in a new window]   Fig. 5.   The amino-terminal His of KIR2DL1 is involved in Co2+ and Cu2+ binding. a, Tris/glycine native gel electrophoresis (15%) of wild type 2DL1T (control) and its histidine mutants (H1A, H5A, H1AH5A, H13A, and H182A) in the presence (+ lanes) or absence ( lanes) of Co2+. The concentration of Co2+ used was 1 mM. The amount of wild type or mutant 2DL1T protein loaded in each lane was 2 µg. b, morpholine/bicine native gel electrophoresis (15%) of wild type 2DL1T (control) and its histidine mutants (H1A, H5A, H1AH5A, H13A, and H182A) in the presence (+ lanes) or absence ( lanes) of Cu2+. The concentration of Cu2+ used was 1 mM. The amount of wild type or mutant 2DL1T protein loaded in each lane was 2 µg.

To test whether the Glu residue at position 2 in the putative zinc-binding motif (HEGVH, residues 1-5) may play a role in Co²⁺ binding, we also made a mutant changing Glu² to Ala (E2A). Like the wild type 2DL1T, the E2A mutant dimerizes in the presence of Co²⁺ (data not shown), indicating that Glu² is not involved in the binding of 2DL1T to Co²⁺.

A sequence alignment of the extracellular domains of different KIRs indicates that the amino-terminal His residue is conserved in all inhibitory KIR molecules containing two Ig-like domains (KIR2DL). All the inhibitory KIR2DL receptors may therefore have the potential to form dimers in the presence of Co²⁺. Although the activating receptor KIR2DS4 does not have an amino-terminal His, and does not bind Co²⁺, several other activating KIR2D molecules (KIR2DS, also known as p50) (17), including KIR2DS2, KIR2DS3, and KIR2DS5 (11, 17-19), all contain the initial His. This suggests that some activating receptors may also be dimerized by Co²⁺. An amino-terminal His is also found in some inhibitory and activating KIR molecules with three Ig-like domains, D0, D1, and D2, (KIR3D) (4, 18, 20-22), such as KIR3DL1 and KIR3DS1. However, KIR3DL2 does not have the amino-terminal His. Whether the initial His residue can cause dimerization of KIR3D molecules and pairing of the D0 domain still remains to be explored. There is no His in the KIR3D molecules at the amino terminus of the D1 domain, the homologous position of the KIR2D amino-terminal His.

KIR2DL1 Can Be Cross-linked into Dimers in the Presence of Zn²⁺ and Related Divalent Metal Ions-- Although we have not observed any binding between KIR2DL1 and Zn²⁺ by native gel shift assays, we cannot rule out the possibility that KIR2DL1 binds Zn²⁺ with a lower affinity. Except for the coenzyme vitamin B₁₂ (23), there are very few biological systems that utilize cobalt. Co²⁺ can often substitute for Fe²⁺, Cu²⁺, and Zn²⁺ in vitro; examples include the replacement of natural iron heme in hemoglobin and myoglobin with a synthetic Co²⁺ heme (24), Co²⁺ derivatives of blue copper proteins (25), and Co²⁺-substituted zinc-binding carboxypeptidase (26). Although the amount of cobalt is hardly detectable in serum, zinc is present at 5-20 µM in serum in adults (27). It is possible that the Co²⁺-mediated dimerization of KIR2DL1 mimics the interactions between KIR2DL1 and Zn²⁺ in vivo.

To detect potential dimer formation of KIR2DL1 in the presence of Zn²⁺ and its related metal ions, soluble KIR2DL1 was incubated with various divalent metal ions and then chemically cross-linked with the reagent BS³ (bis(sulfosuccinimidyl) suberate) to stabilize any dimer formation, if it exists. The physiological concentration of free Zn²⁺ ion has been estimated to be approximately 0.25-1 µM (28), based on the fact that 95% of zinc in serum is complexed with proteins (29, 30). A much higher concentration of Zn²⁺ (0.2 mM) was needed for measurements here to allow visualization of the effect of Zn²⁺ on the oligomeric state of KIR2DL1, although some cross-linking was evident at 10 µM Zn²⁺. In the absence of any metal ions, a very faint dimer band of soluble 2DL1H was observed when it was incubated with the cross-linker (Fig. 6). In contrast, a significant amount of 2DL1H formed dimers when cross-linked in the presence of Co²⁺, Cu²⁺, Zn²⁺, Ni²⁺, and Cd²⁺ (Fig. 6). In addition to the dimer band, soluble 2DL1H also formed higher molecular weight aggregates in the presence of these metal ions (Fig. 6). The divalent metal ions Co²⁺, Cu²⁺, Ni²⁺, and Cd²⁺ have similar chemical properties as Zn²⁺ and can all substitute for Zn²⁺. In the presence of Mn²⁺, Mg²⁺, Ca²⁺, and Fe²⁺, only the background dimer band was observed (Fig. 6), suggesting that dimerization of KIR2DL1 is specifically mediated by Zn²⁺ and its related metal ions Co²⁺, Cu²⁺, Ni²⁺, and Cd²⁺. Like 2DL1H, soluble 2DL1T, which lacks the carboxyl-terminal 24 amino acids, can also be specifically cross-linked into dimers by Co²⁺, Cu²⁺, Zn²⁺, Ni²⁺, and Cd²⁺ (data not shown), confirming our earlier observation that the carboxyl-terminal stem region is not required for Co²⁺-mediated KIR2DL1 dimerization.

View larger version (25K): [in this window] [in a new window]   Fig. 6.   Chemical cross-linking of KIR2DL1 in the presence of divalent metal ions. SDS-PAGE (15%) of 2DL1H cross-linked with 1.5 mM BS3 (bis(sulfosuccinimidyl) suberate) and in the presence of various divalent metal ions, including Co2+, Cu2+, Zn2+, Ni2+, Mn2+, Mg2+, Ca2+, Cd2+, and Fe2+. The metal ions were present at 0.2 mM concentration. Soluble 2DL1H was also cross-linked with BS3 at 0.0 mM ( lane) and 1.5 mM (+ lane) in the absence of any divalent metal ions (control). The amount of 2DL1H protein loaded in each lane was 4 µg. The SDS-PAGE was carried out under reducing conditions.

Aggregation of a Disulfide-linked KIR2DL1 Dimer by Co²⁺-- In addition to the formation of a significant amount of dimers, higher molecular weight aggregates of KIR2DL1 have also been observed in the presence of Zn²⁺, Co²⁺, Cu²⁺, Ni²⁺, and Cd²⁺ (Fig. 6). To explore the mode of divalent metal ion-induced aggregation, we have examined the effect of Co²⁺ on a KIR2DL1 dimer that is covalently linked at the carboxyl-terminal stem region by a disulfide bond (P220C) (13). The disulfide-linked dimer formed a dimer of dimers and even higher order dimer associates in the presence of Co²⁺ (Fig. 7a). Because the P220C dimer is linked at the carboxyl-terminal stem region after the D2 domain, the amino termini in the D1 domains of two neighboring P220C dimer molecules are free to dimerize through Co²⁺, thus resulting in the formation of a dimer of dimers and even higher order aggregates (Fig. 7b). This hypothetical model for the aggregation of P220C dimer by Co²⁺ may represent the mechanism by which Zn²⁺ aggregates KIR·HLA-C complexes for signaling.

View larger version (10K): [in this window] [in a new window]   Fig. 7.   Aggregation of a disulfide-linked KIR2DL1 dimer by Co2+. a, Tris/glycine native gel electrophoresis (15%) of a KIR2DL1 dimer in the absence (control lane) or presence of 0.05 mM Co2+ (Co2+ lane). As in Fig. 1b, the P220C dimer of KIR2DL1 is covalently linked at the carboxyl-terminal stem region through a disulfide bond. b, hypothetical model for the aggregation of KIR2D by divalent metal ions. M2+ represents Zn2+ and related divalent metal ions.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The extracellular domain of the inhibitory NK receptor KIR2DL1 dimerizes in the presence of Co²⁺ as demonstrated by native and SDS-gel electrophoresis as well as gel filtration chromatography. Furthermore, the Co²⁺-bound dimer binds to HLA-Cw4 with higher affinity or avidity than does the KIR2DL1 monomer. Both the soluble 2DL1H (amino acids 1-224) and 2DL1T (amino acids 1-200) constructs form dimers through Co²⁺, indicating that the carboxyl-terminal stem region is not required for dimerization. Further mutational studies have revealed that the Co²⁺-binding site is located at the amino terminus of KIR2DL1; replacement of the initial His¹ residue with Ala prevents the Co²⁺-induced dimerization of KIR2DL1. Because the amino-terminal His¹ residue is conserved in all inhibitory KIR2D molecules, Co²⁺-mediated dimerization may therefore be a common property of inhibitory NK receptors that specifically recognize HLA-C.

The inhibitory receptor KIR2DL1 also binds to Cu²⁺, as shown by morpholine/bicine native gel electrophoresis; it remains monomeric upon Cu²⁺ binding as judged by gel filtration chromatography. However, the fact that Cu²⁺ can enhance chemical cross-linking of KIR2DL1 into dimers suggests that Cu²⁺ may dimerize KIR2DL1 in the same way as Co²⁺ but with lower affinity. Indeed, we have found that the Co²⁺- and Cu²⁺-binding sites on KIR2DL1 overlap, and the same mutation of His¹ with Ala knocked out both the Co²⁺- and Cu²⁺-binding activities of KIR2DL1.

Chemical cross-linking has also allowed us to detect dimerization of KIR2DL1 in the presence of Zn²⁺ and several of its related divalent metal ions (Ni²⁺ and Cd²⁺, in addition to Co²⁺ and Cu²⁺). Increased dimerization was not observed in the presence of other divalent metal ions (Mn²⁺, Mg²⁺, Ca²⁺, and Fe²⁺), suggesting that dimerization of KIR2DL1 is specific for Zn²⁺ and its related metal ions. Binding of KIR2DL1 to Zn²⁺, Ni²⁺, and Cd²⁺ have not been observed on native gels or by gel filtration chromatography, indicating that KIR2DL1 has a much lower affinity for Zn²⁺ than for Co²⁺ and that Co²⁺-mediated dimerization may mirror the interaction between KIR2DL1 and Zn²⁺ in vivo. The weak interaction of Zn²⁺, relative to Co²⁺, with inhibitory NK receptors may be necessary to modulate the negative signaling in NK cells, because a strong interaction may result in the constitutive formation of dimers that are active in signaling.

In summary, we have discovered that the inhibitory NK receptor KIR2DL1 dimerizes in the presence of Co²⁺ and found evidence that Co²⁺-mediated dimerization of KIR2DL1 may mimic the interaction between KIR2DL1 and Zn²⁺ in vivo. Because mutation of the amino-terminal His residues impairs the inhibitory function of KIR2D molecules (8) and disrupts the Co²⁺-mediated dimerization of KIR2DL1 (this study), divalent metal ion (possibly Zn²⁺)-induced dimerization of KIR2D may be necessary for the delivery of an inhibitory signal. In the Co²⁺-bound KIR2DL1 dimer, the cytoplasmic tails may be far apart, because Co²⁺ binds to the amino terminus. Dimerization of the cytoplasmic domains could probably be achieved through a conformation resembling the disulfide-linked P220C dimer. Zn²⁺ may be required for the aggregation of KIR2D·HLA-C complexes in the same way that Co²⁺ causes aggregation of P220C dimers.

	ACKNOWLEDGEMENTS

We thank Dr. J. L. Strominger, Dr. H. Ploegh, and Dr. M. Eck for critical reading of the manuscript as part of Q. R. F.'s thesis; Dr. A. Chrombach for help with designing the morpholine/bicine buffer system for native gel electrophoresis; Dr. D. N. Garboczi for helpful advice; A. Haykov for inclusion body preparation; N. Sinitskaya for oligonucleotide and peptide synthesis; and members of the Wiley lab for discussion.

	FOOTNOTES

^* This work was supported by the National Institutes of Health and the Howard Hughes Medical Institute.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ Supported by a National Science Foundation Predoctoral Fellowship.

An investigator of Howard Hughes Medical Institute. To whom correspondence should be addressed: Dept. of Molecular and Cellular Biology and Howard Hughes Medical Institute, Harvard University, 7 Divinity Ave., Cambridge, MA 02138. Tel.: 617-495-1808; Fax: 617-495-9613; E-mail: dcwadmin@crystal.harvard.edu.

Published, JBC Papers in Press, May 17, 2000, DOI 10.1074/jbc.M003318200

	ABBREVIATIONS

The abbreviations used are: NK, natural killer cells; KIR, killer cell immunoglobulin-related receptor; BS³, bis(sulfosuccinimidyl) suberate.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				H. Li, Y. Zhao, Y. Guo, Z. Li, L. Eisele, and W. Mourad Zinc Induces Dimerization of the Class II Major Histocompatibility Complex Molecule That Leads to Cooperative Binding to a Superantigen J. Biol. Chem., March 2, 2007; 282(9): 5991 - 6000. [Abstract] [Full Text] [PDF]

				B. Treanor, P. M.P. Lanigan, S. Kumar, C. Dunsby, I. Munro, E. Auksorius, F. J. Culley, M. A. Purbhoo, D. Phillips, M. A.A. Neil, et al. Microclusters of inhibitory killer immunoglobulin-like receptor signaling at natural killer cell immunological synapses J. Cell Biol., July 3, 2006; 174(1): 153 - 161. [Abstract] [Full Text] [PDF]

				M. Faure, D. F. Barber, S. M. Takahashi, T. Jin, and E. O. Long Spontaneous Clustering and Tyrosine Phosphorylation of NK Cell Inhibitory Receptor Induced by Ligand Binding J. Immunol., June 15, 2003; 170(12): 6107 - 6114. [Abstract] [Full Text] [PDF]

				J. E. Boyson, R. Erskine, M. C. Whitman, M. Chiu, J. M. Lau, L. A. Koopman, M. M. Valter, P. Angelisova, V. Horejsi, and J. L. Strominger Disulfide bond-mediated dimerization of HLA-G on the cell surface PNAS, December 10, 2002; 99(25): 16180 - 16185. [Abstract] [Full Text] [PDF]

				L. M. Carlin, K. Eleme, F. E. McCann, and D. M. Davis Intercellular Transfer and Supramolecular Organization of Human Leukocyte Antigen C at Inhibitory Natural Killer Cell Immune Synapses J. Exp. Med., November 19, 2001; 194(10): 1507 - 1517. [Abstract] [Full Text] [PDF]

				M. Valés-Gómez, R. A. Erskine, M. P. Deacon, J. L. Strominger, and H. T. Reyburn The role of zinc in the binding of killer cell Ig-like receptors to class I MHC proteins PNAS, February 1, 2001; (2001) 41618298. [Abstract] [Full Text]

				M. Vales-Gomez, R. A. Erskine, M. P. Deacon, J. L. Strominger, and H. T. Reyburn The role of zinc in the binding of killer cell Ig-like receptors to class I MHC proteins PNAS, February 13, 2001; 98(4): 1734 - 1739. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) All Versions of this Article: 275/31/23700    most recent M003318200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Fan, Q. R. Articles by Wiley, D. C. Search for Related Content PubMed PubMed Citation Articles by Fan, Q. R. Articles by Wiley, D. C. Social Bookmarking                      What's this?