HLA-F and MHC-I Open Conformers Bind Natural Killer Cell Ig-Like Receptor KIR3DS1

Abstract

Based on previous findings supporting HLA-F as a ligand for KIR3DL2 and KIR2DS4, we investigated the potential for MHC-I open conformers (OCs) as ligands for KIR3DS1 and KIR3DL1 through interactions measured by surface plasmon resonance. These measurements showed physical binding of KIR3DS1 but not KIR3DL1 with HLA-F and other MHC-I OC while also confirming the allotype specific binding of KIR3DL1 with MHC-I peptide complex. Concordant results were obtained with biochemical pull-down from cell lines and biochemical heterodimerization experiments with recombinant proteins. In addition, surface binding of HLA-F and KIR3DS1 to native and activated NK and T cells was coincident with specific expression of the putative ligand or receptor. A functional response of KIR3DS1 was indicated by increased granule exocytosis in activated cells incubated with HLA-F bound to surfaces. The data extend a model for interaction between MHC-I open conformers and activating KIR receptors expressed during an inflammatory response, potentially contributing to communication between the innate and adaptive immune response.

Fig 1. KIR3DL1 and KIR3DS1 binding to MHC-I Bw4, Bw6, and HLA-F epitopes as measured by surface plasmon resonance (SPR).

Sensorgrams of KIR3DL1 and KIR3DS1 binding–indicated above each column of graphs–to immobilized refolded Bw4 epitopes of HLA heavy chain, β₂M and respective peptide for each HLA complex as indicated. Experiments were performed in triplicate on the same chip before (blue) and after (red) acid treatment. KIR3DL1 and KIR3DS1 concentrations of 1 μM and 2 μM are indicated by light and dark colors respectively. SPR profiles for additional Bw4 and non-Bw4 allotypes, plus controls for the integrity of each complex and free heavy chain/open conformer generated after acid treatment are presented in supplementary S1 Fig and S2 Fig.

Fig 2. Western analysis using the indicated mAbs for detection of gel fractionated protein after pull-down with KIR3DS1-D0-D2stem-His (above).

Four cell lines were incubated with KIR3DS1-His and pull-downs performed followed by Western blot analysis with HCA2 and 3D11. Heterodimerization of KIR-His and HLA-F without tag followed by Ni column purification (below). After gel fractionation, gels were stained with Coomassie blue and proteins visualized. Heterodimerization assays were carried out with KIR alone or with KIR + HLA-F as indicated above each lane. Recombinant HLA-F is included alone for comparison. MW markers are indicated for both gels.

Fig 3

Surface staining with KIR3DS1-His and HLA-F (A) PBMC from three individuals were stained with NK markers (CD3-, CD56+) and with mAb Z27 and DX9 to detect KIR3DL1 (DX9+, Z27+) and KIR3DS1 (DX9-, Z27+), and stained with HLA-F tetramer as indicated. Three individuals were chosen based on their KIR haplotypes, containing homozygous KIR3DL1, heterozygous KIR3DL1, KIR3DS1, and homozygous KIR3DS1 as indicated. (B) Cells described in A were activated and subjected to the same staining protocols 11 days post activation. (C) Isolated T cells from two individuals with the indicated KIR3DS1L1 genotypes were stained with mAb 3D11 and with KIR3DS1-His as marked beneath each pair of before (left) and after (right) activation profiles.

Fig 4. Functional measurement of HLA-F and KIR3DS1 interaction.

PBMC from three donors were activated and stimulated with refolded HLA-F, HLA-control tetramers, or no protein as described in Materials and Methods. (A) A representative FACS staining for the analysis of CD107a expression on cell populations gated in the top panel indicated as KIR3DS1+ and KIR3DS1-. Expression of CD107a was detected with percentages of the total cell population analyzed in bold in the lower right quadrant of each profile. Experiments were performed in the presence of HLA-F and D^b-TRP bound to streptavidin beads and no-protein as indicated above the panels. The percentages CD107a positive cells within the KIR3DS1+ populations (B) and in the KIR3DS1- population (C) for each donor are graphed for comparison. (B and C) Five experiments were performed for each of four KIR3DS1+ donors (Donors CW and EB: KIR3DL1/KIR3DS1, Donors TG and DG KIR3DS1/KIR3DS1). The graphs show the percentage of total cells using the formula $\frac{\%\left(For{D}^b\right)-\%control}{\%control}$ showing the mean of 5 replicates for each individual, the error bars represent the standard error of the mean (SEM) of the replicates, and the p-values calculated by performing a paired t-test comparing all values F and D^b across replicates and individuals.

Fig 5. Steric clashes with presented peptides introduced by KIR substitutions.

(Left) View of the KIR3DL1/HLA-B*5701 interface (PDB accession code 3VH8; www.rcsb.org). (Right) Model of KIR3DS1 in complex with HLA-B*5701, shown as at left. The steric clash between Arg166 with p8 is highlighted in red. One rotomer of Arg166 is depicted; however, all accessible Arg166 rotomers sterically clash with residue p8 of the bound peptide. HLA-B*5701 is shown in cartoon representation, colored teal; the bound peptide (LSSPVTKSF) is shown as a ribbon with side-chains in licorice stick representation, colored yellow. KIR3D molecules are shown in semi-transparent surface representation, colored grey. Residue 166 from the KIR3D molecules is shown in licorice stick representation, colored grey (KIR3DL1) or grey and red (KIR3DS1).