We stained several cell lines with a soluble fusion chimera consisting of the extracellular domain of KIR3DS1 attached to the Fc region of human IgG1 (KIR3DS1-Fc). We chose KIR3DS1 because it does not bind to HLA class I or class II genes [except HLA-F, which has restricted tissue expression (20, 21)]. Surprisingly, KIR3DS1-Fc bound to all human cell lines arising from many cell lineages including those of T-cell (Jurkat), B-cell (721.221, RAJI, and EBV-transformed primary B-cell line), monocytic (THP-1), and erythro-myeloid (K562) origin ( Figure 1A ). Of interest, K562 cells are HLA class I deficient (including HLA-F deficient), which cued us to the existence of non-HLA ligands for KIR3DS1.

Given the widespread expression of KIR3DS1 ligands across various human cell lines, we decided to perform a pooled, genome-wide CRISPR/Cas9-based screen [previously described in (22)] in K562 cells, which are HLA-deficient and exhibited the highest levels of KIR3DS1-Fc binding ( Figure 1B ). Briefly, K562 cells stably expressing Cas9 were transduced with a lentiviral single-guide RNA (sgRNA) library that targeted 18,166 human protein-coding genes, and were then cultured under selection to allow for gene inactivation and turnover of gene products to occur. K562 library cells were then stained with KIR3DS1-Fc, followed by a secondary stain with a PE-conjugated anti-human IgG (Fc-specific) antibody, and a tertiary stain with a magnetic-bead–conjugated anti-PE antibody to perform magnetic enrichment. The magnetically enriched cells were cultured, expanded, and then analyzed by flow cytometry, which showed that whereas almost all cells in the initial population (i.e. pre-enrichment) bound to KIR3DS1-Fc, cells cultured post-enrichment had ~50% of cells staining very dimly for KIR3DS1-Fc binding ( Figure 1C ).

Using high-throughput sequencing, sgRNA barcodes in the cultured post-enrichment cell population were quantified to assess enrichment of specific sgRNA sequences. Remarkably, sequencing showed that the top nine enriched target genes encoded enzymes involved in heparan sulfate biosynthesis ( Figure 1D and Table 1 ). These enzymes were directly involved in either heparan sulfate polysaccharide backbone polymerization (i.e. XYLT2, B4GALT7, B3GAT3, EXTL3, EXT1, EXT2), or in the synthesis of precursor molecules needed for heparan sulfate biosynthesis (i.e. UXS1, UGDH, and PAPSS1) ( Figure 1E ). These results indicated that heparan sulfate proteoglycans were a target ligand of KIR3DS1-Fc on K562 cells.

Given that protein interactions with heparan sulfate are dominated by electrostatic forces between positively-charged basic residues on proteins and negatively-charged sulfate moieties on heparan sulfate (6, 23), we computationally determined the isoelectric point (pI) of various domains and regions of KIR3DS1 and other KIRs to roughly assess their potential for interacting with heparan sulfate. Our analyses revealed that D0 domains, which only some KIRs (including KIR3DS1) contain, have a pI above physiological pH ( Figure 2A ), indicating that the D0 domain of these KIRs has a positive net charge and likely mediates binding to heparan sulfate.

To confirm this, we performed surface plasmon resonance on several KIRs to comparatively assess their affinities to heparan sulfate. We included two additional GAGs, hyaluronic acid and chondroitin sulfate, as controls. Remarkably, we found that all Fc constructs of KIRs containing a D0 domain (KIR3DS1, KIR3DL1, KIR3DL2, KIR2DL4, and KIR2DL5) bound to heparan sulfate but not to chondroitin sulfate or hyaluronic acid, whereas there was no GAG binding observed for KIRs not containing a D0 domain (KIR2DL1, KIR2DL3, and KIR2DS4) ( Figure 2B ). Interestingly, kinetic analyses revealed that these interactions were somewhat higher affinity than previously published affinities of KIR3DS1, KIR3DL1, and KIR3DL2 towards HLA-F open conformers (20) ( Table 2 ). This was in concordance with the results from the CRISPR/Cas9 screen and our prediction of KIRs containing a D0 domain having a unique ability to bind to heparan sulfate via electrostatic interactions.

To further validate our findings, we investigated whether KIR3DS1 binding to cell lines can be abolished by eliminating heparan sulfate proteoglycans on the cell surface. To accomplish this, we treated 721.221 cells, another HLA-deficient cell line that showed KIR3DS1-Fc binding ( Figure 1A ), enzymatically with heparinase II or proteinase K. Heparinase II degrades the heparan sulfate sugar backbone while proteinase K cleaves cell-surface proteins (including the proteoglycans to which heparan sulfate is attached) without significantly affecting cell integrity. Indeed, enzymatic treatment with both heparinase II and proteinase K resulted in a dramatic decrease in KIR3DS1-Fc binding to 721.221 cells compared to untreated controls, consistent with heparan sulfate proteoglycans being a cellular ligand for KIR3DS1-Fc ( Figure 3A ).

Given that PAPSS1 scored highly in the CRISPR/Cas9-based screen, we were also interested in experimentally determining whether binding of KIR3DS1 to surface-expressed heparan sulfate proteoglycans was dependent on the presence of sulfate moieties on heparan sulfate. PAPSS1 encodes for 3’-phosphoadenosine-5’-phosphosulfate (PAPS) synthase, an enzyme required to generate PAPS, which is the universal sulfate donor for all cellular sulfotransferase reactions. PAPS synthase can be chemically inhibited by chlorate ( Cl O 3 − ) , an inhibitor that competitively blocks binding of sulfate ions. To eliminate sulfate moieties from heparan sulfate, we cultured EBV-BCLs, which also exhibited high levels of KIR3DS1-Fc binding, in sulfate-free media containing different concentrations of Cl O 3 − (ranging from 10 to 50 mM) and found that KIR3DS1-Fc binding was lost even at low concentrations of Cl O 3 − ( Figure 3B ). KIR3DS1 binding and heparan sulfation content were then assessed at various time points after culturing cells in sulfate-free media containing Cl O 3 − . Heparan sulfation content was measured using an anti-heparan sulfate antibody (clone: 10E4) that recognizes sulfated GlcN residues (25). Continuous culture in sulfate-free media containing Cl O 3 − resulted in loss of cell-surface heparan sulfation and a concomitant decrease in KIR3DS1-Fc binding ( Figure 3C ). This demonstrated that binding of KIR3DS1 to heparan sulfate is dependent on the presence of sulfate moieties, which has been shown to be the case for other heparan sulfate-binding proteins (16). One previous report demonstrated that Cl O 3 − concentrations of 5–20 mM resulted in selective loss of heparan sulfate 6-O-sulfate moieties (26), which may implicate these moieties specifically in binding of KIR3DS1 to heparan sulfate.

Finally, we wanted to test whether KIR binding to heparan sulfate proteoglycans has an impact on binding to known canonical ligands (e.g. HLA). To do so, we explored the effects of heparan sulfate elimination on the interaction of KIR3DL1 with its ligand HLA-B*57:01. We performed KIR3DL1-Fc staining of 721.221 cells that were untransduced (null) or stably transduced with HLA-B*08:01 (a non-ligand) or HLA-B*57:01 (a known HLA ligand) and cultured in either normal media or sulfate-free medium containing Cl O 3 − . Remarkably, we found that while elimination of heparan sulfate proteoglycans abolished binding of KIR3DL1-Fc to untransduced and HLA-B*08:01-expressing 721.221 cells, binding to 721.221 cells expressing HLA-B*57:01 was maintained at a high level ( Figure 3D ). This supports the notion that KIR binding to HLA is not significantly affected by the absence of heparan sulfate proteoglycans.

K562, Jurkat, THP-1, RAJI, and 721.221 cell lines (including HLA transductants) were grown in RPMI-1640 supplemented with 10% fetal bovine serum (Sigma-Aldrich), 2 mM L-glutamine (Gibco), 100 U/mL penicillin (Gibco), and 100 U/mL streptomycin (Gibco) at 37°C/5% CO₂. EBV-transformed B-cell lines (EBV-BCLs) were generated from peripheral blood mononuclear cells from donors bearing specific HLA genotypes; EBV-BCLs were also grown in the same media and conditions. Human samples were used in this study in accordance to protocols approved by Partners Human Research Committee and Institutional Review Board of Massachusetts General Hospital.

All KIR-Fc constructs used in this study were produced in mammalian cell lines and purchased from R&D Systems; they included the following: KIR3DS1-Fc, KIR3DL1-Fc, KIR3DL2-Fc, KIR2DL1-Fc, KIR2DL3-Fc, KIR2DL4-Fc, KIR2DS4-Fc, and KIR2DL5-Fc. Cells were stained with KIR-Fc constructs by washing extensively with PBS, incubating for 30–45 min with 25 μg/mL KIR-Fc diluted in PBS at 4°C while shaking, followed by a secondary staining with goat anti-human lgG Fc-PE (Life Technologies) for 30 min at 4°C while shaking. After staining, cells were washed and fixed using 4% paraformaldehyde/PBS (Affymetrix) and flow cytometric analysis was performed on an BD LSR II or LSR Fortessa. Measurement of heparan sulfate expression by flow cytometry was done using anti-HS-FITC (clone: 10E4, US Biological; used at 1:10 dilution).

CRISPR/Cas9-based library of pooled mutant cells was generated similarly to previously described work published in (22). sgRNA Library design. The optimized sgRNA library was designed by first computationally filtering out potential off-target matches (except for sgRNAs that targeted multiple homologs), and also to have high cleavage activity (based on rules experimentally determined in another set of experiments). In total, a library containing 178,896 sgRNAs targeting 18,166 protein-coding genes in the human consensus CDS (CCDS) and 1,004 non-targeting control sgRNAs was constructed. Cell library generation. K562 cells were lentivirally transduced with Cas9/eGFP vector. 2.40 × 10⁸ Cas9-transduced cells were transduced with the viral pool containing the sgRNA library to achieve an average 1000-fold coverage of the library after selection with puromycin for 7 d. Screening procedure. 5 × 10⁷ pooled mutant K562 cells were washed in PBS and stained in 1 mL of 25 μg/mL KIR3DS1-Fc (R&D) for 15 min at 4°C. Cells were then washed with PBS and stained with 1 mL of anti-hIgG(Fc)-PE (Life Technologies) diluted 1:50 in PBS. Fluorescently labeled cells were washed with PBS and stained with anti-PE Microbeads (Miltenyi) following manufacturer’s instructions but using twice the recommended volumes. Magnetically labelled cells were then depleted in an LD column (Miltenyi) following manufacturer’s instructions, and flow-through cells were collected and cultured for 7 d. Screen deconvolution. sgRNA inserts were PCR amplified from 5 million genome equivalents of DNA from flow-through cells. The resultant PCR products were purified and sequenced on a HiSeq 2500 (Illumina); primer sequences are as follows:

sgRNA quantification primers:

Illumina sequencing primer:

Illumina indexing primer:

Data analysis. Sequencing reads were aligned to the sgRNA library and the abundance of each sgRNA was calculated. Gene-based CRISPR scores (CS) were defined as the average log2 fold change of all sgRNAs targeting a given gene and calculated for the entire screen. To identify enriched genes, the CS distribution was mean-normalized to zero.

For sulfation inhibition, cells were grown for 48 h (or other time period if specified) in sulfate-free/NaClO₃-containing media, which consisted of custom-made Advanced RPMI-1640 (Life Technologies) which had 407 μM magnesium sulfate replaced with 407 μM magnesium chloride, 3.03 μM zinc sulfate replaced with 3.03 μM zinc chloride, 5 nM copper (II) sulfate replaced with 5 nM copper (II) chloride, and sodium chloride reduced from 103 mM to 53 mM, and was supplemented with 10% dialyzed FBS (Gibco), 2 mM L-glutamine (Gibco), 100 μg/mL primocin (In vivogen), and 50 mM sodium chlorate (NaClO₃; Sigma-Aldrich) or other NaClO₃ concentration if specified, which was balanced with sodium chloride (Sigma-Aldrich).

For surface enzymatic treatment of cells, 2.5 × 10⁵ cells were washed three times with PBS, and then incubated in PBS containing 2 U/mL heparinase II (New England BioLabs), 1 U/mL proteinase K (New England BioLabs), or no enzyme for 1 h at 37°C/5% CO₂. Cells were then washed in ice-cold PBS and stained as indicated.

Isoelectric point analysis was determined by inputing Ig-domain sequences (from Cys to Cys) into ExPASy Compute pI/Mw tool, which calculates pI using pK_a values of amino acids (24). The KIR sequences analyzed were obtained from Immuno Polymorphism Database (IPD) (38, 39) and were the following: KIR2DL1*001, KIR2DL2*001, KIR2DL3*001, KIR2DL4*001, KIR2DL5A*001, KIR2DS1*002, KIR2DS4*001, KIR3DL1*001, KIR3DS1*013, KIR3DL2*001, and KIR3DL3*001.

SPR measurements were conducted in phosphate-buffered saline (PBS; Corning) containing 0.005% v/v surfactant P20 (GE Healthcare) using a Biacore 3000 system (Biacore AB). To assess binding of various KIR-Fc constructs (R&D) to various GAGs, biotinylated hyaluronate (molecular weight = 29 kDa), biotinylated heparin (molecular weight = 18 kDa) and biotinylated chondroitin sulfate (molecular weight = 50 kDa) (all from Creative PEGWorks), were immobilized onto individual flow cells of an SA (streptavidin) sensor chip (GE Healthcare) until saturation. A blank flow cell with no immobilized ligand was used as a reference flow cell. Injections of 60 μL of KIR-Fc constructs diluted in PBS to 25 μg/mL were performed at a flow rate of 20 μL/min, with a subsequent 10 min run of buffer to allow sufficient dissociation. Although not presented here, regeneration after each injection was achieved with two pulses of 100 μL of 0.2 M sodium hydroxide (NaOH) (GE Healthcare) at a flow rate of 100 μL/min. Raw sensograms were corrected by double referencing (subtracting from the reference flow cell response and from PBS injection response). All experiments were done at standard temperature (25°C).

Flow cytometry data was acquired on BD LSR II or LSR Fortessa and analyzed using FlowJo software version 10.1 (FlowJo) and analyses were performed using GraphPad Prism 6 (GraphPad Software). Kinetic analyses of SPR data were performed using TraceDrawer 1.9.2 software (Ridgeview Instruments AB) using bivalent analysis model (due to use of dimeric KIR-Fc) to calculate monomeric affinities.