Degenerate Recognition of MHC Class I Molecules with Bw4 and Bw6 Motifs by a Killer Cell Ig-like 3DL Expressed by Macaque NK Cells This information is current as of September 29, 2021. Sebastien M. Maloveste, Dan Chen, Emma Gostick, Julian P. Vivian, Ronald J. Plishka, Ranjini Iyengar, Robin L. Kruthers, Alicia Buckler-White, Andrew G. Brooks, Jamie Rossjohn, David A. Price and Bernard A. P. Lafont

J Immunol published online 5 October 2012 Downloaded from http://www.jimmunol.org/content/early/2012/10/05/jimmun ol.1201360 http://www.jimmunol.org/ Supplementary http://www.jimmunol.org/content/suppl/2012/10/05/jimmunol.120136 Material 0.DC1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on September 29, 2021

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published October 5, 2012, doi:10.4049/jimmunol.1201360 The Journal of Immunology

Degenerate Recognition of MHC Class I Molecules with Bw4 and Bw6 Motifs by a Killer Cell Ig-like Receptor 3DL Expressed by Macaque NK Cells

Sebastien M. Maloveste,* Dan Chen,* Emma Gostick,† Julian P. Vivian,‡ Ronald J. Plishka,x Ranjini Iyengar,x Robin L. Kruthers,x Alicia Buckler-White,x Andrew G. Brooks,{ Jamie Rossjohn,‡ David A. Price,† and Bernard A. P. Lafont*

The killer cell Ig-like receptors (KIRs) expressed on the surface of NK cells recognize specific MHC class I (MHC-I) molecules and regulate NK cell activities against pathogen-infected cells and neoplasia. In HIV infection, survival is linked to host KIR and MHC-I genotypes. In the SIV macaque model, however, the role of NK cells is unclear due to the lack of information on KIR–MHC interactions. In this study, we describe, to our knowledge, the first in-depth characterization of KIR–MHC inter- Downloaded from actions in pigtailed macaques (Macaca nemestrina). Initially, we identified three distinct subsets of macaque NK cells that stained ex vivo with macaque MHC-I tetramers loaded with SIV peptides. We then cloned cDNAs corresponding to 15 distinct KIR3D alleles. One of these, KIR049-4, was an inhibitory KIR3DL that bound MHC-I tetramers and prevented activation, degranulation, and cytokine production by macaque NK cells after engagement with specific MHC-I molecules on the surface of target cells. Furthermore, KIR049-4 recognized a broad range of MHC-I molecules carrying not only the Bw4 motif, but also Bw6 and non- Bw4/Bw6 motifs. This degenerate, yet peptide-dependent, MHC reactivity differs markedly from the fine specificity of human http://www.jimmunol.org/ KIRs. The Journal of Immunology, 2012, 189: 000–000.

atural killer cells are critical components of the innate and inhibitory signals (5–7). In primates, these receptors comprise that contribute to protection against members of the C-type lectin-like family (NKG2A-E) and a num- N intracellular pathogens and neoplasia (1–4). The prin- ber of Ig superfamily molecules, such as leukocyte Ig-like recep- cipal determinant of NK cell activation and function is the ability tors (LILRs) and killer cell Ig-like receptors (KIRs) (8). MHC-I to discriminate between healthy cells and the altered signature of molecules constitute a major ligand for several of these receptors, pathogen-infected or transformed cells. In particular, viral infec-

including KIRs (9). by guest on September 29, 2021 tion and neoplastic processes frequently downregulate the cell KIR molecules are encoded by several highly polymorphic surface expression of MHC class I (MHC-I) molecules to escape that produce cell surface with either two (KIR2D) or three + immune pressure mediated by CD8 T lymphocytes. These (KIR3D) extracellular Ig-like domains. In addition, the transmem- phenotypic changes can trigger NK cell activation. brane domain and cytoplasmic tail occur in two distinct forms. KIR NK cells recognize MHC-I molecules via a large array of ger- molecules with long tails (KIR2DL, KIR3DL) provide inhibitory mline-encoded cell surface receptors that provide both activating signals due to the presence of ITIMs (10). In contrast, KIR mol- ecules with short tails (KIR2DS, KIR3DS) provide activating *Non-Human Primate Immunogenetics and Cellular Immunology Unit, Laboratory signals mediated by the adaptor DAP12, which associates of Molecular Microbiology, National Institute of Allergy and Infectious Diseases, with a charged residue in the KIR transmembrane domain (5). National Institutes of Health, Bethesda, MD 20892; †Institute of Infection and Im- munity, Cardiff University School of Medicine, Cardiff, CF14 4XN Wales, United In humans, KIRs possess fine specificity for defined groups of Kingdom; ‡Department of Biochemistry and Molecular Biology, School of Biomedical MHC-I molecules characterized by specific motifs at the C-terminal Sciences, Monash University, Clayton, Victoria 3800, Australia; xLaboratory of Mo- lecular Microbiology Core, Laboratory of Molecular Microbiology, National Institute end of the a1 helix. KIR2DL1 recognizes HLA-C molecules that of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD possess asparagine and lysine at positions 77 and 80, respectively. { 20892; and Department of Microbiology and Immunology, University of Melbourne, KIR2DL2 and KIR2DL3 recognize all the other HLA-C mole- Parkville, Victoria 3010, Australia cules, which harbor serine and asparagine at positions 77 and 80, Received for publication May 16, 2012. Accepted for publication September 2, 2012. respectively. In contrast, KIR3DL1 specifically recognizes HLA-B This work was supported by the Intramural Research Program of the National Insti- tute of Allergy and Infectious Diseases, National Institutes of Health, the National molecules with a Bw4 motif (NLR[I/T]ALR) between positions Health and Medical Research Council of Australia (NHMRC), and the Association 77 and 83. Other HLA-B molecules, which lack the Bw4 motif, for International Cancer Research (J.R. and A.G.B.); J.R. is supported by an NHMRC harbor the distinct Bw6 sequence (SLRNLRG); this is not recog- Australia Fellowship. nized by any human KIRs. The vast majority of HLA-A molecules The sequences presented in this article have been submitted to GenBank (http://www. ncbi.nlm.nih.gov/genbank/) under accession number HQ713453–HQ713467. carry neither the Bw4 nor the Bw6 motif and are not recognized Address correspondence and reprint requests to Dr. Bernard A. P. Lafont, Laboratory by KIRs, except for HLA-A*03 and HLA-A*11, which bind of Molecular Microbiology, National Institute of Allergy and Infectious Diseases, KIR3DL2 in a peptide-dependent manner (11). The ligands for National Institutes of Health, 4 Center Drive, Bethesda, MD 20892. E-mail address: most activating KIRs are currently unknown. [email protected] The polymorphic KIR genes have been shown to play a critical The online version of this article contains supplemental material. role during viral infection. For example, genetic studies have dem- Abbreviations used in this article: KIR, killer cell Ig-like receptor; LILR, leukocyte Ig-like receptor; pMHC-I, peptide–MHC class I; RT, room temperature; SHIV, SIV/ onstrated that the presence of KIR2DL3 in hepatitis C virus- HIV chimera. infected patients correlates with improved viral control (12).

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1201360 2 DEGENERATE KIR–MHC INTERACTION IN MACAQUES

Similarly, specific combinations of MHC-I (HLA alleles harboring KIR3D cloning and sequencing the Bw4-80I motif) and KIR (KIR3DL1/KIR3DS1) genes are as- Macaque KIR3D cDNAs were cloned, as described previously (28). Briefly, sociated with slower progression to AIDS in HIV-1–infected 1 mg total RNA, extracted from macaque PBMCs using TRI Reagent patients (13, 14). Furthermore, KIR3DL1/S1+ NK cells expand (Molecular Research Center, Cincinnati, OH), was used to synthesize in vivo during acute HIV-1 infection in individuals carrying HLA- cDNA using random hexamers and Superscript III reverse transcriptase Bw4 alleles (15). However, the precise mechanism of protection is (Invitrogen, Carlsbad, CA), according to the manufacturer’s instructions. KIR cDNAs were amplified by PCR using KIR-S1 (59-CAGCACCATG- still unclear, and ligands for KIR3DS1 have yet to be identified. TCGCTCAT-39) sense and KIR-R1 (59-GGGGTCAAGTGAAGTGGAGA- Studies in animal models could help to elucidate the protective 39) reverse primers with high fidelity Phusion polymerase (New England role of NK cells during HIV-1 infection. Biolabs, Ipswich, MA). The PCRs were heated at 98˚C for 30 s, and then Infections of Asian macaques with SIV or SIV/HIV chimeras amplifications were conducted over 28 cycles of 98˚C for 5 s, 63˚C for 1 s, and 72˚C for 20 s. A final extension was conducted at 72˚C for 5 min. The (SHIV) are classical primate models of HIV disease. Rhesus ma- PCR product (∼1.6 kb) was gel purified using the Qiaquick gel extraction caques (Macaca mulatta), pig-tailed macaques (Macaca nem- (Qiagen, Valencia, CA) and cloned into the pCR4Blunt-TOPO vector estrina), and cynomolgus macaques (Macaca fascicularis) are the (Invitrogen). Between 24 and 60 individual clones were sequenced per most commonly used species for this purpose. In these animals, animal using a 3130xl Genetic Analyzer (Applied Biosystems, Foster City, SIV or SHIV infection recapitulates the pathology and disease CA). The sequences of the pig-tailed macaque KIR3D alleles described in this study have been deposited in GenBank (http://www.ncbi.nlm.nih.gov/ progression observed in HIV-1–infected humans (16). The orga- genbank/) under accession number HQ713453–HQ713467. nization of MHC genes in these three species is markedly different compared with the human MHC region. Specifically, macaques Phylogenetic analysis lack the MHC-C , but possess several copies of both the KIR3DL and KIR3DS sequences were aligned separately using the Cluster Downloaded from MHC-A and MHC-B genes. These are called Mamu, Mane, or W program in the MacVector 11.1.2 software suite (MacVector, Cary, NC) Mafa genes in rhesus, pig-tailed, and cynomolgus macaques, re- with minor manual adjustments. Phylogenetic trees were constructed using the neighbor-joining method. Genetic distances were estimated using spectively. Numerous macaque MHC-I alleles have been sequenced, Kimura’s two-parameter method (37). Bootstrap analysis (1000 replicates) and immunogenic viral epitopes presented by the expressed pro- was performed to assign confidence values to tree nodes (38). teins have been identified in some cases (17–22). In addition, Cell lines certain macaque MHC-I alleles are associated with lower viral http://www.jimmunol.org/ loads and protection against SIV/SHIV disease development (23– The MHC-I–deficient cell line 721.221 (39), provided by E. Long (Lab- 25). oratory of Immunogenetics, National Institute of Allergy and Infectious The role of NK cells during SIV/SHIV infection in macaques Diseases, National Institutes of Health, Rockville, MD), was maintained in RPMI 1640 medium supplemented with 10% heat-inactivated FBS, 2 mM remains unclear. NK cell depletion experiments have failed to L-glutamine, 100 U/ml penicillin-G, and 100 U/ml streptomycin (R10 demonstrate a significant impact on SIV load and disease pro- medium). gression, potentially due to partial targeting of all NK cells (26). To generate target cells expressing a single macaque MHC-I allele, full- However, numerous KIR alleles have been identified in rhesus and length cDNAs encoding the pig-tailed macaque MHC-I alleles Mane- A1*003:01, -A1*082:01, -A1*084:01, -A3*13:01, and -B*109:01 were cynomolgus macaques (27–30), and some studies have reported an cloned into the EcoRI restriction site of pcDNA3.1(+) (Invitrogen). Two effect of specific KIR alleles on viral load, independently of MHC million 721.221 cells were nucleofected with 2 mg plasmid DNA using the by guest on September 29, 2021 background (31, 32). Amaxa Cell Line Nucleofector kit V (Lonza, Cologne, Germany) and the Although macaque MHC and KIR sequence diversity has been T-020 program on a Nucleofector II device (Lonza), and then cultured at studied extensively, little is known about macaque KIR specificity 37˚C in R10 medium lacking antibiotics. At 20 h postnucleofection, cell surface expression of MHC-I molecules was verified by flow cytometry for MHC-I ligands. Only two studies have identified specific KIR– using a PE-conjugated MHC-I–specific mAb (clone W6/32; AbD Serotec, MHC pairs in rhesus macaques (33, 34). Colantonio et al. (34) Raleigh, NC). Cells were cultured for 1 wk in R10 medium supplemented demonstrated that KIR3DL05 recognizes Mamu-A1*002. Simi- with 0.70 mg/ml geneticin (Invitrogen), and then positively selected for larly, Rosner et al. (33) showed that KIR3DL05 and KIR3DLw03 MHC-I expression on the cell surface using PE-conjugated anti–MHC-I mAb and MACS anti-PE microbeads (Miltenyi Biotec, Auburn, CA), as interact with several Mamu-A, but not with Mamu-B molecules specified by the manufacturer. Stable cell lines were maintained afterward in vitro. These KIRs also interact strongly with human HLA-C in R10 medium supplemented with 0.70 mg/ml geneticin (Invitrogen). molecules (35). Macaque KIR3D cDNAs were modified by PCR to insert a Flag epitope In this study, we describe, to our knowledge, the first in-depth (DYKDDDDK) between the leader peptide and the D0 domain. Modified characterization of KIR–MHC interactions in pig-tailed macaques. KIR3D cDNAs were cloned into the EcoRI restriction site of pcDNA3.1(+) (Invitrogen). Plasmids encoding macaque Flag-KIR3D constructs were The identified receptor, named KIR049-4, is a member of the nucleofected into the 721.221 cell line, as described for MHC-I alleles. At inhibitory KIR3DL family and exhibits broad peptide-dependent 20 h postnucleofection, cell surface expression of Flag-KIR3D molecules reactivity against both Bw4 and non-Bw4 MHC-I molecules. was verified by flow cytometry using a PE-conjugated Flag-specific mAb (clone M2; Abcam, Cambridge, MA). Cells were cultured for 1 wk in R10 medium supplemented with 0.70 mg/ml geneticin (Invitrogen), and then positively selected for Flag expressiononthecellsurfaceusingPE- Materials and Methods conjugated anti-Flag mAb and MACS anti-PE microbeads (Miltenyi Animals and viruses Biotec), as described for the MHC-I cell lines. Stable cell lines were Pig-tailed macaques were maintained in accordance with the Guide for the maintained afterward in R10 medium supplemented with 0.70 mg/ml Care and Use of Laboratory Animals (36). Animal handling and experi- geneticin (Invitrogen). ments were approved by the National Institute of Allergy and Infectious MHC genotyping Diseases Institutional Animal Care and Use Committee. Macaques were housed in a biosafety level 2 facility. Biosafety level 3 practices were Mane-A1*082 genotyping was performed on genomic DNA extracted followed. Macaques were anesthetized with i.m. injections of ketamine from macaque PBMCs using the Wizard Genomic DNA purification kit hydrochloride (Ketaject; Phoenix Pharmaceutical, St. Joseph, MO) and (Promega, Madison, WI). Briefly, a sequence-specific primer-PCR ampli- acepromazine acetate (Fermenta Animal Health, Kansas City, MO) during fication was performed using Mane-A3S2 sense (59-CAACACACAGAC- phlebotomy and virus inoculations. Ten macaques had been inoculated i.v. CTACCGAGAGAA-39) and Mane-A3R reverse (59-CTCTGATCTGCTC- several years earlier with CXCR4-tropic SHIVof low pathogenicity. At the CGCCG-39) primers with AmpliTaq DNA polymerase (Perkin Elmer, time of study, plasma RNA viral loads were below the limit of detection in Wellesley, MA). An internal control for amplification was obtained using all infected macaques, and CD4+ T lymphocyte counts were within the Mane-DRBS3 sense (59-GAGTGTCATTTCTTCAACGGGA-39)and normal range. Mane-DRBR2 reverse (59-CCTCGCCGCTGCACTGT-39)primers.The The Journal of Immunology 3

PCRs were heated at 94˚C for 5 min, and then amplifications were con- or anti–TNF-a (clone mAb11-allophycocyanin; BD Biosciences) mAb for ducted over 25 cycles of 94˚C for 30 s, 65.5˚C for 30 s, and 72˚C for 30 s. 30 min at 4˚C. After two final washes with PBS, cells were resuspended in A final extension was conducted at 72˚C for 7 min. PCRs were loaded onto PBS containing 1% formaldehyde. a 1.5% agarose gel and separated by electrophoresis. A 246-bp Mane- To monitor NK cell degranulation, macaque PBMCs were incubated for DRB–specific product was amplified in all macaques, whereas a 493-bp 5 h in the presence of anti-CD107a and anti-CD107b mAbs (clone H4A3- product was generated in Mane-A1*082+ macaques. Mane-A1*082- FITC and clone H4B4-FITC; BD Biosciences) either alone or at a NK:target specific amplicons were gel purified using the Qiaquick gel extraction kit ratio of 1:1, with parental 721.221 cells or 721.221 cell lines expressing (Qiagen) and directly sequenced using a 3130xl Genetic Analyzer (Ap- Mane-A or Mane-B alleles. Monensin (GolgiStop; BD Biosciences) was plied Biosystems). Mane-A4*14 genotyping was performed similarly, as added for the last 4 h. Cells were then washed once with PBS, labeled with described previously (18). PE-conjugated pMHC-I tetramers for 15 min at RT, and stained with anti- CD3 (clone SP34-allophycocyanin; BD Biosciences) and anti-CD8 (clone Tetramer production SK1-PerCP; BD Biosciences) mAbs for 30 min at 4˚C. After two further Fluorochrome-conjugated tetrameric peptide–MHC class I (pMHC-I) com- washes with PBS, cells were resuspended in PBS containing 1% formal- plexes were produced in-house, as described previously (40). dehyde. For all assays, at least 300,000 events were acquired per test using Tetramer staining of macaque PBMCs a FACSCalibur flow cytometer (BD Biosciences). Data were analyzed with FlowJo software version 9.3.2 (Tree Star). One million freshly isolated macaque PBMCs were labeled with pMHC-I tetramers conjugated to either PE or allophycocyanin for 15 min at room Results temperature (RT) in the dark. Cells were labeled further with anti-CD3 2 + (clone SP34-FITC, clone SP34-2-PE, or clone SP34-2-allophycocyanin; Macaque CD3 CD8 lymphocyte subsets bind pMHC-I BD Biosciences), anti-CD8 (clone SK1-PerCP; BD Biosciences), anti- tetramers ex vivo

CD14 (clone RMO52-FITC or clone RMO52-allophycocyanin; Beckman + Downloaded from Coulter), and anti-CD20 (clone B9E9-FITC or clone B9E9-allophyco- In the SIV/SHIV pig-tailed macaque model, CD8 T cell responses cyanin; Beckman Coulter) mAbs for 30 min at 4˚C, washed twice with specific for the SIV Gag-derived epitope KP9 (KKFGAEVVP) PBS, and resuspended in PBS containing 1% formaldehyde. At least contribute to viral control and improved survival in monkeys car- 200,000 events were acquired per test using a FACSCalibur flow cytometer rying the Mane-A1*084 allele (formerly known as Mane-A*10) (BD Biosciences). Data were analyzed with FlowJo software version 9.3.2 (20). We identified a different Gag-derived epitope (DI9: (Tree Star, Ashland, OR). DHQAAMQII) presented by the Mane-A1*082 allele (formerly Tetramer staining of KIR-expressing cell lines known as Mane-A*03). Mane-A1*082+ macaques infected with http://www.jimmunol.org/ + One million 721.221 parental cells and KIR-expressing cell lines were SIV/SHIV were found to harbor CD8 T lymphocytes specific for stained with PE-conjugated pMHC-I tetramers for 30 min at RT in the dark. this epitope, as measured by intracellular staining for IFN-g and KIR surface expression was assessed independently for each cell line by TNF-a (data not shown). Although KP9-specific CD8+ T cell staining with a PE-conjugated Flag-specific mAb (clone M2; Abcam) for responses were detected in all Mane-A1*084+ SIV/SHIV-infected 30 min at 4˚C. Cell samples were washed twice with PBS, resuspended + in PBS, and stained with 7-aminoactinomycin D (Via-Probe; BD Bio- macaques tested, only a fraction of infected Mane-A1*082 ma- sciences) for 10 min at RT in the dark prior to acquisition. At least 60,000 caques responded to the DI9 epitope. events were acquired per test using a FACSCalibur flow cytometer (BD To investigate whether this poor response frequency was due to Biosciences). Data were analyzed with FlowJo software version 9.3.2 the lack of peptide-specific CD8+ T cells, we stained PBMCs from (Tree Star). by guest on September 29, 2021 Mane-A1*082+ SHIV-infected macaques directly ex vivo using Immunophenotyping of macaque NK cells Mane-A1*082 tetramers loaded with the DI9 peptide or an N- terminal truncated variant (HI8: HQAAMQII). Each tetramer Macaque peripheral blood samples were stained at RT with mAbs against + + CD3 (clone SP34-FITC; BD Biosciences), CD8 (clone SK1-PerCP; BD detected a subset of CD8 T cells in Mane-A1*082 macaques that Biosciences), CD14 (clone RMO52-allophycocyanin; Beckman Coulter), exhibited IFN-g responses upon in vitro stimulation with the DI9 CD20 (clone B9E9-allophycocyanin; Beckman Coulter), and NK markers peptide (Fig. 1A). We did not detect specific tetramer staining of coupled to PE. The following mAbs were used: 1) anti-CD16 clone 3G8, CD8+ T cells from Mane-A1*082+ SHIV-infected macaques that anti-CD56 clone MY31, anti-CD2 clone S5.2, anti-NKG2D clone 1D11, anti-CD158a clone HP-3E4, CD158b clone CH-L, and anti-CD158 clones failed to produce IFN-g upon DI9 peptide stimulation, suggesting DX27 and DX9 (BD Biosciences); 2) anti-NKp30 clone Z25, anti-NKp46 that the lack of response was due to the absence of peptide-specific clone BAB281, anti-NKp80 clone MA152, anti-NKG2A clone Z199, and CD8+ T cells. However, in the same samples, a significant fraction anti-CD158b clone GL183 (Beckman Coulter); 3) anti-CD161 clone HP- (13.9–24.7%) of CD32CD8+ cells stained with the Mane-A1*082 3G10, anti-LILRA2 clone 135.4, anti-LILRA4 clone eBio17G10.2, anti- 2 tetramers (Fig. 1B). Both tetramers recognized the same CD3 LILRB1 clone HP-F1, anti-LILRB2 clone 42D1, anti-LILRB3 clone + MKT5.7H5.1, and anti-LILRB4 clone ZM4.1 (eBioscience); and 4) anti- CD8 lymphocyte subset, as demonstrated by costaining experi- KIR2D clone NKVFS1 (Miltenyi Biotec). After 15 min, RBCs were lysed ments using Mane-A1*082 tetramers conjugated with different with FACS lysing solution (BD Biosciences) for 10 min, and the samples fluorochromes (Fig. 1C), suggesting that subsets of CD32CD8+ were washed twice with PBS. Cells were then resuspended in PBS con- lymphocytes express a receptor capable of binding pMHC-I com- taining 1% formaldehyde. At least 200,000 events were acquired per test using a FACSCalibur flow cytometer (BD Biosciences). Data were analyzed plexes. using CellQuest Pro software version 6.0 (BD Biosciences). Tetramer-binding CD32CD8+ lymphocyte subsets are present Functional assays irrespective of MHC genetic background or infection status To monitor NK cell activation or TNF-a production by NK cells, freshly Next, we analyzed freshly isolated PBMCs from a cohort of 20 pig- isolated macaque PBMCs were incubated for 5 h in the presence of brefeldin tailed macaques that included 10 Mane-A1*082+ animals (50%). A (GolgiPlug; BD Biosciences), either alone or at a NK:target ratio of Ten macaques (50%) had been infected with a CXCR4-tropic 1:1 based on the percentage of NK cells in the PBMC sample, with pa- SHIV, and 10 macaques (50%) were uninfected. We identified rental 721.221 cells or 721.221 cell lines expressing Mane-A or Mane-B 2 + alleles. Cells were subsequently washed once with PBS, labeled with either Mane-A1*082 tetramer-binding CD3 CD8 lymphocytes in seven PE-conjugated pMHC-I tetramers for 15 min at RT or PE-conjugated anti- animals, but the presence of these cells was independent of infec- KIR2D mAb (clone NKVFS1; Miltenyi Biotec) for 30 min at 4˚C, and tion status or Mane-A1*082 genotype (Table I). Five macaques then stained with anti-CD3 (clone SP34-FITC; BD Biosciences) and anti- harbored CD32CD8+ lymphocytes that stained similarly with both CD8 (clone SK1-PerCP; BD Biosciences) mAbs for 30 min at 4˚C. After Mane-A1*082 tetramers. Two additional macaques possessed a two further washes with PBS, the cells were permeabilized for 10 min 2 + using Permeabilization Solution 2 (BD Biosciences), washed with PBS, subset of CD3 CD8 cells that stained with the HI8/Mane-A1*082 and stained with anti-CD69 (clone FN50-allophycocyanin; BD Biosciences) tetramer, but not with the DI9/Mane-A1*082 tetramer. 4 DEGENERATE KIR–MHC INTERACTION IN MACAQUES Downloaded from

FIGURE 1. pMHC-I tetramers bind macaque CD32CD8+ lymphocytes ex vivo. (A) Detection of virus-specific CD8+ T cells with Mane-A1*082 tet- ramers in a SHIV-infected, Mane-A1*082+ macaque (PT1246A). (B) Two Mane-A1*082 tetramers bind subsets of CD32CD8+ lymphocytes in a SHIV- infected, Mane-A1*082+ macaque (PT93P049) lacking CD8+ T cell responses to the corresponding peptide epitopes. Lymphocytes are gated as CD32CD8+ after exclusion of CD20+ and CD14+ cells. (C) Costaining of CD32CD8+ lymphocytes with HI8/Mane-A1*082 and DI9/Mane-A1*082 tetramers (PT1670). (D) Costaining of CD32CD8+ lymphocytes with HI8/Mane-A1*082 and AF9/Mane-A4*14 (PT98P033). http://www.jimmunol.org/ To expand our observations, we used Mane-A4*14 (formerly The receptors responsible for binding the AF9/Mane-A4*14 and known as Mane-A*17) tetramers loaded with another Gag-derived Mane-A1*082 tetramers are clearly distinct. Moreover, costaining peptide (AF9: ALAPVPIPF) (41). Four macaques harbored a experiments of CD32CD8+ lymphocytes from one animal subset of CD32CD8+ lymphocytes that bound the AF9/Mane- (PT98P033) with both AF9/Mane-A4*14 and HI8/Mane-A1*082 A4*14 tetramer. As with the Mane-A1*082 tetramers, the pres- tetramers revealed distinct but overlapping subsets of cells reac- ence of a Mane-A4*14 tetramer-binding CD32CD8+ lymphocyte tive with each pMHC-I complex (Fig. 1D). subset was independent of infection status and MHC genotype Thus, macaque CD32CD8+ lymphocytes express receptors with (Table I). diverse specificity for pMHC-I complexes. Indeed, three distinct patterns of tetramer binding were observed, most likely resulting by guest on September 29, 2021 from the expression of three different receptors. One receptor rec- ognizes both Mane-A1*082 complexes, but not AF9/Mane-A4*14. Table I. The presence of tetramer-binding NK cell subsets is A second receptor binds Mane-A1*082 in a peptide-dependent independent of MHC-I genotype and infection status manner. A third receptor is specific for the Mane-A4*14 molecule. Peptide–MHC-I tetramers bind NK cell subsets directly ex vivo Tetramer Staining (NK %) The observation of pMHC-I tetramer staining within the CD32 MHC Genotype A1*082 A4*14 CD8+ cell subset was somewhat unexpected because these reagents Infection are typically used to analyze epitope-specific CD8+ T cells, which Macaque Status A1*082 A4*14 DI9 HI8 AF9 express CD3 as part of the TCR complex. Thus, the absence of CD3 PT1670 Yes Yes No 15.9 17.5 1.4 expression by tetramer-binding cells suggested the involvement PT93P049 Yes Yes No 10.7 20.2 0.7 of a different hematopoietic lineage. We hypothesized that the PT98P021 No Yes No 17.4 22.3 1.6 tetramer-positive cells were NK cells, which do express CD8a,but PT98P056 Yes No No 13.4 14.9 4.8 PT97P027 No No No 18.7 24.0 1.3 lack CD3 molecules. PT99P004 No Yes No 1.1 8.5 3.4 To evaluate the role of macaque NK cells in pMHC-I tetramer PT98P033 Yes No No 1.1 9.0 12.5 binding, we performed immunophenotyping assays on peripheral PT96P067 No Yes No 0.8 0.9 21.4 blood samples from 20 animals using a panel of 10 distinct markers PT98P050 Yes No Yes 0.4 0.5 8.0 (Supplemental Table I). First, CD16 and CD56, two markers used PT93549 Yes No Yes 0.1 0.5 7.9 to distinguish human NK cell subsets, were evaluated (42). A large PT1246A Yes Yes No 0.7 2.6 0.7 2 PT98P031 No Yes No 0.4 0.6 1.3 majority of CD3 CD8+ cells expressed high levels of CD16, but PT99P003 No Yes No 1.4 1.6 1.9 only a small fraction was positive for CD56, consistent with PT210A Yes Yes Yes 0.3 0.7 0.9 previous studies of rhesus macaque NK cells (43). Almost all PT99P015 No Yes Yes 0.6 1.1 0.5 2 + PT93518 Yes No Yes 3.1 3.2 3.1 CD3 CD8 lymphocytes expressed CD2, an adhesion molecule PT96P062 No No Yes 0.1 0.1 3.8 expressed by both T lymphocytes and NK cells. Among natural PT98P049 No No Yes 0.1 0.4 0.4 cytotoxicity triggering receptors, NKp30 and NKp46 were ex- PTA0P045 No No Yes 0.3 0.4 0.6 pressed on most CD32CD8+ cells. In contrast, NKp44, a receptor PT1669 Yes No No 2.9 4.4 1.4 expressed on tissue NK cells, was not detected (mean 0.0% 6 2 2 2 CD14 CD20 CD3 CD8+ lymphocytes were assessed for HI8/Mane-A1*082, 0.1%; data not shown). CD32CD8+ lymphocytes also expressed DI9/Mane-A1*082, and AF9/Mane-A4*14 tetramer staining. Data represent the high levels of the lectin-like receptors NKp80, NKG2A, and mean value of up to five independent experiments. Values in bold represent tetra- 2 mer-binding NK cell subsets. NKG2D (Supplemental Table I). CD161 expression by CD3 The Journal of Immunology 5

CD8+ cells was more variable between individual macaques. The expression of CD16, NKp30, and NKp46 was almost exclusively detected on CD32CD8+ cells and was virtually absent on CD3+ CD8+ cells (Supplemental Table I); the other markers tested were less specific for NK cells. Next, we examined expression of six distinct LILR receptors (LILRA2, LILRA4, LILRB1, LILRB2, LILRB3, LILRB4) on macaque cells. A mAb specific for LILRB3 (CD85a) reacted with monocytes only (data not shown). The five additional human LILR- specific mAbs did not cross-react with macaque molecules. Among six distinct mAbs specific for human KIR2D (HP-3E4, GL183, CH- L, DX27, NKVFS1) and KIR3D (DX9) molecules, only one of them (NKVFS1) was reactive with macaque cells (Supplemental Table I and data not shown). KIR2D was expressed on a subset of CD32CD8+ cells that displayed variable frequencies between macaques (mean 22.1 6 15.6%; range 0.3–68.2%). Three distinct staining patterns were observed. Intense staining of KIR2D+ cells was observed in seven macaques (35%), and dimmer staining was detected in eight ad- Downloaded from ditional macaques (40%) (data not shown). Five macaques (25%) harbored both bright and dim KIR2D+ subsets within the CD32 CD8+ cell population, suggesting that NKVFS1 recognizes mul- tiple macaque KIR2D molecules that are either differentially expressed or bind the mAb with distinct affinities. 2 + Collectively, these findings indicate that the CD3 CD8 cell http://www.jimmunol.org/ population comprises NK cells rather than CD8+ T cells with downregulated CD3 expression (Supplemental Table I). The high frequency of tetramer-binding cells observed in uninfected ma- caques further argues against a T cell origin. However, none of the markers tested displayed concordant expression profiles that could explain the observed patterns of tetramer reactivity. Cloning of KIR3D alleles from pig-tailed macaques

To identify the receptors responsible for pMHC-I tetramer binding, by guest on September 29, 2021 + + we investigated the distribution of tetramer-binding cells among FIGURE 2. pMHC-I tetramer staining of CD16 and KIR2D NK cell A CD16+ and KIR2D+ NK cells. The Mane-A1*082 and Mane- subsets. ( ) Patterns of pMHC-I tetramer staining as a function of CD16 expression. (B) Patterns of pMHC-I tetramer staining as a function of A4*14 tetramer-binding cells were found mainly among CD16+ KIR2D expression. In each case, representative stainings from two ma- NK cells, which are known to preferentially express KIR mole- caques (PT1670, top and middle panels; PT98P033, bottom panel) are cules in humans (44) (Fig. 2A). However, tetramer staining did not shown. Cells are gated as CD32CD8+ lymphocytes. cosegregate with KIR2D expression (Fig. 2B). These results led us to consider KIR3D molecules. On this basis, we cloned KIR3D alleles from four pig-tailed in rhesus and cynomolgus macaques. However, there was no ev- macaques, including one animal (PT93P049) with NK cells that idence for the conservation of activating KIR loci among macaque bound both Mane-A1*082 tetramers and one animal (PT98P033) species, as no clustering of KIR3DS alleles was observed (data not with NK cells that bound both the HI8/Mane-A1*082 and AF9/ shown). Mane-A4*14 tetramers. We identified nine KIR3DL and six KIR3DS alleles in this small group of macaques, with each animal The KIR3DL molecule KIR049-4 recognizes Mane-A1*082 expressing multiple activating and inhibitory receptors. All KIR3DL To identify KIR3D alleles encoding pMHC-I tetramer-binding molecules possessed long cytoplasmic tails containing two receptors, we expressed four KIR3DS and four KIR3DL ITIMs (Fig. 3). In contrast, the KIR3DS molecules had very cDNAs, the latter from distinct allelic clusters, in 721.221 cells. short cytoplasmic tails, and their transmembrane domains har- Due to the lack of available mAbs specific for macaque KIR3D bored an arginine residue that was absent from the KIR3DL molecules, we inserted a Flag epitope between the leader peptide molecules (Fig. 3). and the first extracellular Ig-like domain (D0), and established Next, we compared pig-tailed macaque KIR3DL and KIR3DS stable cell lines expressing high levels of each KIR3D molecule on allele sequences by phylogenetic analysis with alleles previously the cell surface (Fig. 5A, top row). Whereas the KIR3DL alleles identified from rhesus and cynomolgus macaques. Pig-tailed ma- were expressed at high levels and remained stable, expression of caque KIR3DL sequences intermingled with alleles from the two the KIR3DS alleles was somewhat lower and generally less stable. other macaque species to form five distinct clusters containing Each cell line was tested for its ability to bind Mane-A1*082 and representatives from each macaque species (Fig. 4). This obser- Mane-A4*14 tetramers. None of the KIR3D cell lines bound the vation suggests the presence of five conserved KIR3DL loci AF9/Mane-A4*14 tetramer above background levels observed among pig-tailed, rhesus, and cynomolgus macaques. Similar with parental 721.221 cells (Fig. 5B). Cells expressing KIR049-4, phylogenetic comparisons performed with KIR3DS alleles dem- however, bound both Mane-A1*082 tetramers (Fig. 5A [bottom onstrated that the sequence diversity of pig-tailed macaque row], 5B). Consistent with these findings, the KIR049-4 allele KIR3DS alleles was comparable to that of KIR3DS alleles found was cloned from a macaque (PT93P049) with NK cells reactive 6 DEGENERATE KIR–MHC INTERACTION IN MACAQUES

FIGURE 3. Predicted protein sequences of the transmembrane and cytoplasmic domains of pig-tailed macaque KIR3D alleles. The predicted sequences of the transmembrane domain and cytoplasmic tails of nine KIR3DL and six KIR3DS molecules from pig-tailed macaques are aligned with the corre- sponding domains of human KIR3DL1*001 and KIR3DS1*010 molecules. Amino acids identical to residues present in KIR3DL1*001 are indicated by a period. Gaps introduced for the alignment are represented by dashes. The two conserved YxxL sequences forming the ITIM motif in the tail of human and

macaque KIR3DL molecules are boxed. Boxes in the transmembrane domains of human and macaque KIR3DS molecules indicate basic residues (lysine or Downloaded from arginine) required for KIR association with adaptor molecules such as DAP12. against both Mane-A1*082 tetramers (Fig. 1B). In contrast, none receptors were identified by staining with Mane-A1*082 tetramers, of the other 7 KIR3D cell lines bound either of the Mane-A1*082 and activation status was assessed for both the tetramer-binding tetramers. The KIR049-4 cell line also bound weakly to the KP9/ and nonbinding subsets (Fig. 6A, Supplemental Fig. 1A). In the

Mane-A1*084 tetramer (Fig. 5B). absence of target cells, both NK subsets exhibited low activation http://www.jimmunol.org/ levels. Stimulation with 721.221 cells or cells expressing Mane- KIR049-4 interactions with Mane-A1*082 and Mane-A1*084 A1*003, -A3*13, or -B*109 triggered high activation levels in both molecules inhibit NK cell responses NK subsets. In contrast, activation of KIR049-4+ or other Mane- The KIR049-4 allele encodes a KIR3DL molecule, which should A1*082 tetramer-binding NK cells was almost completely inhibi- generate inhibitory signals upon cognate ligand engagement. To ted after stimulation with 721.221 cells expressing Mane-A1*082 formally address this, we monitored the responses of primary or Mane-A1*084. Moreover, tetramer-negative NK cells in the same macaque NK cells after in vitro stimulation with the MHC-deficient samples maintained high levels of CD69 expression, similar to those cell line 721.221, or 721.221 variants expressing single macaque observed after stimulation with other MHC alleles (Fig. 6A, Sup- MHC-A or MHC-B alleles (Mane-A1*082, -A1*084, -A1*003, plemental Fig. 1A). Additional analyses showed that KIR2D+ cells by guest on September 29, 2021 -A3*13, -B*109). After stimulation, NK cells were analyzed for were not inhibited under the same conditions (Fig. 6B, Supple- inducedexpressionofCD69,amarkerofcellularactivation.NK mental Fig. 1A). Thus, the inhibition of NK cell activation by Mane- cells expressing KIR049-4 or other Mane-A1*082 tetramer-binding A1*082 or Mane-A1*084 was restricted to the KIR049-4+ subset. In further experiments, the specific inhibition of NK cell acti- vation by Mane-A1*082 and Mane-A1*084 was associated with reduced degranulation, as monitored by surface mobilization of CD107, and decreased TNF-a production (Fig. 6, Supplemental Fig. 1B, 1C). Collectively, these results demonstrate that KIR049- 4 or related Mane-A1*082 tetramer-binding molecules are in- hibitory KIRs that recognize specific MHC-I molecules. KIR049-4 specificity is defined by the a1 helix of MHC-I molecules To characterize KIR049-4 specificity for MHC-I binding, the five MHC-I alleles tested were divided in two groups based on their capacity to interact with KIR049-4. Accordingly, Mane-A1*082 and Mane-A1*084 were classified as binders, whereas Mane- A1*003, -A3*13, and -B*109 were classified as nonbinders. The protein sequences of all alleles were analyzed for polymorphic sites that cosegregated with the ability to bind KIR049-4. Fifty-six amino acid positions differed between the five alleles, but most of these were specific to one allele (32 positions; 57.1%). Further- more, the amino acids present at 17 additional positions (30.3%) in the binder group were also found among the nonbinder MHC-I alleles. The seven remaining positions mapped to the MHC-I a1 FIGURE 4. Phylogenetic analysis of pig-tailed macaque KIR3DL alleles. Phylogenetic analysis of KIR3DL alleles from pig-tailed (PT, red), and a2 helices at positions 67, 74, 76, 77, 114, 156, and 165 (Fig. rhesus (RH, blue), and cynomolgus (CY, green) macaques was performed 7). Five of these positions (67, 74, 77, 114, 156) contribute to using Kimura’s two-parameter method. The tree is rooted on one allele of peptide-binding pockets and may affect KIR–MHC-I interactions both the human KIR3DL1 and KIR3DL2 loci. Bootstrap values for 1000 by altering peptide selection. The mutation at position 165 from replicates are provided on the node. leucine, which is found in binder MHC-I alleles, to the valine The Journal of Immunology 7 Downloaded from FIGURE 5. KIR049-4 binds Mane-A1*082 tetramers. (A) KIR expression and DI9/Mane-A1*082 tetramer staining are shown for the indicated stable cell lines expressing macaque Flag-KIR3D molecules. Surface KIR expression was assessed by staining with PE-conjugated anti-Flag mAb (top row, white histograms). Staining with DI9/Mane-A1*082 tetramer is shown for the same cell lines (bottom row, white histograms). Parental 721.221 cells stained under the same conditions are depicted in gray. (B) Eight stable cell lines expressing distinct KIR3Ds were stained with the following pMHC-I tetramers: HI8/ Mane-A1*082, DI9/Mane-A1*082, KP9/Mane-A1*084, and AF9/Mane-A4*14. The relative binding index, defined as the mean fluorescence intensity of tetramer binding to the KIR cell line divided by the mean fluorescence intensity of tetramer binding to the parental 721.221 cell line stained in parallel , under identical conditions, is shown in each case. Data shown represent the mean of up to nine experiments (***p 0.0001 by one-way ANOVA). http://www.jimmunol.org/

residue present in the nonbinder MHC-I alleles is unlikely to af- or an alanine in the nonbinder MHC-I alleles. The exchange be- fect KIR binding due to the similar physicochemical properties of tween a charged residue and a small hydrophobic amino acid at these amino acids. In contrast, the glutamic acid present at posi- a position exposed on the surface of the a1 helix could affect the tion 76 in the binder MHC-I alleles is replaced by either a valine KIR3D–MHC-I interaction (Fig. 7). by guest on September 29, 2021

FIGURE 6. KIR049-4 inhibits NK cell functions in a specific MHC-I–dependent manner. (A) Activation of NK cells (left) was monitored by induction of CD69 expression after stimulation with 721.221 cells or 721.221 variants expressing Mane-A or Mane-B alleles. NK cells stimulated under identical conditions were also monitored for degranulation by surface CD107a/b staining (middle) and for the production of intracellular TNF-a (right). Coex- pression of KIR049-4 or related molecules was determined by HI8/Mane-A1*082 tetramer staining after stimulation. NK cells are gated as CD32CD8+ lymphocytes. Data shown represent the mean of five independent experiments, each performed using a sample obtained from a distinct macaque (PT1670, PT93P049, PT97P027, PT98P021, PT98P056). ***p , 0.0001, **p , 0.001 by one-way ANOVA. (B) Activation (left), degranulation (middle), and TNF-a production (right) among KIR2D+ and KIR2D2 NK cell subsets after stimulation with individual MHC-I molecules. These data include tetramer-positive cells present in both the KIR2D+ and KIR2D2 NK cell subsets (Fig. 2). Data shown represent the mean of four independent experiments, each performed using a sample obtained from a distinct macaque (PT1670, PT93P049, PT97P027, PT98P021). PT98P056 was not included because ,2% of the NK cells were KIR2D+ in this animal. Error bars indicate SD. Representative raw data are shown in Supplemental Fig. 1. 8 DEGENERATE KIR–MHC INTERACTION IN MACAQUES

molecules (HLA-B*27:05, -B*51:01, and -B*57:01) with high avidity, but failed to bind another Bw4+ HLA molecule (HLA- B*44:02). These interactions were specific for KIR049-4, as four additional KIR3DL molecules failed to bind the same tetramers (Table II). Furthermore, the interaction was confirmed using pri- mary macaque NK cells expressing KIR049-4 (data not shown). Next, we tested three HLA-B molecules bearing Bw6 motifs and two HLA-A molecules without either the Bw4 or the Bw6 motif (Table II). KIR049-4 bound to HLA-B*07:02 and HLA-A*03:02 tetramers in a peptide-dependent manner, albeit with lower avid- ities compared with the Bw4+ tetramers. No binding was detected with HLA-A*02:01, -B*08:01, and -B*35:01 tetramers loaded with different peptides. Thus, KIR049-4 can recognize MHC-I molecules bearing both Bw4 and Bw6 motifs in peptide-depen- dent manner.

Discussion FIGURE 7. Map of polymorphic sites that distinguish binder from In this study, we characterized the function of a pig-tailed macaque nonbinder MHC-I molecules. The seven polymorphic residues that differ KIR3DL molecule that recognizes a broad range of MHC-I ligands Downloaded from between binder and nonbinder macaque MHC-I molecules are represented in a peptide-dependent manner. Engagement of this receptor by in red on the HLA-B*57:01-KIR3DL1 structure (PDB: 3VH8). The MHC- its cognate ligands inhibits NK cell activation, degranulation, I H chain is shown in green, and b2-microglobulin is depicted in blue. The and cytokine production. Although the inhibitory properties of D0, D1, and D2 domains of KIR3DL1 are shown in pink, yellow, and KIR049-4 are typical of a KIR3DL molecule, its degenerate MHC-I orange, respectively. Amino acids 165–167 in the D1 domain (red) are in recognition profile contrasts with the focused recognition prop- close proximity to the side chain of residue 76 in the MHC-I molecule. erties of human KIR molecules. http://www.jimmunol.org/ In humans, the three KIR3DL genes (KIR3DL1, KIR3DL2, and KIR049-4 interacts with both the Bw4 and Bw6 motifs of MHC-I KIR3DL3) encode polymorphic receptors with distinct MHC-I Human KIR3DL1 recognizes MHC-I molecules bearing the Bw4 specificities (45). KIR3DL1 binds to MHC-I molecules carrying motif at the C-terminal end of the a1 helix, whereas KIR3DL2 the Bw4 motif at the C-terminal end of the a1 helix, particularly specificity is limited to HLA-A*03 and HLA-A*11 in a peptide- those possessing an isoleucine at position 80 (Bw4-80I) (46), dependent manner (45). Surprisingly, both macaque MHC-I whereas KIR3DL2 recognizes only HLA-A*03 and HLA-A*11 in molecules recognized by KIR049-4, Mane-A1*082 and -A1*084, a peptide-dependent manner (11). The ligands for KIR3DL3 are possess the Bw6 motif, rather than the Bw4 motif. currently unknown. Similarly, KIR2D family members recognize by guest on September 29, 2021 To assess KIR049-4 specificity with regard to Bw4/Bw6 motifs, HLA-C molecules based on the presence of motifs, called C1 and we stained the KIR049-4 cell line with multiple human tetramers C2, in the same region of the MHC-I a1 helix as the Bw4 motif comprising HLA molecules with Bw4, Bw6, and non-Bw4/Bw6 (45). There are no known human KIRs specific for MHC-I mol- motifs (Table II). KIR049-4 bound three distinct Bw4+ HLA ecules carrying the Bw6 motif, which represent almost 50% of all

Table II. KIR049-4 interacts with both the Bw4 and Bw6 motifs of MHC-I in a peptide-dependent manner

HLA Peptide Origin/Sequence KIR033-2a KIR033-7 KIR049-4 KIR049-6 KIR049-7 Bw4-80I B*51:01 HIV Env YPPPIQGVI 1.3b 2.0 69.1 1.0 0.7 B*57:01 HIV Gag ISPRTLNAW 1.2 1.8 79.0 1.2 0.9 HIV Gag TSTLQEQIGW 1.2 1.7 71.6 1.2 0.9 HIV Gag KAFSPEVIPMF 1.2 1.7 56.1 1.2 0.9 Bw4-80T B*27:05 HIV Gag KRWIILGLNK 1.3 1.8 203.2 1.1 0.7 B*44:02 HIV RT EELRQHLLRW 1.3 2.1 1.1 1.2 1.0 Bw6 B*07:02 HIV Nef RPMTYKAAL 1.5 2.1 22.0 1.2 0.8 CMV pp65 TPRVTGGGAM 1.2 1.9 6.2 1.0 0.7 CMV pp65 RPHERNGFTV 1.4 1.9 2.4 1.0 0.7 HIV Gag GPGHKARVL 1.3 2.4 1.6 1.2 0.8 B*08:01 HIV Gag GEIYKRWII 1.2 1.1 1.1 1.1 1.2 EBV EBNA3A QAKWRLQTL 1.4 1.6 1.4 1.2 0.9 B*35:01 HIV Vif DAKLVITTY 0.9 1.1 1.6 1.0 0.8 Non-Bw4/Bw6 A*02:01 EBV BMLF1 GLCTLVAML 0.9 1.4 1.0 0.8 1.2 HIV Gag SLYNTVATL 1.1 1.6 0.9 1.0 1.2 A*03:01 HIV Env TVYYGVPVWK 1.0 1.5 4.8 1.0 0.7 HIV Gag RLRPGGKKK 1.4 1.6 0.9 0.9 0.8 Values in bold represent positive tetramer-binding indexes. aThe 721.221 cell line and stable derivatives thereof expressing five different KIR3DL alleles (KIR033-2, 033-7, 049-4, 049-6, and 049-7) from distinct clusters (Fig. 4) were stained with MHC-I tetramers and analyzed by flow cytometry. bRelative binding index, defined as the mean fluorescence intensity of tetramer binding to the KIR cell line divided by the mean fluorescence intensity of tetramer binding to the parental 721.221 cell line stained in parallel under identical conditions. The Journal of Immunology 9

HLA-B allotypes (46). In contrast, the macaque KIR3DL mole- likely that the D0 domain of KIR049-4 is an important contributor cule identified in this study recognizes Bw4+ MHC-I molecules to degenerate MHC-I recognition via stabilization of the KIR– with high avidity, some Bw6+ MHC-I molecules with lower MHC complex. The majority of the variable residues affect the D1 avidity, and some non-Bw4/Bw6 MHC-I molecules. In each case, and D2 domain interactions with the a1 helix, particularly around the nature of the peptide presented by MHC-I affects the strength position 76 and the Bw4/Bw6 motif region. In particular, the D1 of the interaction, with a more profound impact on the lower region at positions 165–167, which contacts both the peptide and avidity interactions. Despite the relatively low binding avidity for the MHC-I molecule around position 76 and 80, differs com- macaque Bw6+ MHC-I molecules detected by pMHC-I tetramer pletely between KIR049-4 and human KIR3DL1. These clustered staining, the interaction is sufficient to block NK cell activation in polymorphisms most likely contribute to the degenerate MHC-I functional assays. recognition profile of KIR049-4 and its associated peptide de- Although peptides have been shown to impact human KIR–HLA pendency. interactions (47–50), particularly in the case of KIR3DL2 mole- The ability of KIR049-4 to recognize Bw6+ MHC-I molecules is cules (11), this is often considered secondary to the role of the not unique. Rhesus macaque NK cells also express specific KIR3DLs presenting MHC-I H chain because few KIR residues interact encoded by Mamu-KIR3DL05 alleles that recognize Bw6+ MHC-I directly with the peptide (51–53). The degenerate MHC-I recog- molecules in a peptide-dependent manner (33, 34). Two such nition characteristics of KIR049-4 suggest that macaque KIR3D KIR3DL molecules, KIR3DLw03*004 and KIR3DL05*007, dis- molecules rely on somewhat different criteria than human play degenerate MHC-I recognition properties and are sensitive KIR3DLs for MHC-I recognition, with the bound peptide perhaps to amino acid changes at positions 77, 80, and 83 within the Bw4/ playing a more influential role. A consequence of this peptide Bw6 motif (33). Furthermore, polymorphic residues located in the Downloaded from selectivity is that some permissive MHC-I molecules are unable to D1 domain encoded by these KIR3DL05 alleles contribute to the interact with KIR049-4 in the presence of antagonist peptides, peptide-dependent recognition of Mamu-A1*002 tetramers, which potentially due to steric interference from peptide amino acid side carry the Bw6+ motif (34). Our analysis extends these findings chains. This is most likely exemplified in our dataset by the dif- beyond the rhesus macaque model. ferential binding observed for HLA-B*07:02 tetramers bearing The characteristics of KIR049-4 described in our study have

distinct peptides derived from HIV and CMV (Table II). Similarly, implications for other macaque species. Indeed, phylogenetic http://www.jimmunol.org/ it is possible that some MHC-I molecules identified as nonbinders analysis identified several KIR alleles from rhesus and cynomolgus in our assays, such as HLA-A*02:01, -B*08:01, and -B*35:01, macaques that are highly similar to KIR049-4 (Fig. 4). In par- could be recognized by KIR049-4 if different, agonist peptides ticular, the cynomolgus macaque KIR3DL allele KIR55 differs were loaded in the binding groove. Accordingly, the range of from KIR049-4 by only one synonymous and six nonsynonymous MHC-I molecules that can bind KIR049-4 is likely larger than mutations affecting the D0 domain, the D2 domain, and the cy- currently described. However, not all MHC-I molecules are able to toplasmic tail (28). As none of these mutations affect the predicted bind KIR049-4, as demonstrated in our functional assays with cell points of contact with MHC-I molecules, KIR55 most likely lines expressing Mane-A1*003, -A3*13, and -B*109 (Fig. 6). In exhibits a degenerate MHC-I recognition profile similar to that of these assays, the MHC-I molecules are loaded with numerous KIR049-4. Additional KIR molecules encoded by the same locus, by guest on September 29, 2021 variable peptides derived from the intracellular compartment and such as KIR07 in cynomolgus (28) and KIR3DLw03*004 in most likely incorporate both agonist and antagonist peptides. The rhesus (29) macaques, possess many additional polymorphic sites. lack of recognition by KIR049-4 indicates that motifs within the However, these mutations map mainly outside of the predicted MHC-I protein itself are necessary to stabilize the KIR–pMHC points of contact, and these KIR alleles most likely behave sim- complex. ilarly to KIR049-4. Therefore, the degenerate MHC-I recognition To gain a deeper understanding at the atomic level, we analyzed profile observed with KIR049-4 is most likely common to the the degenerate MHC-I recognition properties of KIR049-4 based three macaque species for which we have KIR data, due to con- on the structure of human KIR3DL1 complexed with a Bw4+ servation of the encoding locus. MHC-I molecule (HLA-B*57:01) (53). The binding site between Two additional macaque NK cell receptors are apparent from the these two proteins is defined by 17 aa in the MHC-I molecule that macaque pMHC-I tetramer reactivity patterns observed in our contact 21 aa in the KIR3DL1 molecule. Most of the contact study. These are likely KIR molecules that remain to be identified. residues (11 of 17) in HLA-B*57:01 are shared with macaque Based on the MHC-I–binding properties of KIR molecules, it MHC-I molecules that bind KIR049-4. The six residues that differ would be expected that pMHC-I tetramer staining could be used map to positions 19, 76, 80, 83, 142, and 151. The contact residues more generally to characterize NK cells ex vivo. However, only identified in human KIR3DL1 are equally divided between 11 aa one previous report has described such results in rhesus macaques (positions 9, 11, 13, 29, 34, 200, 201, 228, 230, 276, and 277) that (34). Multiple studies have reported staining of human NK cell are conserved in KIR049-4 and 10 aa (positions 138, 140, 165, clones or KIR transfectants with pMHC-I tetramers. Indeed, this 166, 167, 199, 227, 272, 279, and 282) that differ in KIR049-4. approach has allowed the characterization of specific ligands for The pattern of conservation/polymorphism between the KIR3DL1 KIR3DL1 and KIR3DL2 (11, 48, 50, 56). In contrast, no other and KIR049-4 molecules suggests that the interactions mediated report has shown pMHC-I tetramer staining of human NK cells by the D0 and D2 domains of KIR3D with the side of the MHC-I isolated directly from blood or tissue. The reasons for this lack molecule and the a2 helix, respectively, are conserved. Studies of staining remain to be reconciled with our knowledge of KIR using KIR3DL1 mutants have shown that the D0 domain helps to binding to MHC-I molecules. stabilize the interaction mediated by D1 and D2 over the MHC-I Finally, the ability to characterize NK cells using pMHC-I groove (54). Additional studies have shown that KIR2DL1 rec- tetramers provides us with a new approach that enables the ognition can be expanded beyond the confines of HLA-C mole- study of these cells during SIV/SHIV infection. This is an im- cules by fusion with the D0 domain of KIR3DL1 to include some portant advance because relatively few reagents are currently HLA-B and HLA-G molecules (55). As the contact points present available to monitor the diverse array of KIR molecules expressed in the D0 domain of KIR3DL1 are conserved in KIR049-4, it is in macaques. 10 DEGENERATE KIR–MHC INTERACTION IN MACAQUES

Acknowledgments complex class I (A*01) delay simian immunodeficiency virus SIVmac-induced immunodeficiency. J. Virol. 76: 292–302. We are grateful to Boris Skopets, Rahel Petros, and the staff of the National 24. Loffredo, J. T., J. Maxwell, Y. Qi, C. E. Glidden, G. J. Borchardt, T. Soma, Institutes of Health animal facility for taking care of the macaques in this A. T. Bean, D. R. Beal, N. A. Wilson, W. M. Rehrauer, et al. 2007. Mamu-B*08- study. We thank Jason Brenchley, Nichole Klatt, and Que Dang for expert positive macaques control simian immunodeficiency virus replication. J. Virol. advice on flow cytometry; Steven Kent for providing the construct used to 81: 8827–8832. 25. Fernandez, C. S., I. Stratov, R. De Rose, K. Walsh, C. J. Dale, M. Z. Smith, produce the KP9/Mane-A1*084 tetramer; Kristin Ladell for assistance with M. B. Agy, S. L. Hu, K. Krebs, D. I. Watkins, et al. 2005. Rapid viral escape at graphics; and David Margulies, Jack Bennink, Jon Yewdell, and Vanessa an immunodominant simian-human immunodeficiency virus cytotoxic T- Hirsch for comments on the manuscript. lymphocyte epitope exacts a dramatic fitness cost. J. Virol. 79: 5721–5731. 26. Choi, E. I., K. A. Reimann, and N. L. Letvin. 2008. In vivo natural killer cell depletion during primary simian immunodeficiency virus infection in rhesus Disclosures monkeys. J. Virol. 82: 6758–6761. The authors have no financial conflicts of interest. 27. Hershberger, K. L., R. Shyam, A. Miura, and N. L. Letvin. 2001. Diversity of the killer cell Ig-like receptors of rhesus monkeys. J. Immunol. 166: 4380–4390. 28. Bimber, B. N., A. J. Moreland, R. W. Wiseman, A. L. Hughes, and D. H. O’Connor. 2008. Complete characterization of killer Ig-like receptor (KIR) References haplotypes in Mauritian cynomolgus macaques: novel insights into nonhuman 1. Cerwenka, A., and L. L. Lanier. 2001. Natural killer cells, viruses and cancer. primate KIR gene content and organization. J. Immunol. 181: 6301–6308. Nat. Rev. Immunol. 1: 41–49. 29. Kruse, P. H., C. Rosner, and L. Walter. 2010. Characterization of rhesus macaque 2. Caligiuri, M. A. 2008. Human natural killer cells. Blood 112: 461–469. KIR genotypes and haplotypes. Immunogenetics 62: 281–293. 3. Lanier, L. L. 2008. Evolutionary struggles between NK cells and viruses. Nat. 30. Blokhuis, J. H., M. K. van der Wiel, G. G. Doxiadis, and R. E. Bontrop. 2010. Rev. Immunol. 8: 259–268. The mosaic of KIR haplotypes in rhesus macaques. Immunogenetics 62: 295– 4. Vivier, E., D. H. Raulet, A. Moretta, M. A. Caligiuri, L. Zitvogel, L. L. Lanier, 306. W. M. Yokoyama, and S. Ugolini. 2011. Innate or adaptive immunity? The 31. Bostik, P., J. Kobkitjaroen, W. Tang, F. Villinger, L. E. Pereira, D. M. Little, Downloaded from example of natural killer cells. Science 331: 44–49. S. T. Stephenson, M. Bouzyk, and A. A. Ansari. 2009. Decreased NK cell fre- 5. Lanier, L. L. 2008. Up on the tightrope: natural killer cell activation and inhi- quency and function is associated with increased risk of KIR3DL allele poly- bition. Nat. Immunol. 9: 495–502. morphism in simian immunodeficiency virus-infected rhesus macaques with 6. Raulet, D. H., and R. E. Vance. 2006. Self-tolerance of natural killer cells. Nat. high viral loads. J. Immunol. 182: 3638–3649. Rev. Immunol. 6: 520–531. 32. Chaichompoo, P., P. Bostik, S. Stephenson, S. Udompunturuk, J. Kobkitjaroen, 7. Vivier, E., E. Tomasello, M. Baratin, T. Walzer, and S. Ugolini. 2008. Functions K. Pattanapanyasat, and A. A. Ansari. 2010. Multiple KIR gene polymorphisms of natural killer cells. Nat. Immunol. 9: 503–510. are associated with plasma viral loads in SIV-infected rhesus macaques. Cell. 8. Trowsdale, J. 2001. Genetic and functional relationships between MHC and NK Immunol. 263: 176–187. receptor genes. Immunity 15: 363–374. 33. Rosner, C., P. H. Kruse, M. Hermes, N. Otto, and L. Walter. 2011. Rhesus http://www.jimmunol.org/ 9. Parham, P. 2005. MHC class I molecules and KIRs in human history, health and macaque inhibitory and activating KIR3D interact with Mamu-A-encoded survival. Nat. Rev. Immunol. 5: 201–214. ligands. J. Immunol. 186: 2156–2163. 10. Long, E. O. 2008. Negative signaling by inhibitory receptors: the NK cell par- 34. Colantonio, A. D., B. N. Bimber, W. J. Neidermyer, Jr., R. K. Reeves, G. Alter, adigm. Immunol. Rev. 224: 70–84. M. Altfeld, R. P. Johnson, M. Carrington, D. H. O’Connor, and D. T. Evans. 11. Hansasuta, P., T. Dong, H. Thananchai, M. Weekes, C. Willberg, H. Aldemir, 2011. KIR polymorphisms modulate peptide-dependent binding to an MHC S. Rowland-Jones, and V. M. Braud. 2004. Recognition of HLA-A3 and HLA- class I ligand with a Bw6 motif. PLoS Pathog. 7: e1001316. A11 by KIR3DL2 is peptide-specific. Eur. J. Immunol. 34: 1673–1679. 35. Older Aguilar, A. M., L. A. Guethlein, M. Hermes, L. Walter, and P. Parham. 12. Khakoo, S. I., C. L. Thio, M. P. Martin, C. R. Brooks, X. Gao, J. Astemborski, 2011. Rhesus macaque KIR bind human MHC class I with broad specificity and J. Cheng, J. J. Goedert, D. Vlahov, M. Hilgartner, et al. 2004. HLA and NK cell recognize HLA-C more effectively than HLA-A and HLA-B. Immunogenetics inhibitory receptor genes in resolving hepatitis C virus infection. Science 305: 63: 577–585. 872–874. 36. Committee on Care and Use of Laboratory Animals. 1985. Guide for the Care 13. Martin, M. P., X. Gao, J. H. Lee, G. W. Nelson, R. Detels, J. J. Goedert, and Use of Laboratory Animals. National Institutes of Health, Department of by guest on September 29, 2021 S. Buchbinder, K. Hoots, D. Vlahov, J. Trowsdale, et al. 2002. Epistatic inter- Health and Human Services Publication 85-23, Washington, DC. action between KIR3DS1 and HLA-B delays the progression to AIDS. Nat. 37. Kimura, M. 1980. A simple method for estimating evolutionary rates of base Genet. 31: 429–434. substitutions through comparative studies of nucleotide sequences. J. Mol. Evol. 14. Martin, M. P., Y. Qi, X. Gao, E. Yamada, J. N. Martin, F. Pereyra, S. Colombo, 16: 111–120. E. E. Brown, W. L. Shupert, J. Phair, et al. 2007. Innate partnership of HLA-B 38. Felsenstein, J. 1985. Confidence limits on phylogenies: an approach using the and KIR3DL1 subtypes against HIV-1. Nat. Genet. 39: 733–740. bootstrap. Evolution 39: 783–791. 15. Alter, G., S. Rihn, K. Walter, A. Nolting, M. Martin, E. S. Rosenberg, 39. Shimizu, Y., and R. DeMars. 1989. Production of human cells expressing indi- J. S. Miller, M. Carrington, and M. Altfeld. 2009. HLA class I subtype- vidual transferred HLA-A,-B,-C genes using an HLA-A,-B,-C null human cell dependent expansion of KIR3DS1+ and KIR3DL1+ NK cells during acute hu- line. J. Immunol. 142: 3320–3328. man immunodeficiency virus type 1 infection. J. Virol. 83: 6798–6805. 40. Price, D. A., J. M. Brenchley, L. E. Ruff, M. R. Betts, B. J. Hill, M. Roederer, 16. Letvin, N. L., M. D. Daniel, P. K. Sehgal, R. C. Desrosiers, R. D. Hunt, R. A. Koup, S. A. Migueles, E. Gostick, L. Wooldridge, et al. 2005. Avidity for L. M. Waldron, J. J. MacKey, D. K. Schmidt, L. V. Chalifoux, and N. W. King. antigen shapes clonal dominance in CD8+ T cell populations specific for per- 1985. Induction of AIDS-like disease in macaque monkeys with T-cell tropic sistent DNA viruses. J. Exp. Med. 202: 1349–1361. retrovirus STLV-III. Science 230: 71–73. 41. Pratt, B. F., D. H. O’Connor, B. A. Lafont, J. L. Mankowski, C. S. Fernandez, 17. Boyson, J. E., C. Shufflebotham, L. F. Cadavid, J. A. Urvater, L. A. Knapp, R. Triastuti, A. G. Brooks, S. J. Kent, and M. Z. Smith. 2006. MHC class I allele A. L. Hughes, and D. I. Watkins. 1996. The MHC class I genes of the rhesus frequencies in pigtail macaques of diverse origin. Immunogenetics 58: 995– monkey: different evolutionary histories of MHC class I and II genes in primates. 1001. J. Immunol. 156: 4656–4665. 42. Reeves, R. K., J. Gillis, F. E. Wong, Y. Yu, M. Connole, and R. P. Johnson. 2010. 18. Otting, N., C. M. Heijmans, R. C. Noort, N. G. de Groot, G. G. Doxiadis, CD162 natural killer cells: enrichment in mucosal and secondary lymphoid J. J. van Rood, D. I. Watkins, and R. E. Bontrop. 2005. Unparalleled complexity tissues and altered function during chronic SIV infection. Blood 115: 4439– of the MHC class I region in rhesus macaques. Proc. Natl. Acad. Sci. USA 102: 4446. 1626–1631. 43. Webster, R. L., and R. P. Johnson. 2005. Delineation of multiple subpopulations 19. Lafont, B. A., A. Buckler-White, R. Plishka, C. Buckler, and M. A. Martin. of natural killer cells in rhesus macaques. Immunology 115: 206–214. 2003. Characterization of pig-tailed macaque classical MHC class I genes: 44. Cooper, M. A., T. A. Fehniger, and M. A. Caligiuri. 2001. The biology of human implications for MHC evolution and antigen presentation in macaques. J. natural killer-cell subsets. Trends Immunol. 22: 633–640. Immunol. 171: 875–885. 45. Parham, P., L. Abi-Rached, L. Matevosyan, A. K. Moesta, P. J. Norman, 20. Smith, M. Z., C. J. Dale, R. De Rose, I. Stratov, C. S. Fernandez, A. G. Brooks, A. M. Older Aguilar, and L. A. Guethlein. 2010. Primate-specific regulation of J. Weinfurter, K. Krebs, C. Riek, D. I. Watkins, et al. 2005. Analysis of pigtail natural killer cells. J. Med. Primatol. 39: 194–212. macaque major histocompatibility complex class I molecules presenting immu- 46. Parham, P., P. J. Norman, L. Abi-Rached, and L. A. Guethlein. 2011. Variable nodominant simian immunodeficiency virus epitopes. J. Virol. 79: 684–695. NK cell receptors exemplified by human KIR3DL1/S1. J. Immunol. 187: 11–19. 21. Krebs, K. C., Z. Jin, R. Rudersdorf, A. L. Hughes, and D. H. O’Connor. 2005. 47. Rajagopalan, S., and E. O. Long. 1997. The direct binding of a p58 killer cell Unusually high frequency MHC class I alleles in Mauritian origin cynomolgus inhibitory receptor to human histocompatibility leukocyte antigen (HLA)-Cw4 macaques. J. Immunol. 175: 5230–5239. exhibits peptide selectivity. J. Exp. Med. 185: 1523–1528. 22. Evans, D. T., P. Jing, T. M. Allen, D. H. O’Connor, H. Horton, J. E. Venham, 48. Thananchai, H., G. Gillespie, M. P. Martin, A. Bashirova, N. Yawata, M. Yawata, M. Piekarczyk, J. Dzuris, M. Dykhuzen, J. Mitchen, et al. 2000. Definition of five P. Easterbrook, D. W. McVicar, K. Maenaka, P. Parham, et al. 2007. Cutting new simian immunodeficiency virus cytotoxic T-lymphocyte epitopes and their edge: allele-specific and peptide-dependent interactions between KIR3DL1 and restricting major histocompatibility complex class I molecules: evidence for an HLA-A and HLA-B. J. Immunol. 178: 33–37. influence on disease progression. J. Virol. 74: 7400–7410. 49. Fadda, L., G. Borhis, P. Ahmed, K. Cheent, S. V. Pageon, A. Cazaly, 23. Pal, R., D. Venzon, N. L. Letvin, S. Santra, D. C. Montefiori, N. R. Miller, S. Stathopoulos, D. Middleton, A. Mulder, F. H. Claas, et al. 2010. Peptide E. Tryniszewska, M. G. Lewis, T. C. VanCott, V. Hirsch, et al. 2002. ALVAC- antagonism as a mechanism for NK cell activation. Proc. Natl. Acad. Sci. USA SIV-gag-pol-env-based vaccination and macaque major histocompatibility 107: 10160–10165. The Journal of Immunology 11

50. Fadda, L., G. M. O’Connor, S. Kumar, A. Piechocka-Trocha, C. M. Gardiner, immunoglobulin-like receptor 3DL1-mediated recognition of human leukocyte M. Carrington, D. W. McVicar, and M. Altfeld. 2011. Common HIV-1 peptide antigen B. Nature 479: 401–405. variants mediate differential binding of KIR3DL1 to HLA-Bw4 molecules. J. 54. Khakoo, S. I., R. Geller, S. Shin, J. A. Jenkins, and P. Parham. 2002. The D0 Virol. 85: 5970–5974. domain of KIR3D acts as a major histocompatibility complex class I binding 51. Fan, Q. R., E. O. Long, and D. C. Wiley. 2001. Crystal structure of the human enhancer. J. Exp. Med. 196: 911–921. natural killer cell inhibitory receptor KIR2DL1-HLA-Cw4 complex. Nat. 55. Fu, L., B. Hazes, and D. N. Burshtyn. 2011. The first Ig domain of Immunol. 2: 452–460. KIR3DL1 contacts MHC class I at a secondary site. J. Immunol. 187: 1816– 52. Boyington, J. C., S. A. Motyka, P. Schuck, A. G. Brooks, and P. D. Sun. 2000. 1825. Crystal structure of an NK cell immunoglobulin-like receptor in complex with its 56. Stewart-Jones, G. B., K. di Gleria, S. Kollnberger, A. J. McMichael, E. Y. Jones, class I MHC ligand. Nature 405: 537–543. and P. Bowness. 2005. Crystal structures and KIR3DL1 recognition of three 53. Vivian, J. P., R. C. Duncan, R. Berry, G. M. O’Connor, H. H. Reid, T. Beddoe, immunodominant viral peptides complexed to HLA-B*2705. Eur. J. Immunol. S. Gras, P. M. Saunders, M. A. Olshina, J. M. Widjaja, et al. 2011. Killer cell 35: 341–351. Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021