Jeffrey P. Houchins

A multigene family on human chromosome 12 encodes natural killer-cell lectins

We previously isolated a series of cDNA clones designated NKG2-A, B, C, and D from a human natural killer (NK) cell library. These transcripts encode a family of type II integral membrane proteins having an extracellular Ca2+-dependent lectin domain. The predicted peptides share structural similarities and amino acid sequence similarity with known receptor molecules. In this report, the genomic organization and mRNA expression of each of the genes were studied by using transcript-specific probes. Southern blot experiments reveal that the probes cross-hybridize with a maximum of five genes at high stringency. By probing a Southern blot prepared from a series of hamster/human hybrid somatic cell lines, we demonstrated that all of the hybridizing fragments occur on human chromosome 12. No gene rearrangement and little restriction fragment length polymorphism (RFLP) was observed with these probes. mRNA expression of the NKG2 genes occured in NK cells and some T cells but not in other hematopoietic cell types or in other tissues tested. Each of the transcripts occurred in all three of the NK cell lines tested: however, the genes were differentially regulated in T cells. NKG2-D was expressed in nine of fourteen T-cell clones or lines in the panel, whereas NKG2-A/B was expressed in three and NKG2-C was expressed in only one. Expression of each of the transcripts was upregulated following T-cell growth factor (TCGF)-induced activation of a cloned NK cell. The limited distribution of these proteins and their sequence similarity with known receptor molecules suggest that they may function as receptors of human NK cells.

Natural killer lectin-like receptors have divergent carboxyl-termini, distinct from C-type lectins

The mode of recognition of target cells by natural killer (NK) lymphocytes is unknown. Three related groups of NK cell-specific molecules may have receptor function (Ly49, NKR-P1 in rodents and NKG2 in humans). These NK lectin-like receptors (NKLLR) are type II integral membrane proteins with extraeellular domains homologous to Ca2+-dependent animal lectin carbohydrate recognition domains (C-type CRDs; Chambers et al. 1993). The NKG2 family consists of eDNA clones NKG2-A and -C and the more distantly related NKG2-D. Other NKG2 genes may exist (Yabe et aI. 1993). In order to isolate additional gene products related to the NKG2 family, an NK cell eDNA library (Yabe et al. 1993) was screened with the NKG2-C-specific probe at low stringency and with a nearly fulMength NKG2-A probe. Four cDNA clones with a novel sequence designated NKG2-E were isolated and sequenced. Other clones isolated were identified as either NKG2-A, C [including an NKG2-C clone lacking the transmembrane domain, NKG2-C 2 2 0 2 8 4 base pairs (bp)] or E. NKG2-E encodes a protein 261 amino acid long, sufficiently divergent from NKG2-C to be a product of a different gene. NKG2-E is 95% homologous to NKG2-C over the first 191 amino acid residues, then both sequences diverage (76% homologous over the 29 subsequent residues). An Alu repeat encodes the 15 carboxy-terminal residues of NKG2-E, including an arginine instead of a terminal cysteine present in NKLLRs and C-type CRDs. Identical NKG2-Esegments with anAlu repeat ( 600 770 bp) were sequenced from a different cDNA library by TA cloning of nested polymerase chain reactionderived products. Alu repeats have been reported in coding regions of a few other proteins (Claverie 1992). In a northern blot panel, the NKG2-C specific probe (98% bp homology with NKG2-E) hybridized to mRNAs equivalent in size to NKG2-C or E ordy in NK ceils and not other ceil lines (Yabe et al. 1993). C-type lectins have been classified into groups based on CRD sequence similarity. This classification coincides with the type of protein of which the CRDs are a part (i. e., group I: proteoglycans, group n : type I/receptors, group llI: collagen domains, group IV: selectins). Group I and II CRDs are encoded by three exons, whereas a single exon encodes group 111 and IV CRDs. The exon boundaries within these groups are highly conserved (Bezonska et aI. 1991). NKLLRs share several characteristics with group II C-type lectins. Both groups are type 11 integral membrane proteins with a smiliar gene structure. The CRD of Ly49, NKR-P1 and group ]1 CRDs are encoded by three exons with the same boundary positions (Wang et al. 1991; Giorda et aI. 1992; Bezouska et aI. 1991). Despite the apparent similarity between NKLLRs and group lI C-type iectins, a previous computergenerated comparison of lectin domains has demonstrated that NKLLRs constitute a separate group from C-type Iectins (Chambers et al. 1993). The reason for this discrepancy is revealed in a sequence comparison of NKG2-E to related proteins (Fig. 1). The lectin domain of C-type lectins (groups I I V ) and NKLLRs was first divided into two portions. The beginning of the NKG2-E/C divergence (which corresponds to the last coding exon of Ly49, NKR-P1 and group 1I C-type) was used as the dividing point. Only the COOH-terminal segments of NKLLR are clearly distinct from the C-type Iectins. The NH2 segments of NKLLR CRDs are related to groups I and II C-type CRD segments, as expected. Of note, highly conserved calcium and carbohydrate binding residues found in C-type CRDs (Drickamer 1992) are absent from the COOH segment of NKLLRs. In summary, NKLLRs are closely related to group II C-type lectins, although a significant evolutionary divergence seems to have occurred to the portion corresponding to the carboxy terminal coding exon. Should the carboxy terminal segment of NKLLRs serve as the Iigand portion of the molecule, as is the case for C-type lectin (Weis et al. 1992), one can predict different and varied NKLLR binding specificities from C-Type CRDs.

Candidate natural killer cell receptors.

Among the high points of immunological discovery has been the identification of antigen-recognizing receptors on B and T cells. Of the lymphocyte populations, only the NK cell receptor remains unknown. Consequently, any newly-recognized, cell-surface molecules expressed selectively on NK cells, especially ones that can transmit a signal to the cell upon appropriate ligand interaction, are possible candidates. This article describes such candidates.

Characterization of a novel gene (NKG7) on human chromosome 19 that is expressed in natural killer cells and T cells

NKG7 is a cDNA clone generated from a human NK-cell clone. The DNA and predicted aa sequence of NKG7 is not homologous with any previously reported genes or peptides. NKG7 mRNA is expressed in activated T cells and in A-LAK cells isolated from the peripheral blood of normal individuals, and in normal human kidney, liver, lung and pancreas. Furthermore, NKG7 mRNA is expressed at high levels in TCR gamma delta-expressing CTL clones, and in some TCR alpha beta-expressing CTL clones (both CD4+ and CD8+), but is not expressed in other TCR alpha beta-expressing CTL clones and in cell lines representing B cells, monocytes, and myeloid cells. NKG7 mRNA is not expressed in normal human brain, heart, or skeletal muscle. Southern hybridization of NKG7 suggests that NKG7 is a single-copy gene localized to chromosome 19. A hydropathicity profile of the predicted 148 aa polypeptide indicates that NKG7 is a type-I integral membrane protein with a 38-aa extracellular domain and a 61-aa cytoplasmic domain. These results indicate that the NKG7 gene encodes a novel cell surface protein expressed in several cell types, including NK cells and T cells.