Proc. Nati. Acad. Sci. USA Vol. 90, pp. 542-546, January 1993 A -related involved in the function of natural killer cells (natural killer tumor-recognition molecule/histone-related/arginine- and serine-rich domains) STEPHEN K. ANDERSON*t, STEVEN GALLINGER*, JOHN RODER*, JOYCE FREYt, HOWARD A. YOUNGt, AND JOHN R. ORTALDOt *Division of Molecular Immunology and Neurobiology, Mount Sinai Hospital Research Institute, 600 University Avenue, Toronto, Ontario, M5G1X5, Canada; and tLaboratory of Experimental Immunology, Biological Response Modifiers Program, Division of Cancer Treatment, National Cancer Institute-Frederick Cancer Research and Development Center, Frederick, MD 21702-1201 Communicated by Edward M. Scolnick, October 19, 1992

ABSTRACT Natural killer cells are non-major histocom- that these molecules also were linked to the triggering of NK patibility complex-restricted large granular lymphocytes that cells. This hypothesis was supported by cross-linking anti-ID can recognize and destroy tumor cells without prior stimula- to anti-2,4-dinitrophenol, which redirected NK cytotoxicity tion. A 150-kDa molecule on the surface of human natural against a NK-resistant target precoated with 2,4-dinitrophe- killer cells was identified as a component of a putative tumor- nol. On the basis of this evidence, the anti-ID antiserum was recognition complex. We report here the isolation of cDNAs proposed to identify a putative NK receptor (pNKR; ref. 12) coding for the 150-kDa tumor-recognition molecule from hu- complex on the surface of NK cells that is involved in the man and mouse cDNA libraries. The amino terminus of the recognition of tumor targets and is coupled to the cytolytic predicted protein contains a large hydrophobic region followed pathway. Therefore, we have named the 150-kDa member of by a domain that is highly homologous to cyclophilin/pepti- the complex described herein as a NK-tumor recognition dylprolyl ci-trans . The remainder of the protein is molecule (NK-TR). extremely hydrophilic and contains three homologous posi- The present study describes the isolation ofa cDNA coding tively charged clusters. There are also three regions that for a portion of the proposed NK-TR from a human NK-cell contain extensive arginine- and serine-rich repeats. Compari- Agtll expression library by screening with the anti-pNKR son of the human and mouse predicted amino acid sequences antiserum. Overlapping cDNAs containing the complete cod- revealed >80% homology. ing sequence were isolated from mouse and human cDNA libraries by DNA hybridization.§ The NK-TR gene is a Natural killer (NK) cells are defined as CD3-, CD16+, large single-copy nonrearranging gene specifically expressed in granular lymphocytes (LGL) capable of killing tumor targets NK cells. The predicted amino acid sequence of the NK-TR in a non-major histocompatibility complex-restricted fashion indicates that it represents a discrete class of protein con- (1). The NK cell is thought to constitute a first line of defense taining multiple domains with homology to cyclophilin and against microorganisms and tumor cells (2). Although the NK DNA/RNA-binding . phenomenon has been extensively studied, the molecule(s) responsible for the recognition of tumor cells by the NK cell has not been precisely defined. Numerous reports have MATERIALS AND METHODS implicated various NK cell and target-cell proteins in the Synthesis of NK-Cell cDNA Library and Plaque Screening. recognition events that occur (3-6), but there is still no Human NK cells were purified from human peripheral blood explanation ofthe process of specific recognition and lysis of (donated by the Canadian Red Cross). Peripheral blood was tumors by the NK cell. Studies have shown that the T-cell depleted of monocytes by adherence to plastic tissue culture receptor is not involved in target-cell recognition by NK cells plates, and B cells were removed by passage ofthe cells over and that CD3- LGL do not rearrange and express the a, 83, nylon wool columns. Two cycles of T-cell depletion with or 'y chains of the T-cell receptor (7-10). anti-CD3 and rabbit complement (Cedarlane Laboratories, To define the molecule(s) responsible for tumor-cell rec- Hornby, ON, Canada) were done. Activated NK cells were ognition by NK cells, a technique was used that involved the obtained by culture in the presence of at 10 production of anti-idiotypic (ID) antisera to a monoclonal units/ml (Cetus) for 2 days. Total cellular RNA was isolated antibody that recognized a variety of NK-cell-susceptible by using the method of Chirgwin et al. (13), and mRNA was tumor targets and blocked NK-cell-mediated lysis (11). The purified by using two rounds of selection on oligo(dT)- anti-ID antiserum was shown by immunofluorescence to bind cellulose. Double-stranded cDNA was synthesized by using specifically to CD3, CD16+ NK cells, and it blocked the a modification of the method of Gubler and Hoffman (14). binding of LGL to NK-susceptible targets, thereby prevent- After the addition of open-ended EcoRI linkers (Pharmacia), ing lysis (12). Two proteins of 80 and 150 kDa were immu- the cDNA was inserted into EcoRI-cut Agtll (Promega). The noprecipitated from 125I surface-labeled NK cells but not resulting library contained 6 x 106 independent clones. from T cells or monocytes. The ability ofNK cells to perform Antibody screening of the Agtll library was conducted by antibody-dependent cellular cytotoxicity was not affected by using the procedure of Young and Davis (15). Additional the anti-ID antiserum, indicating that the proteins recognized by the anti-ID antiserum were not associated with the CD16- Abbreviations: ID, idiotype; LGL, large granular lymphocytes; NK, mediated killing pathway. Prolonged pretreatment (18 hr) of natural killer; NK-TR, tumor-recognition molecule. effector cells with the anti-ID antiserum enhanced lytic tTo whom reprint requests should be sent at present address: activity and interferon y production by NK cells, suggesting Laboratory of Experimental Immunology, National Cancer Insti- tute-Frederick Cancer Research and Development Center, Building 560, Room 31-93, Frederick, MD 21702-1201. The publication costs of this article were defrayed in part by page charge §The sequences reported in this paper have been deposited in the payment. This article must therefore be hereby marked "advertisement" GenBank data base [accession nos. L04288 (human cDNA) and in accordance with 18 U.S.C. §1734 solely to indicate this fact. L04289 (mouse cDNA)]. 542 Downloaded by guest on October 1, 2021 Immunology: Anderson et al. Proc. Natd. Acad. Sci. USA 90 (1993) 543 clones were isolated by plaque screening of the human NK sites, no appreciable N-linked glycosylation of the NK-TR library, a human bone-marrow AgtlO cDNA library (Clon- was detected by treatment of protein lysates with N-glycosi- tech), a human thymus Agtl0 cDNA library (from T. Mak, dase (data not shown). The NK-TR contains several distinct University of Toronto), or mouse peripheral blood lympho- domains, as diagrammed in Fig. 2A. The first 58 amino acids cyte Agt10 cDNA library (Clontech) with DNA probes fol- of the NK-TR constitute a hydrophobic segment that resem- lowing the procedure of Benton and Davis (16). bles the multiple membrane-spanning region of the CD20 Radiolabeling of Proteins and Immunoprecipitation. Intact protein amino terminus (27). This result suggests that the CD3- LGL, T cells, and monocytes were surface-labeled NK-TR is attached to the membrane via its amino terminus, with 1251 by using lactoperoxidase that selectively labels only and most of the protein is expressed on the cell surface. The surface proteins (17). Proteins were immunoprecipitated proposed orientation of the molecule in the membrane is from cell lysates with anti-peptide antiserum or normal rabbit supported by the ability of the anti-ID and antipeptide serum, and protein A-Sepharose CL-4B. The complete cod- antibodies to bind to intact cells (12), indicating that the ing region of the human cDNA was cloned into pBluescript cyclophilin-related domain and the second positively charged KS(+) (Stratagene) and used as a template for a coupled in domain are extracellular. There is a 24-amino acid hydro- vitro transcription/translation system (Promega). Protein phobic region at position 641, and a 19-amino acid hydro- samples were analyzed by 7.5% SDS/PAGE (18). phobic stretch at position 1286. Approximately half of the DNA Sequencing. Direct sequencing of Agtll clones was amino acids in these regions are polar, negatively charged, or achieved through the use of forward and reverse sequencing residues, so they do not conform with typical trans- primers (New England Biolabs). The EcoRI-cleaved cDNA membrane segments. However, we cannot exclude the pos- inserts of individual A clones were purified and digested with sibility that they cross the membrane and result in the HaeIII, Alu I, or Mbo I and cloned into the appropriate expression of residues 665-1285 in the . restriction site of M13mp18 or M13mp19 (19). Single- Of particular interest is the presence of a domain that is stranded template DNA was purified and sequenced with homologous to cyclophilin/peptidylprolyl cis-trans isomer- Sequenase (United States Biochemical). Both strands of all ase (28, 29) immediately following the hydrophobic amino possible fragments were sequenced. terminus of the protein. Cyclophilin is the name given to a Nucleic Acid Hybridization. RNA and Southern blotting cytoplasmic protein that represents the major cyclosporin techniques were essentially as described (20). All radioactive A-binding protein isolated and protein sequenced from hu- DNA probes were synthesized by the random-primed man and bovine thymus (30). Protein sequencing of the method (BRL) using [a-32P]dATP as label (Amersham). peptidylprolyl cis-trans isomerase (28, 29) revealed that it corresponded to the cyclophilin gene, and cyclosporin A was shown to inhibit the activity of this enzyme. Fig. 2B RESULTS AND DISCUSSION shows a comparison of the NK-TR cyclophilin-related do- Molecular Cloning of the NK-TR. Affinity-purified anti-ID main with human thymus cyclophilin (23), the Drosophila antiserum was used to screen a human NK-cell Agtll cDNA ninaA protein (21), N. crassa cyclophilin (22), tomato cyclo- library. Of =500,000 plaques screened, six positive clones philin (25), and S. cerevisiae cyclophilin (24). The NK-TR were purified. These clones were used to produce Agtll sequence possesses approximately the same degree of ho- lysogens in Escherichia coli strain Y1089. The lysogens were mology (from 47% to 53% amino acid homology) to each of then used to overproduce /3-galactosidase fusion proteins by the previously described cyclophilin-related sequences. induction with isopropyl f3-D-thiogalactoside at 42°C. Fusion Greater than 70%o of the NK-TR amino acid residues in this proteins were purified on anti-f-galactosidase columns and region are homologous to one or more of the known cyclo- tested for their ability to block immunoprecipitation ofthe 80- philin proteins. The NK-TR is functionally more related to and 150-kDa proteins by the anti-ID serum. One clone (NK-2) the ninaA protein because it contains a membrane-bound produced a fusion protein that inhibited the immunoprecipi- cyclophilin-like domain, and the ninaA gene product is tation of the 80- and 150-kDa molecules and therefore was expressed as a transmembrane protein with the cyclophilin chosen forfurther study. A synthetic peptide corresponding to domain contained in the (31). a positively charged region of the predicted NK-2 fusion The remainder of the predicted NK-TR protein is ex- protein (peptide 104; residues 476-497 in Fig. 1) was injected tremely hydrophilic, containing >20% serine residues and into rabbits and produced the anti-p104 antiserum that reacted 37% charged residues. The positively charged amino acids with the 150-kDa molecule (12). The antipeptide antiserum are far more abundant, resulting in a net positive charge of possessed properties similar to the original anti-ID antiserum. 147 for the NK-TR protein. Three homologous positively These results indicate that clone NK-2 expresses an epi- charged clusters are present in the NK-TR (Fig. 2C). These tope contained in the 150-kDa molecule recognized by the positively charged regions of the molecule bear homology to anti-ID antiserum. The cDNA insert ofclone NK-2 contained several DNA-binding proteins, but the greatest homology 250 bp that possessed no significant homology to any known was seen with sperm histone (ref. 26; Fig. 2C). Synthetic sequence in the GenBank data base. The predicted amino peptide p104 from the second positively charged region was acid sequence fused to /-galactosidase in NK-2 is underlined able to induce an anti-NK-TR antiserum that reacted with the in Fig. 1. The NK-2 cDNA insert was used as a probe to surface of freshly isolated NK cells and possessed functional screen the Agtll NK cell cDNA library as well as a AgtlO properties similar to the original anti-ID antiserum (12). human bone-marrow library and a human thymus cDNA Because the anti-peptide antiserum could block killing of library in AgtlO. Overlapping cDNAs that spanned a total of tumor targets, this result suggests that the positively charged 4272 bp were isolated. domains are involved in triggering the NK lytic mechanism. Nudeotide and Deduced Amino Acid Sequence of the NK- The functional significance of the high density of positive TR. The nucleotide sequence and deduced amino acid se- charge in these regions is presently unclear. In addition, there quence of the human NK-TR are shown in Fig. 1. The human are three large clusters (-100 amino acids) containing nu- NK-TR cDNA possesses a continuous open reading frame of merous arginine-serine repeats. These regions have a general 1403 amino acids (predicted Mr 157,713) that begins with an motif of alternating positive and polar amino acids. Arginine- ATG initiation codon immediately preceded by an in-frame and serine-rich domains have been reported in a number of stop codon. The open reading frame is sufficient to code for proteins that regulate RNA splicing. The arginine-serine the 150-kDa molecule observed by immunoprecipitation. domains of the Drosophila suppressor of white apricot and Although there are several potential N-linked glycosylation transformer gene products target the proteins to a subcom- Downloaded by guest on October 1, 2021 544 Immunology: Anderson et al. Proc. Natl. Acad. Sci. USA 90 (1993)

(o0) 200

Aloawrq~rgrqPro A (117) 400 In C C CA A A C 0 A

(114) 600 (250) CI C Q A~~~~~~~~~~~~~~~~~~~~~~~Lo'tI CT Soo

Argie~ ii~la hol~ual*Srl~olVo~l Ar;uln r~lol~n~~shArp; aaeArgr~ rgjlootAr;VIrvlles~sl~l~ulTrysfrl~&svarhGLuolys~rg01. yerorisrlu (317) 1000 01C A AACA A 0 ?C CC 10- - CTUA CAA 10CARAART CC AICOA Vol As A A a A A TVol a a a a VolCI (11) 1200 T o AG0aTC a Ca a O CA C c 0 a0 CO CCOC aCgGO a CC a CO I2 Arg VlArrAr-gArg Thril, 0 o le n~o~oliy~otGy.y~o~oi~o Arg Arg~hoV<ryoLuoCA Tyr olor (450) 1400

--?l C A0 CCAACC 0 ? A CC C CC ACC Aoa 10 C CACTo GAA (516) Val r r; PO RoVo r l Ror GbAallAVal Alo~s r AValo 1597

(503) 1797 A OCA C aT C? CO A OCaOC C 0 a C G10 C A CAC1a

(649) 1997

C 0 0 C A 0 A C? -?T ACa aGA GA CT A0 Ca OAhaC C ProC~~~yRio Vol Not - t~~~~~~~o r A; Ar; Aoo Og l lyo At (710) Ao~~ohrAs;Riorhol 2197

CAC TTaC aTC A a C CA a---Oc? C CCAGM ACT cc 0000 In Oh 0 C ? a 0 ORh (761) 2391

---au P o Arg Pro pNis ValAgl&SrOy Co1. Nrg O or lnoArg Oly Glsb NrOylArgAs (647) ?%AAArAACCCACWIOCAAACCACICCTC10AAAOZCI0AAAOCACACOCA AQ2,CC ACCPr¶GAhQQ*. .1C.A . TIAAA¶OGGAG %CAhACACAOACAIGOCCAA0OAAA 2591 C2 COA C1 T ? TI CO 7AchO CA T OG AAAI OC --GA AC C CA C?0 --- A 0 COCAC--A10000 a Q0C 01Ar~l rlroa Nor Ph OlyrAwhp ArgrRo~~~ Pr lo-- Nor I* LYS Pyrfo-- Gb Gb ---rspho yop G rg Lm (014) ThOIRCAACCROTTOTAOCAMMAIM CCCCPrC10OCAOCC¶?CAACC70RAGARAC.. TAICCA10AAQChAG~CAG GIIAC1?A0GAAJ"hACCC OArGAAAAAA---OAAA4 rhAC0OAACAQT CACAAIIDAAGAMCAT1?10A AC1AG 2791 O OTC TC? T0T C QCO C T A 0 C 00CAC GAZO GA COO C1001 au Nor Ol AlNtnoop VAiPOarlo Loy Aloho Gly VialOo (961) 2991 A ?h AG TCCAGRACCA GAA00~QMT'RTi ?---TAG CCACAhT4aRGOOMGAAOC rCo 0C CO A010~mTC CRATCIIC ACCRCACROT --- C OCAFMCIZCAOC ITCMC0 0 0'1

TTC0c C ?0 00 0a T G ACA AC 000 AC aG Ac GA0T C aAG OTMGA0--- AG??Q CGAC hO r Alo o No orob Oboyoot o r; h (1047) Pro Alya Voll y Ob~Lys gloly---Go rrgNor 3191

GA? ---1 hC CAC C ? ?? A GC CA------a? O AC ? ChAC? W TO CaA? A 0 --- ro r Aro Nor VolA (1113) Vrol Cy His~ Valo Cynot ------sfkg eel 0.rgolArg Ar snLos 3366

hhhCACh00?1GPhhUhACAA TOCA C-2. ..M.0GQhcOA0?AACOAC GGAAA1AIGGAGAAT~CA~kM. -CTGIGA1ACCAA C WGA1CCTATWICAAMTIAAACQRAA (U79) 3595

(1245) 3765 0CC1CA10CTCCACOCGACCCCCA¶OCCTWICGATh 7 CaG¶cAcCA1000TARAGICACR?C1OCGAACA0IAhA06CT1ACOCC. .

(1311) AGCA~cOATI~hrUOAWCIIOCCAGAICO1?IC~roCAACA *. ...- ..~ C0GA .....TM1hmACcAAC7C0TIGCRACACOMWAACRATI ITACA110COCAAAOAA¶¶NAA10CA0AGrCCIOAAAAICAA1 3062

(1370) 4102

WI0AA0A1?AACCA0VoAAtCA. .R.. T6KATAI0P000Ir100GARh0---OA. - OAAG0CA1CACCAC ACA6KC1C10ACA¶TA00OCTMACCWIACTAAAOCA0 Ca amIo00AThC¶1C0IAACTAr~¶0AAACTyIArT1a1CAIVAAC aov~C10mrrrcrAccTr2ATCracraT2CM0¶10rIn (1403) 4302

0175

4375 1443) 4575 4775 497S 5175 5375 5575 5775 S975 U75 6375 6700

FIG. 1. NK-TR cDNA sequence. The sequence of the human NK-TR mRNA was determined from overlapping cDNAs. Nucleotides or amino acids of the mouse cDNA that differ from the human sequence are shown below the human sequence. Nucleotide numbers of human cDNA are shown at right; amino acid numbers are shown in parentheses. The symbol I above a vertical line indicates the position of an intron in the human NK-TR gene. The heavy underlining of amino acids 404-486 indicates the region expressed in Agt11 clone NK-2. The dashed underlinings indicate hydrophobic segments. Boxed regions represent clusters of positively charged amino acids.

partment ofthe nucleus thought to be the site ofRNA splicing homology to protein kinases or protein was (32, 33). detected. Perhaps the NK-TR mediates signal transduction The possible mechanism of signal transduction associated through other associated proteins. Also the cyclophilin/ with the NK-TR is presently unclear because no significant peptidylprolyl cis-trans isomerase region of the NK-TR is Downloaded by guest on October 1, 2021 Immunology: Anderson et al. Proc. Natl. Acad. Sci. USA 90 (1993) 545 A I 11 III

HOMOLOGY ___m__|_

NK-TR DR.F1( Ft*;R IC2

NinaA V TS R I Y M4 ; IF V K T1.F'. G F' G A: F:

NC-CYC L GP N N KP T S E I K A Q S N F r D,D V V T A RP NF Ei T HU CYC V N V F F i AV 3 G PIL ; V S F FIG. 2. of the NK-TR F R A Comparison yeast M Y FAD G CJ with other known proteins. (A) Schematic Tohaato A N K 'V F F DiL GG AA .RY ANIT TGA .V T A E IR A of NK-TR structure. A linear map of :P]KNK.!mK CY' F GG S T %';'K N iFS ! EG N G K NK-TR protein structure is shown. Black 1~~ rectangles indicate hydrophobic regions. NC-CYC F 'T The cyclophilin-related domain is labeled. Y-CYC K IGN , T ; S Shaded rectangles indicate the three ho- Ye as t RAET'L _Yi S V .A r: mologous positively charged clusters. Ar- to M KP! H G P A Tom K CFT ginine- and serine-rich domains are indi- '~~~~~~~~~~~~ cated by boxes containing horizontal N-TR K L H A F F I -H A lines. of NK-TR and known Ni n&A F .K A r. A V~gN!E :...... H.N R G:Pi A|';; '. V >NvWLa...... v k (B) Alignment NC -CYC .-A :: r V, R P G :.'M A N G P ;;.; W I.- .! cyclophilin genes. The cyclophilin-related -:_- riK A:j N 'F T V P T S 'w CYC E . I L K H T G GI S :^ k ' . ,F domain of NK-TR is with amino HU F.Es N - P L- M AL AlG P N'' N ,,7 F' I C A K T' W aligned acid sequences of ninaA Yeast P GLA iGMA A P N .:. I T Drosophila pro- Tomato .N V G A! P G T ,JA K TE - tein (NinaA; ref. 21), Neurospora crassa cyclophilin (NC-CYC; ref. 22), human NK -TR F < v ||F E F A s F Y A DIV'RF]IR thymus cyclophilin (HU-CYC; ref. 23), NinaA V K VL E M E.LN'T V E I.T E Saccharomyces cerevisiae cyclophilin DD M K NC-CYC E;VLf3 E S VIV K A F A S SizA I YS KK p v A (Yeast; ref. 24), and tomato cyclophilin HU-CYC F KV KNEE G M N I IJ FA R F N T S K I A L- L (Tomato; ref. 25). Only the residues ho- YeaOt !Di ! D I! R I \~L ,E: "V~~~~E~ NK mologous to the corresponding NK-TR Tomato i J END,D Wi 3 iqlveQ Ewt-At pi1 t residues are boxed. (C) Alignment of pos- itively charged regions with histone C;-Ig HP'R | Kv'J iR S:YF. He 2S9 5 (HIST). The three positively charged do- C I js'v[Jsi' Il ffE MV N H K!hr R K E 2K5K H j F. H K 5125 mains of NK-TR are aligned with sperm C F 1 - N |K KG1: -F. H GIS K E L 'K5E I' K K H Ki FP 1065 histone (26). Identical residues are boxed. N P ST . V[ MK |K r[A ' A. HI :iN TA A A K AA-V K A YT t- Amino acid numbers are shown at right.

involved in a signal-transduction process that occurs during quence analysis of these clones demonstrated that the mouse NK-cell-target cell interaction. and human genomic loci share an identical exon structure for Molecular Cloning of the Mouse Homologue of the NK-TR. all intron/exon junctions examined (data not shown). Screening of 500,000 clones from a mouse peripheral blood The NK-TR cDNA was used to investigate the expression lymphocyte AgtlO cDNA library (Clontech) with DNA probes of NK-TR mRNA in both human and mouse. Fig. 3B shows derived from the 5' and 3' regions ofthe human NK-TR cDNA the results of RNA blot analysis of human NK cells and T resulted in the isolation of 30 overlapping clones representing cells as well as several mouse tissues. A 7-kb mRNA was a 6.7-kb cDNA. The nucleotide sequence and proposed amino present in purified human NK cells but not in T cells. In acid sequence of the mouse NK-TR that differs from the humans, no significant expression was found in B cells human NK-TR sequence are shown in Fig. 1. The 5' initiation purified by anti-immunoglobulin panning, in monocytes iso- codon of the mouse gene was not contained within any of the clones sequenced; however, some sequence from the amino- A terminal hydrophobic region was determined. The predicted I B3 L \ I open reading frame of the mouse NK-TR cDNA is 1460 amino 1: acids with a predicted molecular mass of 160 kDa for the unmodified protein. The cyclophilin-related domain is 100%O LSW 9. conserved between the human and mouse NK-TR. The re- mainder of the two proteins is highly conserved, except for V5-~ several regions of elevated diversity. The mouse NK-TR possesses an additional 104 amino acids at the carboxyl 4 _I's '- tS terminus numerous that contains arginine-serine repeats. - Analysis of the NK-TR Gene and mRNA Expression. South- ern blotting analysis of human and mouse genomic DNAs was conducted to investigate the nature of the loci coding for the NK-TR mRNA. Fig. 3A demonstrates that a single-copy cellular gene homologous to the NK-TR cDNA present is in FIG. 3. Analysis of NK-TR gene. (A) Southern blot of human and both human mouse and genomic DNA. Analysis of various mouse genomic DNA. BamHI, EcoRI, HindIll, or Sac I-digested mouse tissues failed to reveal any gene rearrangements (data genomic DNA (10 Ag per lane) was probed with [32P]dATP-labeled not shown). In addition, a human genomic DNA clone was cDNA insert from clone NK-2 for the human blot (Left) or a 1.1-kb isolated from a human fetal liver library (34). The DNA mouse cDNA insert (PBL-21) for the mouse blot (Right). Lanes: B, sequence ofthe 3' region ofthe human NK-TRgenomic locus BamHI-digested human and mouse DNA; E, EcoRI digest; H, has been determined, and the positions of introns in the gene HindIl digest; S, Sac I digest. Molecular sizes in kbp are shown at left. (B) RNA are indicated in Fig. 1. The nucleotide sequence of the 3' blot of human and mouse NK-TR. Poly(A)+ RNA (5 coding region of the cDNA, including the observed ,ug per lane) was electrophoresed on a 0.7% agarose gel containing termina- formaldehyde. (A) Human mRNA. NK, purified NK cell mRNA; T, tion codon, has been confirmed by sequencing the genomic purified T-cell mRNA. (B) Mouse tissue mRNA. Sp, spleen mRNA; DNA. Several unspliced cDNAs were isolated from the Lu, lung mRNA; Li, liver mRNA. Positions of 18S and 28S ribosomal mouse peripheral blood lymphocyte cDNA library, and se- RNAs are shown at left. Downloaded by guest on October 1, 2021 546 Immunology: Anderson et al. Proc. Natl. Acad Sci. USA 90 (1993) A 13 acknowledge the work of Drs. D. Hoskin and J. Stankova in the -l purification of human NK cells used for cDNA library construction. This work was supported in part by a grant from the National Cancer Institute of Canada. S.K.A. was the recipient of a National Cancer Institute of Canada Fellowship while in the laboratory of J.R. The mouse cDNA was cloned and sequenced by S.K.A. at the National Research Council of Canada, Biotechnology Research Institute, Montreal.

7..4 -- 1. Trinchieri, G. (1989) Adv. Immunol. 47, 187-376. 2. Herberman, R. B. & Ortaldo, J. R. (1981) Science 214, 24-30. 3. Roder, J. C., Rosen, A., Fenyo, E. M. & Troy, F. A. (1979) Proc. Nati. Acad. Sci. USA 76, 1405-1409. FIG. 4. (A) Immunoprecipitation of NK-TR by anti-peptide an- 4. Fast, L. D., Beatty, P., Hansen, J. A. & Newman, W. (1983) J. LGLs were precipitated with normal Immunol. 131, 2404-2410. tiserum. 1251-membrane-labeled 5. Gilbert, C. W., Zaroukian, M. H. & Esselman, W. J. (1988) J. rabbit serum (lane 1) or anti-peptide 106 antiserum (lane 2), and Immunol. 140, 2821-2828. separated by SDS/PAGE. (B) In vitro translation of human cDNA. 6. Herberman, R. B., Reynolds, C. W. & Ortaldo, J. R. (1986) Annu. Coupled transcription/translation was done on vector alone (lane 1) Rev. Immunol. 4, 651-680. or vector containing complete human NK-TR cDNA (lane 2). 7. Ritz, J., Campen, T. J., Schmidt, R. E., Royer, H. D., Hercend, T., Positions of Mr standards (X10-3) are shown at left. Hussey, R. E. & Reinherz, E. L. (1985) Science 228, 1540-1543. 8. Lanier, L. L., Cwirla, S., Federspiel, N. & Phillips, J. H. (1986) J. lated by adherence to plastic, in platelets, or in thymocytes Exp. Med. 163, 109-214. not expression in mouse tissues correlated 9. Young, H. A., Ortaldo, J. R., Herberman, R. B. & Reynolds, (data shown). The C. W. (1986) J. Immunol. 136, 2701-2704. approximately with the levels ofNK cells present; the highest 10. Borst, J., van de Griend, R. J., van Oostveen, J. W., Ang, S. L., expression was found in spleen followed by liver and lung. Melief, C. J., Seidman, J. G. & Bolhuis, R. L. (1987) Nature Analysis of NK-TR Protin. Previous studies have indicated (London) 325, 683-688. that and other are intracellular 11. Bino, T., Frey, J. L. & Ortaldo, J. R. (1992) Cell. Immunol. 142, 28-39. proteins (35). The observation that the NK-TR cyclophilin 12. Frey, J. L., Bino, T., Kantor, R. R. S., Segal, D. M., Giardina, domain was part of a membrane-associated protein was inter- S. L., Roder, J., Anderson, S. & Ortaldo, J. R. (1991) J. Exp. Med. esting, particularly with regard to a recent report (36) of a 174, 1527-1536. cell-surface cyclosporin A-binding protein. Because anti- 13. Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J. & Rutter, peptide antibodies directed against the second positively W. J. (1979) Biochemistry 18, 5294-5299. 14. Gubler, U. & Hoffman, B. J. (1983) Gene 25, 263-269. charged domain reacted with the cell surface, the cyclophilin- 15. Young, R. A. & Davis, R. W. (1985) in Genetic Engineering, eds. related domain of the NK-TR should be expressed on the cell Setlow, J. K. & Hollaender, A. (Plenum, New York), Vol. 7, pp. surface. To confirm surface expression, several synthetic 29-41. peptides from the cyclophilin-related region were used to 16. Benton, W. D. & Davis, R. W. (1977) Science 196, 180-182. immunize rabbits. One antipeptide antiserum (p106; amino 17. Marchalonis, J. J., Cone, R. E. & Santer, V. (1971) Biochem. J. 124, 921-925. acids 95-115 in Fig. 1) was found weakly reactive with the 18. Laemmli, U. K. (1970) Nature (London) 227, 680-685. surface ofCD3-, CD56+ NK cells by flow cytometry (data not 19. Norrander, J., Kempe, T. & Messing, J. (1983) Gene 26, 101-106. shown) to a degree similar to that reported (12) with peptide 20. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular p104. To further examine the surface expression of this Cloning: A Laboratory Manual (Cold Spring Harbor Lab., Cold NK cells were surface labeled with Spring Harbor, NY). cyclophilin-related region, 21. Shieh, B.-H., Stamnes, M. A., Seavello, S., Harris, G. L. & Zuker, 125I and immunoprecipitated with anti-p106. The resultant C. A. (1989) Nature (London) 338, 67-70. 150-kDa band is shown in Fig. 4A. These results support the 22. Tropschug, M., Nicholson, D., Hartl, F.-A., Kohler, H., Pfanner, contention that this cyclophilin-related region is on a NK N., Wachter, E. & Neupert, W. (1988) J. Biol. Chem. 263, 14433- cell-surface-associated protein. The presence of a putative 14440. cyclosporin A-binding domain on a molecule expressed pri- 23. Haendler, B., Hofer-Warbinek, R. & Hofer, E. (1987) EMBO J. 6, 947-950. marily in LGL suggests that it may represent a physiologically 24. Haendler, B., Keller, R., Hiestand, P. C., Kocher, H. P., Weg- important target for cyclosporin A-mediated immunosuppres- mann, G. & Rao Movva, N. (1989) Gene 83, 39-46. sion. To confirm that the cDNA isolated could code for the 25. Gasser, C. H., Gunning, D. A., Budelier, K. A. & Brown, S. M. 150-kDa molecule, the complete human cDNA was cloned (1990) Proc. Natl. Acad. Sci. USA 87, 9519-9523. into pBluescript KS(+) and used in a coupled transcription/ 26. Kmiecik, D., Sellos, D., Belaiche, D. & Sautiere, P. (1985) Eur. J. translation system. The expected 150-kDa product was ob- Biochem. 150, 359-370. 27. Einfeld, D. A., Brown, J. P., Valentine, M. A., Clark, E. A. & served (Fig. 4B). Ledbetter, J. A. (1986) EMBO J. 7, 711-717. In conclusion, we have identified a NK-cell-specific gene 28. Takahashi, N., Hayano, T. & Suzuki, M. (1989) Nature (London) in mouse and human. The protein produced by this gene plays 337, 473-475. an important role in NK-cell cytotoxicity (12). The NK-TR 29. Fischer, G., Wittmann-Liebold, B., Lang, K., Kiethaber, T. & possesses a very unusual multi-domain structure composed Schmid, F. X. (1989) Nature (London) 337, 476-478. a amino terminus, a 30. Harding, M. W., Handschumacher, R. E. & Speicher, D. W. (1986) of hydrophobic cyclophilin-related do- J. Biol. Chem. 261, 8547-8555. main, three histone-like domains, and three arginine- and 31. Schneuwly, S., Shortridge, R. D., Larrivee, D. C., Ono, T., Ozaki, serine-rich domains. Although a direct role for the 150-kDa M. & Pak, W. L. (1989) Proc. Natl. Acad. Sci. USA 86,5390-5394. NK-TR protein in target-cell recognition has not yet been 32. Bingham, P. M., Chou, T.-B., Mims, I. & Zachar, Z. (1988) Trends demonstrated, the NK-TR gene will be a useful tool for Biochem. Sci. 4, 134-138. studying the ontology of NK cells as well as the molecular 33. Li, H. & Bingham, P. M. (1991) Cell 67, 335-342. 34. Lawn, R. M., Fritsch, E. F., Parker, R. C., Blake, G. & Maniatis, basis of tumor-cell recognition and lysis by the natural T. (1978) Cell 15, 1157-1174. immune system. 35. Schreiber, S. L. & Crabtree, G. R. (1992) Immunol. Today 13, 136-142. We thank B. Bere, C. Collins, A. Mason, D. Reynolds, and R. 36. Cacalano, N. A., Chen, B.-X., Cleveland, W. L. & Erlanger, B. F. Winkler-Pickett for their excellent technical assistance. We also (1992) Proc. Natd. Acad. Sci. USA 89, 4353-4357. Downloaded by guest on October 1, 2021