Identification of upstream cis-acting regulatory elements controlling lineage-specific expression of the mouse NK cell activation receptor, NKR-P1C.

Mouse NKR-P1C (NK1.1) is a homodimeric type II transmembrane protein expressed on natural killer (NK) cells, NKT cells, and on CD117+ progenitor thymocytes capable of giving rise to cells of the T and NK lineages. Although its physiological ligands remain unknown, NKR-P1C engagement with a monoclonal antibody (mAb) leads to interferon-gamma (IFN-gamma) production and the directed release of cytotoxic granules from NK cells. We have cloned and sequenced a approximately 10-kb genomic fragment corresponding to the 5'-flanking region of the C57Bl/6 mouse NKR-P1C gene. A transcriptional initiation site has been mapped in NK cells and an NK1.1+ T cell line by primer extension and rapid amplification of 5'-cDNA ends (5'-RACE) techniques. Although the 5'-flanking region of NKR-P1C is TATA-less, we have identified an initiator region and a downstream promoter element, which together constitute the principal minimal functional promoter. Computational analysis of the 10-kb 5'-flanking region revealed potential regulatory factor binding sites. DNaseI hypersensitivity assays identified a single hypersensitive site (HS) about a 9-kb upstream of the transcriptional initiation site. This site, termed HS1, was able to act as a transcriptional enhancer element in an NK cell line, while minimally affecting transcription in non-NK cell lines. Moreover, the HS1 element was shown to function as a promoter, with a transcript detected only in fetal NK1.1+ cells. An additional promoter and two non-coding exons were also characterized. These results identify the minimal upstream cis-acting elements, and point to a complex regulatory mechanism involved in the lineage-specific control of NKR-P1C expression in NK lymphocytes.

The classic NK1.1 antigen (NKR-P1C) is one member of a group of disulfide-linked homodimeric type II transmembrane C-type lectin-like receptors encoded by the NKR-P1 gene family (1,2). The NKR-P1 gene cluster is found within a single genetic region designated the natural killer gene complex (NKC), 1 which is located on chromosome 6 in mice, chromosome 4 in rats, and chromosome 12 in humans (2,3). The NKC region is strikingly conserved among species, even though not all NKCencoded molecules have been identified in all species (4).
The NKC can be divided into several families of genes: NKRP1, Ly49, CD69, NKG2, and CD94. The prototype of the NKR-P1 family, NKR-P1A, was first cloned in rats using a mAb screened for its ability to redirect NK cell lysis of Fc receptor (FcR) ϩ targets (5,6). In mice, the NKR-P1 gene family consists of at least five distinct genes: NKR-P1A, NKR-P1B, NKR-P1C, NKR-P1D, and NKR-P1F (while the NKR-P1E locus represents a pseudogene) (1,(7)(8)(9). The initial cloning of the mouse NKR-P1 genes revealed that they consist of six exons separated by five introns (7). The first exon of the NKR-P1C gene encodes the N-terminal region of the protein, corresponding to the cytoplasmic domain. The second exon encodes the transmembrane region, exon 3 encodes a variable stalk region, and exons 4 -6 contain the putative carbohydrate recognition domain (CRD) (7). Although NKR-P1C had been originally cloned as the NK1.1 antigen (2), we and others have recently demonstrated that the mouse "NK1.1 antigen" actually represents the products of at least two distinct genes, NKR-P1B and NKR-P1C, differentially expressed across mouse strains (10,11).
NKR-P1C is expressed by all functional NK cells (7) and by a subset of mature T cells termed NKT cells (12). In selected strains of mice, virtually all NK activity is contained within the NKR-P1C ϩ subset (13), and in vivo administration of anti-NK1.1 mAb eliminates NK activity but has no effect on cellmediated or humoral immunity (14). Moreover, engagement of NKR-P1C with a mAb mediates NK cell redirected lysis of FcR-expressing targets (15), indicating that NKR-P1C is capable of activating and directing the cytolytic process. To date, however, physiological NKR-P1 ligands on target cells have not yet been identified although recombinant NKR-P1 proteins were suggested to bind synthetic carbohydrates (16 -20). Thus, the NKR-P1 proteins remain orphan receptors. In addition, our recent work has shown that a population of CD117 ϩ progenitor thymocytes expressing NK1.1 is capable of giving rise to the cells of T and NK cell lineages (21)(22)(23). Importantly, a single CD117 ϩ NK1.1 ϩ cell was shown to give rise to both T and NK cells, establishing that NK1.1 expression can serve to identify a common progenitor for both lineages (24).
In light of the developmentally controlled appearance of the NKR-P1 proteins during lymphocyte lineage commitment (21-* This work was supported by funds from the Canadian Institute for Health Research (CIHR). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§  24) and their lineage-restricted expression by mature NK cells, we wondered how their expression is regulated. To address this, we investigated the molecular cis-acting elements responsible for their transcriptional regulation. This report describes the cloning and detailed analysis of the upstream region of the mouse NKR-P1C gene. We have identified the location of the major NKR-P1C transcriptional initiation site; two new noncoding exons in the 5Ј-region of the gene (exons 1.1 and 1.2); three promoters; and putative transcription factor binding sites in the upstream flanking region. Furthermore, we describe the detection and analysis of a DNaseI hypersensitive site (HS1), located 9-kb upstream of the major transcriptional initiation site, which exhibits promoter and enhancer-like activity. Taken together, our present findings identify the minimal functional promoter and a unique enhancer/promoter element, HS1, which appear to confer NK lineage-specific expression of the NKR-P1C gene.

EXPERIMENTAL PROCEDURES
Screening of a Mouse Genomic Library-A mouse -phage genomic library derived from C57Bl/6xC3H F1 mice (generously provided by Dr. Gillian Wu, York University, Toronto, ON, Canada) was screened using an NKR-P1C gene-specific probe corresponding to the B6 NKR-P1C cDNA coding sequence. Three clones were identified spanning various regions within the mouse NKR-P1C locus and were subcloned. Clone P13 was characterized in this study.
Primer Extension Analysis-mRNAs used in primer extension assays were generated from two different cell lines: CTLL-2, and MNK-1 9.76kb/GFP . For the CTLL-2 mRNA, primer extension assays were performed using two 18-bp NKR-P1C antisense primers: E1A (5Ј-CA-GGGTTGGATAAGACAG-3Ј) complementary to bases Ϫ403 to Ϫ420; and E1B (5Ј-GTGTCCTGGGTGGCTTTA-3Ј) complementary to bases ϩ28 to ϩ45 (both primer locations are relative to the initiator methionine codon ATG in exon 1, Fig. 2, underlined, bold). For MNK-1 9.76kb/GFP mRNA, the same E1A primer was used in addition to a GFP primer (5Ј-CTCCTCGCCCTTGCTCACCA-3Ј) complementary to bases ϩ2 to ϩ20 of GFP coding sequence. Primers were labeled using [␥-32 P]dATP (specific activity Ն6000 Ci/mmol) and T4 polynucleotide kinase (New England Biolabs, Beverly, MA) to a specific activity of 1 ϫ 10 9 cpm/g. 1 ϫ 10 5 cpm of each primer (0.1 ng) was mixed with 1.5 g of oligo(dT)-selected RNA from CTLL-2 cells or 2 g of MNK-1 9.76kb/GFP mRNA. 10 g of total yeast RNA or 2 g of MNK-1 mRNA were used as negative controls. The RNA samples were reverse-transcribed at 42°C for 1 h using AMV reverse transcriptase (Invitrogen). Radioactive primer extension products were resolved in a sequencing gel. The same primers were used for sequencing reactions using Sequenase Version 2 Sequencing Kit (Amersham Biosciences, Piscataway, NJ). The template for sequencing reactions was a genomic fragment corresponding to the upstream region of the NKR-P1C gene.
Luciferase, ␤-Galactosidase, and Protein Assays-Cells transfected with the luciferase and ␤-galactosidase reporter plasmids were assayed using the Dual-Light reporter system (Tropix Inc. Applied Biosystems, Bedford, MA), as previously described (25). Results represent the average luciferase activity indexed for ␤-galactosidase activity and protein concentration per sample.
DNaseI Hypersensitivity Assay-MNK-1 9.76kb/GFP cells (10 8 ) were pelleted and resuspended in reticulocyte standard buffer (RSB; 10 mM Tris HCl (pH 7.5), 10 mM NaCl, 3 mM MgCl 2 ). Cells were lysed in 0.5% (v/v) Nonidet P-40, and nuclei were collected and resuspended in RSB buffer. Digestions were carried out at 37°C with 0.25-8.0 units of DNaseI (Invitrogen) for 3 min then terminated with an equal volume of DNaseI stop buffer (0.6 M NaCl, 1% (w/v) SDS, 20 mM Tris-HCl (pH 8.0), 10 mM EDTA). Samples were treated with proteinase K overnight, and DNA was recovered by phenol/chloroform extraction and ethanol precipitation, then resuspended in 100 l of dH 2 O. 10 g of each sample was digested with NotI followed by Southern blotting and hybridization using a 32 P-labeled GFP probe overnight at 65°C. Blots were exposed to PhosphorImager screens (Kodak, Rochester, NY) overnight, and scanned using a PhosphorImager (BioRad).
Computational Analysis of the 5Ј-Flanking Region Sequence-Putative regulatory elements in the 5Ј-flanking regions of the mouse NKR-P1C gene were identified by searching the transcription factor binding site data base (TRANSFAC 4.0 and MatInspector V2.2; transfac.gbf.de/ TRANSFAC/). Parameters used were: core similarity, 1.0; matrix similarity, 0.85. The numerical position of each cis-element corresponds to the 5Ј-end of each motif relative to the major transcriptional initiation site (bold arrow, Fig. 2).

Isolation of Genomic Clones
Containing the 5Ј-Flanking Region of Mouse NKR-P1C-A -phage mouse genomic library was screened by Southern hybridization using a PCR-generated genomic probe complementary to the NKR-P1C 5Ј-coding region. Three positive clones were obtained. One clone, P13, had a 13-kb genomic insert (GenBank TM accession number AY316227), which was further characterized with restriction enzymes. To exclude recombination artifacts, restriction enzyme digestion patterns of mouse genomic DNA from C57Bl/6 splenocytes were compared with those of the P13 phage clone and found to be identical (data not shown). To subclone the 5Ј-flanking region, subgenomic plasmid libraries were gener-ated and screened with a probe corresponding to the NKR-P1C coding region.
Identification of the Major NKR-P1C Transcriptional Initiation Site-To define the transcriptional initiation site of the NKR-P1C gene, primer extension was performed on RNA from CTLL-2 cells (an IL-2-dependent mouse T cell line that expresses NKR-P1C) (26) and MNK-1 9.76kb/GFP (a mouse pre-NK cell line, MNK-1 (27), expressing GFP under the control of the 9.76 kb NKR-P1C 5Ј-upstream region; see Fig. 6A). Extended products were separated on a denaturing polyacrylamide/urea gel alongside a sequence of the genomic clone to measure the size of extended products. Results depicted in Fig. 1A shows that transcription in MNK-1 9.76kb/GFP cells is initiated at a single site 156 bp upstream of the starting ATG codon (Fig. 2  underlined, bold). This site is also observed in CTLL-2 cells (Fig. 1B, upper strong band). In CTLL-2 cells, transcription is also initiated, although at a much lower level, from another site ( Fig. 1B; lower band). The specificity of the primer was confirmed using untransfected MNK-1 cells and yeast RNA as a negative controls (Fig. 1A, right lane and Fig. 1B, left lane).
Similar results were obtained utilizing the 5Ј-RACE technique (Fig. 1C). Here, RNA sources included CTLL-2 cells and splenocytes (NK cell-enriched) obtained from RAG-2-deficient mice (Fig. 1C). Although both major and minor initiation sites were detected by this technique, a set of shorter transcripts was also detected. These shorter fragments are likely due to FIG. 1. Mapping of the mouse NKR-P1C transcriptional initiation site. A, primer extenstion was performed using MNK-1 9.76kb/GFP mRNA. An oligonucleotide corresponding to the 5Ј-end of the GFP gene was hybridized to MNK-1-9.76/ GFP mRNA and extended with reverse transcriptase. Extension products were size fractioned on a denaturing polyacrylamide gel alongside a Sanger dideoxynucleotide sequencing ladder primed as above. mRNA from untransfected MNK-1 cells was used as a negative control. The major transcription initiation site is indicated with an arrow placed next to the corresponding genomic sequence. B, primer extension was performed using CTLL-2 mRNA. An oligonucleotide corresponding to the 5Ј end of the mouse NKR-P1C cDNA (shown in Fig. 2) was hybridized to CTLL-2 mRNA, and extended as described above. Total yeast RNA was used as a control for the specificity of the reaction. The major and minor transcription initiation sites are indicated with the arrows placed next to the corresponding genomic sequence. C, mapping of the transcription initiation sites by 5Ј-RACE. The nucleotide sequence for the genomic portion of mouse 5Ј-flanking region and part of exon 1 (starting ATG is underlined) is shown. The symbols * and # denote the transcription initiation sites of unique clones obtained by 5Ј-RACE and characterized by DNA sequence analysis from CTLL-2 cell line, and RAG-2-deficient splenocytes (NK cell-enriched), respectively. The major transcription initiation site as defined by primer extension analysis (A and B, bold) is shown and numbered ϩ1. inefficient reverse transcriptase function during the first strand synthesis.
Thus, two complementary techniques independently confirmed that the major NKR-P1C transcriptional initiation site is located at position Ϫ156 (relative to the translational initiator methionine ATG codon).
Detection of NKR-P1C mRNAs-containing Novel 5Ј-Exons-Both 5Ј-RACE (Fig. 1C) and primer extension (Fig. 3A) analyses revealed the presence of a novel exon, designated exon 1.1 (E1.1), located upstream of the originally described exon 1 (E1) (7). Fig. 2 shows the transcriptional initiation site for this novel exon (small thin right arrow) and the location of the novel splice donor/acceptor sites (Ngt...agN) corresponding to exons 1.1/2.1. To confirm the physiological existence of exon 1.1, we performed RT-PCR analysis with specific primers (F1B and F1C) in Fig. 2. The results demonstrate that exon 1.1 is transcribed significantly, albeit to a lesser extent than exon 1 (Fig.  3B). Interestingly, transcripts containing exon 1.1 could be detected among adult tissues (CTLL-2 cells, splenic NK cells), but not during fetal ontogeny (day 15 fetal liver and thymus are shown) (Fig. 3B). Moreover, primer extension analysis revealed that NKR-P1C transcripts initiated from the upstream site are not very abundant (Fig. 3A). Although exon 1.1 contains one in-frame ATG start codon (Fig. 2, italicized), no easily discernable structural motifs are present within this sequence.
To investigate whether a transcript containing exon 1.1 could encode for a protein, we generated plasmids containing either transcript 2 cDNA or the native NKR-P1C cDNA (transcript 1) (Fig. 4A). Transfection of these vectors into Jurkat cells (human T cell line) yielded detectable surface expression (as determined by NK1.1 staining) only with the original NKR-P1C exon 1-containing cDNA (Fig. 4B). Notably, even the addition of a proteasome inhibitor (MG-132, Calbiochem, San Diego, CA) could not allow for the detection of a product from the novel exon 1.1-containing cDNA, either on the surface or intracellularly (data not shown). The lack of detectable surface or intracellular staining suggests that the newly identified exon 1.1 is non-coding. In this regard, it is interesting to note that although exon 1.1 possesses an ATG (Fig. 2, italicized) that is in-frame with the rest of the gene, several other start sites, which are followed by multiple in-frame stop codons, precede the in-frame ATG in exon 1.1 (Fig. 4A). The presence of multiple stop codons downstream of the first ATG in exon 1.1 is likely to contribute to the failure to detect NK1.1 expression encoded by transcripts containing exon 1.1.
Promoter/Enhancer Activity of the NKR-P1C 5Ј-Flanking Region-The functional promoter activity of the NKR-P1C up-stream region was assessed by transient expression of luciferase reporter constructs ( Fig 5A). Initially, a genomic fragment containing a full 9.76 kb of the NKR-P1C upstream region was cloned into the pGL-2 vector (pGL-2/9.76kb). To assess lineagespecific expression, this construct was introduced into MNK-1 cells, and luciferase activity in cell extracts was measured. As a non-NK lineage control, the construct was also transfected into SL-12 cells (an NK1.1 Ϫ mouse pre-T cell line) (29). To determine the minimal promoter function, a shorter 600 bp NKR-P1C upstream flanking sequence was also inserted into the pGL-2 vector (pGL-2/600bp). In addition, a pGL-2/SV40 construct (luciferase driven by the SV40 promoter/enhancer) was used as a positive control for luciferase activity, while a CMV-␤-galactosidase vector (CMV promoter/enhancer driving ␤-galactosidase) was used to normalize transfection efficiency. Fig. 5 shows the results of the transfections, normalized for transfection efficiency and fold induction of luciferase activity relative to the promoterless/enhancerless pGL-2/Basic vector. As expected, the control pGL-2/SV40 vector produced a large induction in luciferase activity in both cell lines (Fig. 5A). Importantly, the pGL-2/600bp NKR-P1C promoter region alone also yielded a significant increase in luciferase activity in both cell lines (Fig. 5A). However, when the activity of the large genomic fragment containing the 9.76-kb NKR-P1C upstream region was assessed, a ϳ30-fold higher luciferase activity was observed in MNK-1 cells relative to that of SL-12 cells (Fig. 5A). Strikingly, the 9.76-kb insert drove expression of luciferase in MNK-1 cells almost equal to that of the SV40 construct. In contrast, the activity of the 9.76-kb upstream region was actually reduced 3-fold in SL-12 cells relative to the activity of the shorter 600-bp fragment. Collectively, these results suggest that the 9.76-kb upstream region contains elements necessary and sufficient for lineage-specific expression of the NKR-P1C gene, and that the 600-bp element has core promoter activity.
To further refine the elements in the upstream region responsible for lineage-specific expression, we explored the minimal DNA fragment size of the 5Ј-region that is necessary for full tissue-specific transcriptional regulation of the NKR-P1C gene. We achieved this by performing deletion analysis of the cloned 9.76-kb 5Ј-region, and subcloning these fragments into the pGL-2/Basic vector. Fig. 5B shows functional analysis of the 5Ј-region broken down into 10 different reporter constructs. As noted in Fig. 5A, the 600-bp construct was able to induce luciferase activity in both cell types, thus defining a core promoter region containing the cis-acting regulatory elements responsible for directing NKR-P1C expression. In addition, transfection of reporter constructs ranging from 600 bp to the full 9.76 kb revealed a pattern of enhanced activities in MNK-1 cells, yet suppressed activities in SL-12 cells. This suggests that additional binding sites for regulatory proteins exist in the 5Ј-regions of the NKR-P1C gene. Significantly, constructs containing a minimum of the first 1 kb of flanking sequence enhanced lineage-specific expression (augmented levels in MNK-1 Ն SL-12), while fragments 3 kb and larger silenced ectopic expression (depressed levels in SL-12 Ն MNK-1). In general, larger fragments further augmented the difference between MNK-1 and SL-12 activities. This pattern suggests the exist-ence of both enhancer and silencer elements upstream of NKR-P1C.
The identification of a longer NKR-P1C transcript containing novel 5Ј-untranslated exons suggested that an additional promoter existed. To survey cell lines for the activity of this putative new promoter, termed promoter 2, we designed a reporter luciferase construct (Fig. 5C, 500bp-P2) containing the putative promoter 2 in the absence of the 600 bp NKR-P1C minimal promoter (600 bp promoter 1). As can be observed from Fig. 5D, the promoter 2 (500bp-P2) failed to induce significant luciferase activity, and thus appears to be inactive in MNK-1 cells. This observation helps explain our inability to detect a product in the primer extension analysis of exon 1.1 in MNK-1 cells (data not shown). Interestingly, the 500 bp promoter 2 construct displayed low activity in the mature NK1.1 ϩ T cell line, CTLL-2 (Fig. 5D), suggesting that it is a very weak promoter and explaining the faint signal obtained from primer extension experiments using CTLL-2 cells.

Identification of a DNaseI Hypersensitive Site (HS1) in the Upstream NKR-P1C Region That Exhibits Enhancer/Promoter
Activity-In order to identify 5Ј DNA elements that functionally bind regulatory proteins, we employed a DNaseI hypersensitivity assay. To facilitate this, the full 9.76-kb NKR-P1C upstream region was cloned in front of GFP as a reporter, and this vector was used to generate stable transfectants of MNK-1 cells. Flow cytometric analysis reveals that more than 90% of these cells were GFP ϩ (Fig. 6A), demonstrating that the 9.76-kb fragment drives high level GFP expression in stable MNK-1 transfectants. Similarly transfected SL-12 cells failed to show any expression of GFP (data not shown). To analyze the binding of MNK-1 transcription factors to the 9.76-kb upstream DNA, nuclei from these cells were prepared and treated with different amounts of DNaseI. Recovered DNA was digested with NotI and assayed for DNaseI hypersensitive sites by Southern blotting using a GFP-specific probe. These experiments revealed a single hypersensitive site ϳ9-kb upstream of

1) by primer extension and RT-PCR.
A, primer extension products were size-fractionated on a denaturing polyacrylamide gel alongside a Sanger dideoxynucleotide sequencing ladder primed on an NKR-P1C genomic DNA subclone. The G nucleotide corresponds to position Ϫ540 relative to the major transcriptional initiation site (large right arrow in Fig. 2). B, RT-PCR of various tissues showing expression levels of NKR-P1C transcripts corresponding to exon 1, exon 1.1, or ␤-actin as a control. cDNA was generated from total RNA isolated from C57BL/6 (B6) splenic NK cells, CTLL-2 cells, and B6 day 15 fetal thymus and fetal liver. PCR was performed using 5Ј-primers that correspond to either exon 1.1 (RT-PCR oligo F1B; Fig. 2) or exon 1 (RT-PCR oligo F1C; Fig. 2), and a common 3Ј-primer (RT-PCR oligo R1B/C; Fig. 2). the starting ATG codon (Fig. 6B). The region of DNaseI hypersensitivity appeared to span about 500 bp of upstream region and was termed HS1. Computational analysis of HS1 for DNA binding sites revealed multiple high scoring GATA, Ikaros, and AP-1 sites (core similarity ϭ 1.0, matrix similarity ϭ 0.95) (Fig.   7). This suggested that HS1 might exhibit functional enhancer activity.
To test whether HS1 exhibits enhancer activity, this 500 bp genomic region was cloned into a luciferase backbone containing a minimal promoter (TATA box only). This construct was transfected into MNK-1 and BW5417 cells (Fig. 8A). The HS1 induced 7.5-fold increase in luciferase activity when trans-

FIG. 5. Functional analysis of the NKR-P1C 5-flanking region.
A, luciferase reporter assays comparing the relative activities of the entire 9.76-kb upstream region and the 600-bp core promoter. NKR-P1C upstream regions were cloned in front of the luciferase reporter gene in pGL-2/Basic. Constructs were assayed for luciferase activity upon transfection into MNK-1 cells and SL-12 cells. The mean values (Ϯ S.E.) of three experiments are shown. Bar graphs show the relative luciferase activity of each construct, plotted as fold induction relative to that of the promoterless luciferase construct (pGL-2/Basic). B, ten constructs containing NKR-P1C upstream regions ranging from 600 bp to 9 kb as shown were generated in pGL-2 and assayed as in A. C, the generation of constructs containing promoter 2 (500bp-P2) and promoter 1 (600bp) in pGL-2 vector backbone are shown. D, the 500-bp-P2 construct and the 600-bp construct containing promoter 1 were transfected into MNK-1 and CTLL-2 cells and assayed as in A.

FIG. 6. Analysis of NKR-P1C DNaseI hypersensitive sites in MNK-1 cells.
A, MNK-1 cells were stably transfected with the 9.76-kb NKR-P1C upstream fragment driving GFP and analyzed for GFP expression after selection. B, nuclei from MNK-1 9.76kb/GFP stable transfectants were isolated and treated with increasing concentrations of DNa-seI. DNA isolated from these nuclei was subsequently digested with NotI, and Southern blotted using a GFP-specific probe. The upper arrow corresponds to the NotI restriction endonuclease fragment, while the lower arrow identifies the DNaseI hypersensitive site, termed HS1.

FIG. 7.
Nucleotide sequence of the NKR-P1C 5-HS1 site. The nucleotides are numbered relative to the starting ATG codon (Fig. 2,  bold, underlined). Putative DNA binding sites are underlined with those in reverse orientation indicated by an asterisk. Locations of the Inr, and TATA motifs are underlined and in bold letters. The putative transcriptional initiation site is indicated by the forward arrow. The intron/exon boundary between exons 1.2 and 2.2 (Fig. 2) is represented by dotted vertical lines. Primers used for PCR amplification of HS1 fragment for cloning into pGL-2 backbone (FHS1 and RHS1) and for RT-PCR of promoter 3-derived transcript (F1A) are indicated. fected into MNK-1 cells compared with the TATA box promoter alone (Fig. 8A), while only a moderate 2-fold increase was observed in BW5417 cells (Fig. 8A).
Since the HS1 enhanced the expression of luciferase from a heterologous promoter, we decided to test its ability to increase luciferase expression from the NKR-P1C minimal functional promoter. Thus, the HS1 fragment was cloned into pGL-2 directly in front of the 600 bp NKR-P1C upstream core promoter region. This construct (pGL-2/HS1/600bp) was transfected into the MNK-1, SL-12, and BW5417 cell lines (Fig. 8B). The parent vector without HS1 (pGL-2/600bp) was used as a promoteralone control. Interestingly, the HS1 construct induced a 4-fold increase in luciferase activity when transfected into MNK-1 cells, compared with the promoter alone (Fig. 8C). In contrast, a more moderate 1.5-2-fold increase was seen in SL-12 and BW5417 cells, respectively (Fig. 8C). This suggests that the 500-bp HS1 element of the NKR-P1C gene contains enhancerlike activity, which is at least partially cell-type specific. Thus, although other sites not identified in this study may be responsible for further lineage-specific control of NKR-P1C expression, these experiments collectively outline a minimal functional NKR-P1C promoter and a unique enhancer element, HS1.
To further characterize HS1, this element was cloned into an empty luciferase backbone (HS1 construct). This construct and the construct containing the NKR-P1C minimal promoter (600 bp) were transfected into MNK-1 and CTLL-2 cells (Fig. 9A). Although the 600-bp minimal promoter is capable of driving luciferase expression in both lines, the HS1 construct was capable of inducing luciferase activity only in MNK-1 cells. Nevertheless, the observed luciferase activity induced in MNK-1 cells by HS1 was significantly lower than that induced by the 600-bp NKR-P1C promoter (Fig. 9A). This level of luciferase expression in MNK-1 cells could be due to the presence of multiple NK cell-specific elements (Fig. 7) that might be responsible for inducing luciferase expression in MNK-1 cells even in the absence of a defined promoter, as has been reported for other cell lines (30), or perhaps the HS1 element contained a novel early NK cell-specific promoter. To address the possibility that HS1 could serve as promoter, we examined whether we could detect transcripts initiated by this element. Fig. 9B shows an RT-PCR analysis using a forward primer positioned downstream of the potential initiator consensus sites (Fig. 7, underlined, bold) and a reverse primer located within exon 2 (Fig. 2, GSP nested). Using this approach, a barely-detectable transcript derived from HS1 was observed only in sorted NK1.1 ϩ cells from the fetal thymus (Fig. 9B), suggesting that HS1 contained a novel promoter (promoter 3). The PCR product was cloned and sequenced to verify that the observed band resulted from HS1/promoter-3 activity. The sequence analysis revealed a novel exon (E1.2) that spliced to exon 1, creating a novel exon 2.2 (splice donor and acceptor sites are shown in Fig.  7 and Fig. 2, respectively, as dotted vertical lines). Similarly to exon 1.1 from promoter 2, the initial ATG codon in E1.2 was followed by several in-frame stop codons, suggesting that this exon was also noncoding. DISCUSSION In this study, we have cloned and sequenced about 10 kb of the upstream region of the mouse NKR-P1C gene. We have defined the transcriptional initiation sites, a minimal functional promoter, and an enhancer region responsible for regulating expression of NKR-P1C in a lineage-specific manner. Using a combination of primer extension and 5Ј-RACE methodologies, a single major transcriptional initiation site was identified 156 bp upstream of the ATG coding the N-terminal methionine of NKR-P1C. This initiation site contains the TCA (ϩ1) GTCA sequence, which shows remarkable homology to the initiator (Inr) consensus. There is no canonical TATA box within normally expected distances from the transcriptional initiation site of the mouse NKR-P1C. Although the majority of TATA-less genes are generally thought to be constitutively FIG. 8. Functional analysis of the NKR-P1C HS1 DNA element. A, a 500-bp HS1 fragment was cloned in front of the minimal promoter (TATA box)-containing luciferase backbone plasmid. Constructs were transiently transfected into MNK-1 and BW5147 cells and assayed for luciferase activity. Bar graphs represent fold induction of luciferase activity relative to that obtained from cells transfected with the TATA construct alone. The mean values (ϮS.D.) of two experiments are shown. B, the 500-bp HS1 DNA fragment was cloned in front of the 600-bp minimal NKR-P1C promoter driving a luciferase reporter gene (pGL-2/600bp). Constructs were transiently transfected into MNK-1, SL-12, and BW5147 cells and assayed as in A. Bar graphs represent the relative luciferase activities for each cell line. The values (ϮS.E.) of three independent experiments are shown. C, relative luciferase values from B were plotted as fold induction relative to those of the pGL-2/600bp construct.
The biological significance of the absence of a TATA box among such lymphocyte specific genes is not yet understood. However, studies of numerous TATA-less genes have demonstrated that the Inr element constitutes the simplest functional promoter identified and provides one explanation for how TATA-less promoters direct transcriptional initiation (28). About 30-bp downstream of the NKR-P1C transcriptional start site, we observed the sequence 5Ј-AGACAGG-3Ј, which shows striking homology to the downstream promoter element (DPE) (37). Both Inr and DPE are required for sequence-specific binding of TFIID in TATA-less promoters (37). Furthermore, at positions Ϫ30 to Ϫ35, there is an A/T-rich sequence (5Ј-TTA-AGT-3Ј), which, in addition to Inr and DPE, may contribute to the assembly of the transcription preinitiation complex.
Examination of the nucleotide composition of the NKR-P1C 5Ј-flanking region reveals the region to be depleted in CpG dinucleotides. This is a characteristic of many differentiationassociated and developmentally regulated genes with TATAless promoters (28). There are no CpG dinucleotides in any of the putative cis-elements in the 5Ј-flanking region of mouse NKR-P1C. The NKR-P1C promoter region contains multiple consensus binding sites for the Ets-1, Ikaros, GATA, and TCF-1 factors important in regulating expression of many lymphocyte-specific genes. It is not surprising to find several (at least 5) Ets-1 binding sites in the 100-bp region immediately upstream of the transcriptional initiation site, as mice with a disrupted Ets-1 gene lack NK cells due to impaired survival and/or development of NK cells (38,39).
In expression experiments, the NKR-P1C 5Ј-flanking region confers significant reporter activity when introduced into NK lineage cells. The proximal promoter region (600-bp construct) drove transcription in both NK lineage (MNK-1) and non-NK (SL-12 and BW5147) cell lines. The observation that this core promoter activity was obtained in a non-NK cell line, in which NKR-P1C is not expressed, indicates that the tissue-specific expression of NKR-P1C involves other factors, such as enhancer or silencer cis-elements, that may act in concert with the promoter region. Consistent with this, constructs containing the upstream NKR-P1C region, including the HS1 DNaseI hypersensitive site alone, impart a cell-type specific expression pattern. Further biochemical and functional studies are currently in progress to identify the precise cis-elements contributing to the observed enhancer/silencer functions of the 9.76-kb upstream fragment.
We provide evidence for the existence of several novel promoters (promoters 1-3), with the 600 bp 5Ј-flanking region serving as the principal promoter (promoter 1) active in all NK1.1 ϩ cells (Fig. 10A). The additional upstream promoters (2 and 3) appeared to be developmentally regulated, with promoter 2 active in mature NK cells and promoter 3, which is FIG. 9. Characterization of promoter 3 function and detection of promoter 3-derived transcripts. A, the 500-bp HS1 fragment was cloned into an empty luciferase backbone (HS1 construct). HS1 and 600 bp minimal promoter construct were transiently transfected into MNK-1 and CTLL-2 cells and assayed for luciferase activity. Bar graphs represent fold induction of luciferase activity relative to that obtained from cells transfected with an empty luciferase backbone (pGL-2/Basic). The mean values (ϮS.D.) of three experiments are shown. B, RT-PCR of various tissues showing expression levels of NKR-P1C transcript 3-containing exons 1.2 (E1.2) and 2.2 (E2.2), or ␤-actin as a control. cDNAs were generated from total RNA isolated from C57BL/6 day 15 fetal thymus, NK1.1 ϩ sorted day 15 fetal thymus, CTLL-2 cells, and MNK-1 cells. PCR was performed using 5Ј-primer F1A (Fig. 7), and 3Ј-primer GSP nested (Fig. 2). C, genomic fragment of the NKR-P1C upstream region showing exons 1.2, 2.2, and 2. Large arrow represents the major transcription initiation site, while the small arrow represents the upstream transcription initiation site. Schematic representation of the E1.2-and E1-containing transcripts (transcript 3 and transcript 1, respectively) and the corresponding protein regions are indicated. * represents ATG codons and # represents stop codons. located within the HS1 site, active in fetal-derived NK cells. Additionally, transcripts from both of these promoters contained noncoding exons, due to the presence of multiple stop codons following the first few ATGs. The above findings suggest that mouse NKR-P1C gene shares many similar features with the genes encoding the other NKC-encoded lectin-like proteins, such as the Ly49 gene family (30, 40 -42). Indeed, it has recently been shown that in addition to the Ly49 promoter used by mature NK cells (Pro-2 or Pro-3) there is also a fairly upstream Ly49 promoter (Pro-1) that is primarily active in tissues where Ly49 gene expression is first detected during ontogeny (40) (Fig. 10A). Moreover, as shown for NKR-P1C, the first exon of Ly49a is non-coding, followed by the cytoplasmic (second exon), transmembrane (third exon), stalk (fourth exon), and putative carbohydrate recognition domains (CRD; fifth to seventh exons), respectively (Fig. 10B) (41). These findings suggest a clear commonality in the structure and transcriptional regulation of these NK cell related genes that may have been conserved throughout evolution in order to maintain important features of NK cell function.
In conclusion, the experiments outlined in this study provide an insight into the molecular basis of NKR-P1C expression in NK cells, NK1.1 ϩ T cells, and progenitor thymocytes. We have shown that the 600 bp 5Ј-flanking region of the NKR-P1C gene contains cis-elements required for basal transcription or core promoter function. In addition, we have identified a DNA element (HS1) ϳ9-kb upstream that functions to enhance transcription in a lineage-specific manner and may also serve as a developmentally-regulated promoter. Studies are currently underway to define tissue-and stage-specific elements regulating NKR-P1C expression in vivo.