| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Department of Immunology, University of Toronto, Sunnybrook and Women’s Research Institute, Toronto, ON, Canada;
Laboratory of Molecular Immunology, Institut de Recherches Cliniques de Montréal, Montréal, QC, Canada
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Nonetheless, the underlying molecular basis for the lack of NK1.1 reactivity of NK cells from BALB/c and other mouse strains remains an enigma. Previous studies have suggested that BALB/c NK cells either lack or possess low-level expression of the Nkrp1 genes, as detected by Northern blotting (10); however, no genetic basis for a BALB/c defect in Nkrp1 expression has been established. Thus, lack of NK1.1 reactivity could be due to deletion of Nkrp1 genes, defective gene expression, or allelic polymorphism in BALB/c mice. Extreme variation in gene content between the BALB/c and B6 haplotypes has been observed previously for the related Ly49 gene family (11). This suggests that other NK gene complex (NKC)4 regions, including the Nkrp1-Ocil/Clr region, also may be subject to rapid evolutionary divergence and/or polymorphism.
Importantly, since cognate NKR-P1 ligands have recently been identified (12, 13, 14), a BALB/c defect in NKR-P1 expression could be functionally significant for NK cell function and innate immunity. Moreover, while ligands for the stimulatory NKR-P1A/C receptors remain elusive, ligands for the inhibitory NKR-P1B/D receptors (12) and stimulatory NKR-P1F (13) receptors have been identified as products of the Ocil/Clr family of genes (15, 16, 17), which are intermingled with the Nkrp1 genes themselves in the NKC (18). Thus, determination of the gene content of the BALB/c Nkrp1-Ocil/Clr region and the basis of the BALB/c defect in NK1.1 expression could have implications for the importance of the NKR-P1–Ocil/Clr system in self-nonself discrimination in mice and other species. Therefore, we have undertaken NK1.1 epitope mapping and genomic analysis of the BALB/c Nkrp1 region to determine the functional significance of the NKR-P1 recognition system in NK1.1– mouse strains.
In this study, we demonstrate that the BALB/c genome possesses a full complement of Nkrp1 and Ocil/Clr genes, including novel family members present in both B6 and BALB/c mice. Furthermore, BALB/c NK cells possess normal Nkrp1 expression relative to B6 NK cells, and the BALB/c NKR-P1B receptor functionally binds cognate Ocil/Clr-b ligand. Absent NK1.1 reactivity of BALB/c NK cells can be explained by Nkrp1 allelic divergence, specifically a single amino acid substitution (S191T) present in the BALB/c NKR-P1B/C receptors. In fact, divergence between the B6 and BALB/c Nkrp1-Ocil/Clr regions appears to be localized to the Nkrp1b/c genes, and the B6 Nkrp1d gene appears to be an allele of the Nkrp1b gene found in BALB/c and other mouse strains. These results confirm the importance of the NKR-P1–Ocil/Clr recognition system across strain boundaries, and suggest that this mode of self-nonself discrimination is conserved in other NK1.1– mouse strains, as well as other species.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Mice were purchased from the Charles River Laboratories National Cancer Institute–Frederick Animal Production Area (Frederick, MD), and/or maintained in our own animal facilities.
Cells
BWZ.36 cells (19) were obtained from Dr. N. Shastri (University of California, Berkeley, CA). The 293T cells were obtained from Dr. D. Raulet (University of California, Berkeley, CA). Cells were grown in complete DMEM-HG (high glucose, 10% FCS). All cells were maintained at 37°C in a humidified 5% CO2 atmosphere.
Expression constructs and NKR-P1 mutants
Mouse Nkrp1b (National Institutes of Health–Swiss–Sw) and Nkrp1a/c/d (C57BL/6–B6) constructs were generated previously (5). Mouse Nkrp1e/f cDNAs were amplified from NK libraries using the following primers: P1E-F, ATGGACACAGAAAGGATCTACCTC; P1E-R, TCAGGAGTCATGAAATATGGTTTC; and P1F-F, ATGGACACATCAAAGGTCCATG; P1F-R, TCAGACATGTATCAGGGTCTTTTG. Rat Nkrp1a/b cDNA clones (20) were generously provided by Dr. R. G. Miller (Ontario Cancer Institute, Toronto, ON). Site-directed mutagenesis was performed by the method of gene splicing by overlap extension of DNA ends, as described previously (12). Briefly, known Nkrp1 sequences were amplified by PCR using specific 5' or 3' outside primers and complementary internal primers with specified point mutations; the resulting products were then mixed and amplified using the outside primers alone to obtain the full-length mutant cDNA. All PCR products were directly cloned into pcDNA3.1/V5/HIS/TOPO (Invitrogen Life Technologies), and constructs were sequenced to confirm their identities. The resulting constructs were cotransfected into 293T cells using Effectene reagent (Qiagen) with a pCMV-GFP-nuclear localization sequence reporter, or subcloned into the pMSCV2.2-CMV-IRES-GFP (pMCIG) retroviral vector (12) before transfection into 293T cells. Transfection efficiencies were monitored by GFP expression, and transfection results are shown gated on GFP+ cells for all contstructs tested.
Flow cytometry and cell sorting
Cells were stained as described previously (21). Stained cells were analyzed using a FACSCalibur flow cytometer (BD Biosciences) or were sorted on a FACSDiVa (BD Biosciences). Sorted cells were >99% pure, as determined by postsort analysis.
Receptor fusions, retroviral infections, and BWZ assay
A pMSCV2.2-CMV-IRES-GFP (pMCIG) retroviral vector (12) was modified to include a CD3
/NKRP1B cassette (intracellular region of CD3, membrane proximal and transmembrane regions of NKR-P1B). The CD3/NKRP1B fusion cassette was generated by gene SOEing using the following primers: SALCD3Z-F, GTCGACATGAGAGCAAAATTCAGCAGGAGT G; CD3ZP1B-R, GCGAGGGCACCGACAGCGAGGGGCCAGGGTCTGC; CD3ZP1B-F, ACCCTGGCCCCTCGCTGTCGGTGCCCTCGCTGGCATC; XHOP1BTM-R, CTCGAGTGATGATTTTTGTACTGATAG. The resulting product was TOPO TA-cloned (Invitrogen Life Technologies) and cut with SalI and XhoI (New England Biolabs), then ligated into pMCIG that had been linearized with XhoI and treated with CIP. cDNA inserts were subcloned into the modified cassette vector, pMC3BIG, using XhoI and NotI sites. BWZ.36 cells were infected with retroviral supernatants (24–48 h) from transient triple-transfected 293T cells, as described previously (12), then sorted at day 3–4 following infection. Stable BWZ transductants were analyzed using plate-bound mAb or cell mixtures, as described previously (12).
cDNA libraries
BALB/c DX5+ NK cell (22) and BALB/c-congenic C.B-17/SCID LAK (23) cDNA libraries were reported previously. The NKC of C.B-17/SCID (BALB/c.C57BL/Ka-Igh-1b scid/scid) mice has been confirmed to be BALB/c in gene content (Ref. 24 and data not shown). A (B6 x BALB/c)F1 LAK library was generously provided by Dr. R. G. Miller (Ontario Cancer Institute, Toronto, ON). A B6 DX5+ NK cell cDNA library was provided by Dr. H. Arase (Osaka University, Japan).
Evaluation of BAC gene content
BAC clones from the CHORI-28 BALB/c library (BACPAC Resources) containing Nkrp1 or Ocil/Clr genes were identified and size-estimated as described previously (25). The gene content of BACs was determined by PCR and sequencing of products. Briefly, 
10 ng of BAC DNA was subjected to PCR (94°C 30 s, 57°C 30 s, 72°C 1 min, 30 cycles) with primers capable of amplifying known and predicted BALB/c fragments of Nkrp1 and Ocil/Clr gene sequences (primers are shown in Table I). PCR products were separated on 1% agarose gels and visualized with ethidium bromide. PCR products of each size were cloned by TOPO TA-cloning into pCR2.1 (Invitrogen Life Technologies) and sequenced in-house using T7 and M13 reverse primers. All BAC PCR amplifications were performed using Taq polymerase on enriched, high-copy BAC DNA. Importantly, no mutations were detected between distinct PCR (i.e., at least two clones of each product from separate PCR were sequenced). Partial exon sequence allowed the identification of the Nkrp1 or Ocil/Clr gene that was amplified from each BAC clone. The partial gene sequences from the BALB/c BACs were deposited in GenBank under the following accession numbers: Nkrp1a (exons 3–5, DQ143102); Nkrp1b (exons 3–5, DQ143103); Nkrp1c (exon 1, DQ143106); Nkrp1c (exons 3–5, DQ336140); Nkrp1e (exon 2, DQ143104); Nkrp1f (exons 3–5, DQ143105); Nkrp1g (exons 3–5, DQ336141); Clr-a (exons 3–4, DQ143108); Clr-b (exons 3–4, DQ143111); Clr-c (3' end, DQ143114); Clr-d (exons 4–5, DQ143107); Clr-e (exons 4–5, DQ143109); Clr-f (exons 3–4, DQ143110); Clr-g (exons 3–4, DQ143112); and Clr-h (exons 3–4, DQ143113). Sequence-confirmed gene fragments were used in Southern blot analyses of BAC DNA to confirm Nkrp1-Ocil/Clr gene content as described previously (25).
|
Full-length cDNA clones were obtained using fresh cDNA isolated from d6 BALB/c LAK cells (plastic-adherent IL-2 lymphokine-activated killer cells). RNA was isolated using TRIzol (Invitrogen Life Technologies) and reverse-transcribed into cDNA with SuperScript first-strand cDNA synthesis kit (Invitrogen Life Technologies). PCR conditions, visualization, cloning, and sequencing were performed as outlined above. Library cDNA PCR was performed using primers specific for either the full-length Nkrp1a/b/c coding regions (5) or the corresponding extracellular/3'-untranslated regions (21), under limiting cycle conditions (25 cycles PCR with 50 ng cDNA library, representing 10–20 ng cDNA equivalents). For mammalian expression studies, PCR products were cloned directly into pcDNA3.1/V5/HIS/TOPO (Invitrogen Life Technologies) and sequenced to confirm identity and orientation. All cDNA PCR amplifications were performed using Expand High Fidelity enzyme (Roche Diagnostics), and sequences were determined from multiple clones from at least two separate PCR. Sequences were deposited in GenBank under the following accession numbers: Nkrp1aBALB, DQ237927; Nkrp1bBALB, DQ237928; Nkrp1cBALB, DQ237929; Nkrp1eBALB, DQ237930; and Nkrp1fBALB, DQ237931). Phylogenetic analysis was performed as described previously (26).
BAC end-sequence characterization
BAC DNA (1 µg) was sequenced in-house with T7 and SP6 primers. Based on end-sequence results, primers were designed to PCR-amplify the respective BAC ends. Products were cloned using a TOPO-pCR2.1 kit (Invitrogen) and sequences were confirmed. Cloned end-sequence fragments were radioactively labeled and used as probes in Southern blots of EcoRI-digested BAC DNA.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
One simple explanation for the NK1.1– phenotype of BALB/c NK cells is that the two genes known to confer NK1.1 reactivity, Nkrp1b and Nkrp1c, like Ly49d and Ly49h, may not be present in the BALB/c genome (11). BALB/c NK cells have been reported to be deficient in Nkrp1 transcripts, as detected by Northern blotting (10). Thus, lack of NK1.1 reactivity of BALB/c NK cells could be due to deletion of Nkrp1 genes or defective gene expression.
To investigate this possibility, we constructed a physical map of the BALB/c Nkrp1-Ocil/Clr region. To this end, a BALB/c genomic library, CHORI-28, was probed with a mixture of known Nkrp1 and Ocil/Clr cDNAs. Positive BAC clones were isolated and further screened by PCR using gene-specific primers for the known Nkrp1 and Ocil/Clr genes (Table I). PCR products were sequenced to confirm gene identity, and Southern blots of EcoRI-digested BACs were probed to confirm the presence of specific genes in each BAC clone (Table II). In some instances, the primer pairs cross-amplified closely related sequences (see Table I), resulting in the discovery of new gene fragments. One such novel sequence, cross-amplified using Clr-b primers, corresponds to an exon-3/intron-3/exon-4 (e3-i3-e4) genomic fragment resembling a near-identical (>99%) match to a sequence available from the latest B6 genome assembly (designated Clr-h; GenBank accession no. XM487965). The matching region in the B6 genome does not correspond to the location of any other known Clr gene. This suggests that the Clr-h sequence represents a new Clr family member conserved in the B6 and BALB/c genomes. In addition, a new gene was detected in the region between Clr-f and Nkrp1c based on Southern hybridization using an Nkrp1f probe. This sequence also resembles a near-identical (>99%) match in the B6 genome (designated Nkrp1g; GenBank accession no. XM355818) that does not correspond to the location of any other known Nkrp1 gene. Thus, two new genes, Clr-h and Nkrp1g, are conserved in the B6 and BALB/c genomes.
|
In contrast with genomic PCR, PCR of cDNA libraries derived from BALB/c NK cells revealed the existence of an Nkrp1c-like transcript; however, only a partial extracellular domain fragment could be amplified, due to the location of the primers (data not shown). Southern analysis using this cDNA product, new fragments derived from the cDNA, and the previous genomic Nkrp1cBALB exon-1 fragment as probes (Fig. 1 and data not shown) revealed that the Nkrp1c-like gene resided centromeric to Nkrp1bBALB, approximating the location of Nkrp1c relative to Nkrp1d in the B6 genome (27). Hybridization with Nkrp1bBALB and Nkrp1cBALB probes showed that several BACs contained these genes (Fig. 1). Similarly, hybridization of the same BALB/c BACs with Nkrp1dB6 and Nkrp1cB6 probes revealed the same bands, albeit with weaker intensity (Fig. 1).
|
A relative gene order of Nkrp1a, Clr-h, Clr-f, Nkrp1g, Nkrp1c, Nkrp1b, Clr-g, Clr-d, Clr-e, Clr-c, Nkrp1f, Clr-a, Nkrp1e, Clr-b, and Cd69 was revealed after assembling the BACs in order of gene content (Table II). There are some ambiguities in this assembly. Notably, the order of Clr-h and Clr-f in the BALB/c genome is unknown because of their identification on shared BACs; however, in the B6 genomic database, Clr-h lies centromeric to Clr-f and is transcribed in the opposite orientation. Similarly, Clr-c and Nkrp1f also were found on the same BACs, but their order is known in the B6 haplotype, with Clr-c lying centromeric to Nkrp1f.
To generate a physical map of the approximate relative location of each gene on the chromosome, the size of each BAC clone was ascertained by pulsed-field gel electrophoresis (Fig. 2 and data not shown). Using the gene content and size of each BAC, the approximate location of each gene on chromosome 6 of BALB/c mice was predicted (Fig. 2). The BALB/c Nkrp1-Ocil/Clr region is predicted to be 640 kb in length (Nkrp1a exon 3 to Clr-b exon 4), with a possible range from 575 to 700 kb (Fig. 2). In previous gene-mapping studies, the entire B6 Nkrp1-Ocil/Clr cluster was predicted to be 700 kb in length (16, 27); however, the exact length is unknown, as the B6 genome assembly for this region contains several gaps.
|
Direct comparison of the gene content between BALB/c and B6 mice shows that at least 13 genes are conserved, including Nkrp1c and the novel, but shared, Clr-h and Nkrp1g genes (Fig. 3). The only BALB/c-specific gene was Nkrp1b, but this gene corresponded to the approximate location of Nkrp1d in the B6 genome (Fig. 2), and both the Nkrp1bBALB and Nkrp1dB6 probes hybridized to the same genes (Fig. 1). Thus, it is likely that the two designations actually represent different alleles of the same gene, especially since their transmembrane and cytoplasmic coding sequences are identical. Therefore, like the B6 and BALB/c Ly49 gene clusters, the Nkrp1-Ocil/Clr regions represented in these mice possess shared genes and have presumably diverged from an ancestral haplotype. However, unlike the Ly49 gene repertoire, that of the Nkrp1-Ocil/Clr genes appears to be highly conserved, perhaps with the exception of a seemingly directed divergence of Nkrp1b and Nkrp1c. These similarities and differences are apparent in Southern analyses of genomic DNA (Fig. 4A). The RFLP patterns generated by the Nkrp1f and Clr-f probes are similar in the two mouse strains, but the banding patterns resulting from the Nkrp1cB6 probe are distinct (Fig. 4A). Hybridization of the Nkrp1cB6 and Clr-f probes was conducted at lower stringency than that of Nkrp1f and confirms that the number of Nkrp1 and Ocil/Clr genes, as evidenced by cross-hybridizing genes, is similar in these two strains.
|
|
BALB/c NK cells express abundant Nkrp1b and Nkrp1c transcripts
To obtain cDNA clones of the Nkrp1 genes from BALB/c mice, RT-PCR was performed on fresh BALB/c A-LAK cDNA and BALB/c NK cell cDNA libraries using multiple primer sets. This approach yielded full-length or near full-length cDNA clones for Nkrp1aBALB, Nkrp1bBALB, Nkrp1cBALB, Nkrp1eBALB, and Nkrp1fBALB (Fig. 5 and data not shown). This was unexpected, given that BALB/c NK cells were reported previously to lack detectable Nkrp1 expression in Northern blotting experiments (10). To determine whether Nkrp1b/c expression was specifically deficient in BALB/c mice, RT-PCR was applied to both BALB/c and B6 NK cell cDNA libraries using limited amounts of cDNA and amplification cycles. The Nkrp1 primer sets used have been previously reported, one set corresponding to the extracellular/3'-untranslated region (EC/UTR) (21), a second set corresponding to the full-length coding sequence (FL/CDS) regions (5). As shown in Fig. 5, Nkrp1 transcripts were easily detected in both strains. However, while the EC/UTR primers yielded a significant signal for Nkrp1cBALB, a FL/CDS Nkrp1cBALB signal could not be detected. This has been observed before for Sw strain NK cells (5, 21), and suggests that the primer sets used may be divergent compared with the Nkrp1cBALB sequence. Interestingly, note that the Nkrp1b FL/CDS primer set amplifies Nkrp1bSw (5), Nkrp1bSJL (6), and Nkrp1bBALB (Fig. 5; see below), whereas it also cross-amplifies Nkrp1dB6 (5, 6). Therefore, to confirm the cDNA identities, and determine the discrepancy of the Nkrp1cBALB results, all PCR products were sequenced.
|
Interestingly, full-length transcripts for Nkrp1cBALB were detected equally in BALB/c, B6, and F1 LAK libraries by this method (Fig. 5B). Moreover, primers spanning exons 3–5 were capable of specifically detecting Nkrp1cBALB transcripts in fresh BALB/c but not B6 NK cells (Fig. 5C). Sequencing confirmed that the Nkrp1cBALB EC/UTR product, the P1A-5'/P1C-3' product, the Nkrp1cBALB exon 3–5 product, and the gene-40 sequence (GenBank accession no. X64720) (10) were all identical with one another. These results confirm the existence of a full-length coding Nkrp1c cDNA in BALB/c NK cells. However, comparison of the Nkrp1cBALB cDNA with the published Nkrp1cB6 cDNA reveals that they differ by 20 nonsynonymous substitutions in the extracellular coding region alone (Fig. 6A). Such divergent sequences would be expected to confound attempts using Northern blotting to compare expression of Nkrp1c transcripts between the two mouse strains (10). Notably, the similarity of the Nkrp1cBALB 5' region to Nkrp1aB6 is reminiscent of Ly49O129/C57L, which resembles Ly49AB6 upstream and Ly49DB6 downstream (29, 30).
|
99% identical with Nkrp1bSw/SJL, differing by only three nonsynonymous substitutions in its extracellular coding region (Fig. 6A). The ITIM motif in the Nkrp1bBALB coding sequence is intact, and it lacks a charged residue in its transmembrane region, so this cDNA is predicted to code for a functional inhibitory transmembrane receptor. A near full-length (exon 5 deleted) in-frame cDNA clone of Nkrp1eBALB was also isolated by this method (Fig. 6A); in contrast, Nkrp1e is only known from single genomic exons in B6 mice (27). The Nkrp1fBALB cDNA sequence is identical with the published Nkrp1fB6 sequence (Fig. 6A). Three independent primer sets failed to amplify an Nkrp1gB6/BALB cDNA (Fig. 5C and data not shown), so this gene may represent a pseudogene, at least in the strains tested. When these Nkrp1 sequences (and the B6 Ocil/Clr sequences) are compared over their coding regions, the Nkrp1bBALB and Nkrp1bSw/SJL sequences are closely associated with each other and with Nkrp1dB6 (Fig. 6B). The Nkrp1cBALB sequence is most closely related to the allelic sequence, Nkrp1cB6. All other BALB/c Nkrp1 genes are closely related to their B6 alleles. The novel gene, Clr-h, groups with the Clr-b/c/d/g clade (Fig. 6C). All Clr genes are shown without allele designation, as the BALB/c full-length coding sequences were not available for direct comparison.
Thus, while our results using RT-PCR differ from previous Northern blot analysis of Nkrp1 transcript expression in BALB/c mice, our sequencing analysis confirms and extends previous findings, where our Nkrp1a/b/cBALB sequences correspond to the BALB/c partial sequences of gene 2 (GenBank accession no. X64723), gene 34 (GenBank accession no. X64719), and gene 40 (GenBank accession no. X64720), respectively (10). Thus, BALB/c mice are not deficient in Nkrp1 expression, they simply possess divergent Nkrp1 sequences compared with those of B6 mice.
BALB/c NKR-P1B/C proteins lack NK1.1 reactivity due to allelic divergence
Our results demonstrate that BALB/c mice possess a full Nkrp1 gene content and normal Nkrp1b/c transcript expression in NK/LAK cells when compared with NK1.1+ mouse strains, suggesting that the basis of NK1.1 alloreactivity between these mouse strains may be due to divergence in NKR-P1 protein sequences. As shown in Fig. 7, NK1.1 reactivity is specific to NKR-P1BSw/SJL and NKR-P1CB6 among known NKR-P1 proteins from both mice and rats. However, sequence alignment of these proteins does not directly reveal an amino acid sequence unique to NK1.1+ vs NK1.1– isoforms (Fig. 7A and data not shown). Therefore, we limited our analysis to the mouse NKR-P1 isoforms, and further focused on differences between the mouse NKR-P1BSw/SJL and NKR-P1BBALB sequences, because the two differ in NK1.1 reactivity, but only differ in sequence by three amino acids (see Fig. 7A). Of these three substitutions in the NKR-P1BSw/SJL sequence, two were expected to be nonconservative, resulting in charge alterations in the NKR-P1BBALB isoform (D183V, E217K). However, the D183V substitution also is present in the NK1.1+ NKR-P1CB6 isoform, while the E217 amino acid is intact in the NK1.1– NKR-P1CBALB isoform. This left a seemingly conservative S191T substitution as a candidate determinant for NK1.1 epitope reactivity (Fig. 7A). Interestingly, both NK1.1-reactive isoforms (NKR-P1BSw/SJL, NKR-P1CB6) possess an S residue at this position, whereas all other mouse NKR-P1 sequences have substitutions, including NKR-P1BBALB (T), NKR-P1CBALB (T), NKR-P1DB6 (A), NKR-P1D129 (T), NKR-P1AB6/BALB (A), NKR-P1EB6/BALB (T), NKR-P1FB6/BALB (I), and the predicted NKR-P1GB6 (L). Therefore, to test the significance of this residue in conferring NK1.1 reactivity, the NKR-P1BBALB sequence was altered by site-directed mutagenesis.
|
-CE immunizations used to generate both the polyclonal -NK-1 (3) and monoclonal -NK1.1 (PK136; Ref. 4) specificities, is due to a single amino acid substitution in either NKR-P1B, NKR-P1C, or both receptors. Due to a lack of information about NKR-P1 expression in the CE mouse strain, we cannot distinguish these possibilities at present. Nonetheless, these results demonstrate that BALB/c mice are not deficient in NKR-P1 expression, rather they possess divergent NKR-P1 sequences sufficient to generate alloantibodies when immunized with NK cells from other mouse strains. The BALB/c NKR-P1B receptor is functional and recognizes cognate Ocil/Clr-b ligand
Although BALB/c mice express functional NKR-P1B/C transcripts, it remains formally possible that BALB/c NK cells may be deficient in NKR-P1-mediated recognition, if substitutions in their NKR-P1 coding sequences lead to loss of ligand recognition. Although cognate ligands for the NKR-P1A/C receptors remain unknown, NKR-P1-mediated missing-self recognition of tumor cells has been shown to depend on functional recognition of Ocil/Clr-b ligand by the NKR-P1B/D inhibitory receptors in various mouse strains (12). Therefore, we tested whether the NKR-P1BBALB receptor is capable of recognizing Ocil/Clr-b ligand.
To do this, we used our previous modification of the BWZ reporter-cell assay (12, 19). In this study, expression of a chimeric CD3/NKR-P1B fusion receptor on BWZ.36 reporter cells signals functional recognition of ligand expressed on the surface of intact target cells, via NFAT-driven 
-galactosidase enzyme expression. As shown previously, BWZ cells expressing the CD3/NKR-P1BSw/SJL receptor are capable of recognizing target cells (either BWZ or 293T) expressing Ocil/Clr-b ligand, but not target cells alone (BWZ.P1BSw/SJL; Fig. 8A). Moreover, this interaction is specific for Ocil/Clr-b ligand, as it can be blocked using the -Ocil/Clr-b mAb, 4A6 (Fig. 8A). Importantly, BWZ cells expressing the NKR-P1BBALB receptor are also capable of specifically recognizing Ocil/Clr-b ligand (BWZ.P1BBALB; Fig. 8A), while BWZ cells expressing a control CD69 fusion receptor do not recognize Ocil/Clr-b (BWZ.CD69; Fig. 8A). Furthermore, BWZ.P1BBALB reporter cells also recognize native Ocil/Clr-b ligand on the surface of various hemopoietic cells ex vivo, including bone marrow (BM), lymph node (LN), spleen (SP), and thymus (TH) (Fig. 8B). These results demonstrate that the BALB/c NKR-P1B receptor is functional and recognizes cognate Ocil/Clr-b ligand on the surface of normal cells. Thus, the NKR-P1B–Ocil/Clr-b missing-self recognition system is intact in BALB/c mice despite a single amino acid substitution that is responsible for the lack of NK1.1 reactivity of BALB/c NK cells.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
-NK1.1 mAb, PK136 (4) (Fig. 7). Despite allelic differences, the BALB/c NKR-P1B receptor is fully functional and recognizes Ocil/Clr-b ligand on the surface of transfected and normal cells (Fig. 8). Our finding that the BALB/c and 129 strains appear to share a related Nkrp1 haplotype (Fig. 4) suggests that other NK1.1– mouse strains possess similar allelic differences. These results indicate that the NKR-P1–Ocil/Clr receptor-ligand system plays a broader and more important role in target cell recognition than previously thought. Insight from the BALB/c Nkrp1 region: Nkrp1b and Nkrp1d delineate divergent murine NKC haplotypes
A significant finding of this study is that the mouse Nkrp1b and Nkrp1d genes appear to represent alleles of the same genetic locus (Fig. 3). It has long been suspected that this might be the case, for a number of reasons. Although the original Nkrp1b cDNA sequence (gene 34; Ref. 31) was reportedly derived from B6 mice, along with Nkrp1a (gene 2; Ref. 31); mNKR-P1.7 (9, 28) and Nkrp1c (gene 40; Ref. 31); mNKR-P1.9 (9), we and others have been unable to identify an Nkrp1b genomic or cDNA sequence, or an NKR-P1B protein, from the B6 background (5, 6, 27). However, our group and Kung et al. (6) were able to isolate an identical Nkrp1b cDNA sequence derived from the Sw (5) and SJL strains (6), respectively. Thus, it is possible, due to the large numbers of NK cells required for cDNA library preparation and Northern blotting, that a mixed strain may have been included in the original identification of the mouse Nkrp1 cDNA family. In any case, we and others were successful in cloning an Nkrp1b-related cDNA from B6 NK cells by RT-PCR (5, 6). Due to significant coding sequence divergence in the extracellular region (21 nonsynonymous substitutions, a difference similar to that observed between other Nkrp1 genes), we gave this cDNA a novel designation, Nkrp1d (GenBank accession no. AF338321; Ref. 5); however, Kung et al. (6) only detected 15-aa substitutions and designated their cDNA sequence, Nkrp1bB6 (GenBank accession no. AF354260). Notwithstanding these differences, our collective cDNA findings were later confirmed at the genomic level through random sequencing of the B6 Nkrp1 region: no evidence for the existence of a B6 Nkrp1b sequence could be found, resulting in the designation of the B6 gene, Nkrp1d (GenBank accession no. AF324825; Ref. 27). In the absence of information about the Nkrp1 genomic region from other strains, including Sw or SJL mice, this issue has remained unresolved until now.
Our new data from the BALB/c genome, however, sheds light on the allelic nature of Nkrp1b and Nkrp1d: (1) only a single inhibitory-class cDNA and/or gene could be identified in BALB/c, similar to B6 and other strains analyzed to date; (2) the genomic location of the BALB/c Nkrp1b gene is similar to that of the B6 Nkrp1d gene (3); the Nkrp1bBALB and Nkrp1dB6 probes both hybridize to the same bands on genomic Southern blots; and (4) the Nkrp1bBALB gene shares a highly conserved e3-i3-e4 organization and sequence with the Nkrp1dB6 locus (Fig. 9). If Nkrp1b and Nkrp1d are indeed allelic, they show much greater allelic divergence than other Nkrp1 genes among B6, BALB/c, and other mouse strains (with exception of the Nkrp1c locus). The reason for this specific and directed divergence is unclear, but it is even more intriguing considering that the NKR-P1B/D receptors share a common ligand, the product of the Ocil/Clr-b locus, which is identical among the B6, BALB/c, and 129 strains (12, 13). Differential binding of NKR-P1B and NKR-P1D to a single allele of Ocil/Clr-b has been observed before in functional assays (12); however, this could reflect a differential affinity of each gene product, as a consequence of their divergent sequences, for a common ligand. In turn, such differential binding of NKR-P1 receptors to their Ocil/Clr ligands could have implications for the role of this system in missing self recognition and transplantation biology (12). Confirmation of such a role will have to await the elucidation of Ocil/Clr alleles and haplotypes, as well as functional analyses of the role of Ocil/Clr proteins in allotransplantation.
|
Nevertheless, before any final conclusions can be made concerning Nkrp1 (and Ocil/Clr) haplotype variation, more haplotypes will have to be characterized. The genomic organization and content of the BALB/c and B6 Nkrp1-Ocil/Clr regions are more similar than the Ly49 regions between these strains. If one assumes that the BALB/c and B6 Nkrp1-Ocil/Clr and Ly49 haplotype association is a natural one and not a consequence of the original derivation of inbred mice, then it would appear that the Nkrp1-Ocil/Clr region is evolving less rapidly than the Ly49 region. Although the reason for this remains speculative at present, a striking difference between the two NK cell receptor systems is that the Nkrp1 genes are genetically intermingled in the NKC with the Ocil/Clr genes that encode their ligands (12, 13); thus, both are coinherited. In contrast, the Ly49 genes segregate independently of the genes encoding their ligands, because the MHC is located on a different chromosome than the NKC (18). Coinheritance of ligand-encoding genes may negatively affect the rate of accumulated receptor gene mutations (i.e., receptor evolution), and vice-versa, because amino acid changes that decrease receptor-ligand binding affinities might be selected against. In contrast, the observed level of allelic divergence between Nkrp1b and Nkrp1d (and within the Nkrp1c locus) suggests that this may not always be the case. Further characterization of Ocil/Clr haplotypes is necessary to clarify the relationship between allelic divergence of the Nkrp1 and Ocil/Clr genes and how that influences their function. Intriguingly, the underlying basis for the evolution of numerous Nkrp1 genes in rodents and only a single gene (NKRP1A; Ref. 32) in humans is unknown. Yet like the Ly49 receptor system, this dichotomy is probably a functional consequence of differences in receptor usage by rodent vs human NK cells. Only a single Ocil/Clr-b-like gene exists in humans, designated LLT1 (18, 33). Indeed, the recent finding that human NKR-P1A–LLT1 interact and inhibit NK cell function demonstrates that the NKR-P1 missing-self recognition system is intact in humans (34, 35).
Identification of novel Nkrp1 and Ocil/Clr genes
Southern cross-hybridization and PCR cross-amplification of BAC clones and cDNA libraries identified sequences in the BALB/c genome that appear to represent novel Nkrp1 and Ocil/Clr genes. The first of these, designated Clr-h, was identified from BAC analysis using primers intended for Clr-b (cross-reactive to Clr-d). This new gene is not unique to the BALB/c genome; rather, the sequence is readily identifiable in the latest assembly of the public B6 genome database (99% identical across e3-i3-e4; GenBank accession no. XM487965). Moreover, the corresponding region in the B6 genome does not fit the location of any other known Clr gene (18). Such a high degree of sequence conservation suggests that Clr-h either represents a ligand for a highly conserved receptor, or that it is not subject to selection pressure and thus may represent a nonfunctional gene/pseudogene (at least in the strains tested; Fig. 5C), or encode a ligand for a nonfunctional receptor. In this light, it is interesting to note that the Nkrp1a sequence is highly conserved, the Nkrp1f sequence is identical between B6 (27) and BALB/c (this study), and the Nkrp1e sequence appears to be nonfunctional in both strains (this study, and Ref. 27). As the NKR-P1F receptor binds to Ocilrp2/Clr-g (13, 16, 17, 36, 37), it is likely that at least some of the remaining NKR-P1 receptors and Ocil/Clr ligands bind to one another. Although the Clr-c/d/e genes are not novel, this study elucidates their relative order within the NKC (i.e., Clr-g, Clr-d, Clr-e, Clr-c, and Nkrp1f)
A second new gene, Nkrp1g, is predicted to exist in the interval between Clr-f and Nkrp1c, based upon Southern cross-hybridization with an Nkrp1f cDNA probe. Although it is not known whether the sequence encodes a functional receptor, a sequence is currently available for this putative new gene in the B6 genome database (GenBank accession no. XM355818). The coding sequence bears some similarity to the NKR-P1F receptor in the extracellular region (16), but the predicted protein appears to lack known signaling motifs, including a cytoplasmic ITIM or Y residue, a putative CxCP recruitment motif for the Lck tyrosine kinase, and a charged transmembrane R residue (5, 38). Thus, full BAC sequencing may be required to assess the significance of the BALB/c cross-hybridization results, if no related cDNA can be isolated. This approach may also provide a clue to the origin and relationship of this sequence to the other Nkrp1 genes and/or pseudogenes. Although attempts to isolate an Nkrp1g cDNA were unsuccessful (at least in the strains tested; Fig. 5C), the Nkrp1g sequence appears to be conserved in the rat genome (Ensembl ID ENSRNOT00000035766; 
www.ensembl.org
), along with Nkrp1a (GenBank accession no. M62891), Nkrp1b (GenBank accession no. U56936), and Nkrp1f (GenBank accession no. X97477).
Conformational nature of the NK1.1 epitope
As mentioned previously, the NK1.1 epitope is specific to NKR-P1BSw/SJL and NKR-P1CB6 among known NKR-P1 proteins from both mice and rats (Fig. 7). However, sequence alignment of these NKR-P1 proteins does not provide a clue as to the context of the NK1.1-reactive sequence. Our identification of a substitution in the NKR-P1BBALB vs NKR-P1BSw/SJL sequence that confers NK1.1 reactivity by no means indicates that this amino acid alone is sufficient to generate the NK1.1 epitope. In contrast, both the rat NKR-P1A and NKR-P1B sequences contain the NK1.1-reactive S residue at the correct position, yet neither receptor bears the NK1.1 epitope. This is not surprising considering that NKR-P1B/C can be immunoprecipitated but not Western blotted using PK136 mAb (J.R.C., unpublished observations). This suggests that the NK1.1 epitope is conformational in nature, relying on additional context-dependent amino acid residues, and that denaturation of the three-dimensional structure of the epitope destroys its reactivity.
It was surprising that the seemingly conservative S191T substitution found in both NKR-P1BBALB and NKR-P1CBALB could abolish NK1.1 reactivity, whereas reversing this single amino acid substitution could confer NK1.1 reactivity (Fig. 7). The chemical nature of S and T residues is quite similar: both residues contain a hydroxymethyl group, yet the T residue has one additional methyl group. The implications of this are not entirely clear, but since both S and T residues can be posttranslationally modified by O-linked polysaccharides, it is possible that a specific O-linked glycosyltransferase may be partially responsible for either conferring or destroying the NK1.1 epitope on these receptors. Carbohydrate modification of the NKR-P1 receptors is not functionally required for ligand binding, because tetramers of NKR-P1 proteins produced in bacteria bind to their cognate ligands (12, 13); however, such modifications could alter the affinity of the receptor–ligand interaction, because bacterially produced NKR-P1 tetramers only bind their ligands weakly (12, 13). This is evident in the findings that NKR-P1D tetramers fail to visualize native Ocil/Clr-b ligand on normal cells ex vivo (13), and NKR-P1B tetramers bind more weakly to Ocil/Clr-b ligand than the -Ocil/Clr-b mAb, 4A6 (which is a low-affinity IgM) (12). Nonetheless, our present findings using the NKR-P1BBALB receptor in the BWZ reporter system recapitulate our in vitro and ex vivo results using an NKR-P1BSw/SJL receptor reporter cell (12). These results confirm that the cognate Ocil/Clr-b ligand recognized by these inhibitory receptors is expressed in a broad manner reminiscent of MHC class I (12), in contrast with the more restricted expression pattern suggested by other reports (13). The correlation between Ocil/Clr-b transcript expression (15, 16) and surface protein staining using -Ocil/Clr-b 4A6 mAb (12) seems to be in agreement with our current findings.
Thus, the NKR-P1B–Ocil/Clr-b missing self recognition system is intact in BALB/c mice, despite their well-known NK1.1– strain designation. The generation of mAbs to the BALB/c NKR-P1 proteins should facilitate direct analyses of receptor expression and function. In addition, further analysis of the known and novel NKR-P1–Ocil/Clr interactions in various mouse strains should elucidate the complex nature of this system of at least 6 potential receptor and 8 potential ligand genes linked to one another in the NKC; heterodimerization of these proteins could offer as yet unappreciated additional complexity to NK cell recognition of target cells.
Note added in proof. The genomic sequences and intron–exon structure of the Nkrp1 genes in this study have been modified from Ljutic et. al. (39) to conform to the Ly49 gene nomenclature (i.e., exon 2, cytoplasmic; exon 3, transmembrane, exon 4, stalk; exons 5–7, C-type lectin-like domain).
|
Acknowledgments |
|---|
|
Disclosures |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
|
Footnotes |
|---|
1 This work was supported by grants (to J.R.C. and A.P.M.) from the Canadian Institutes of Health Research (CIHR). J.R.C. is supported by a Career Development Award from the International Human Frontier Science Program Organization. A.M. was supported by a Life Sciences Award from the University of Toronto. A.P.M. is supported by a New Investigator Award from the CIHR. 
2 Address correspondence and reprint requests to Dr. James R. Carlyle, Department of Immunology, University of Toronto, Sunnybrook and Women’s Research Institute, 2075 Bayview Avenue (A-331), Toronto, ON, Canada M4N 3M5. E-mail address: james.carlyle{at}utoronto.ca or Dr. Andrew P. Makrigiannis, Laboratory of Molecular Immunology, Institut de Recherches Cliniques de Montréal, 110 Avenue des Pins Ouest, Montréal, QC, Canada H2W 1R7. E-mail address: makriga{at}ircm.qc.ca
3 Current address: Immunology Research CA, Sanofi-Pasteur, 1755 Steeles Avenue West, Toronto, ON, Canada M2R 3T4
4 Abbreviations used in this paper: NKC, NK gene complex; BAC, bacterial artificial chromosome; EC/UTR, extracellular/3'-untranslated region.
Received for publication December 21, 2005. Accepted for publication March 24, 2006.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
T cell differentiation in mouse fetal thymic ontogeny. J. Immunol. 160: 744-753. 2 integrin, very late antigen-2). J. Immunol. 167: 1141-1144. This article has been cited by other articles:
|
|
 |
|
  M. T. Norton, K. A. Fortner, P. Bizargity, and E. A. Bonney Pregnancy Alters the Proliferation and Apoptosis of Mouse Splenic Erythroid Lineage Cells and Leukocytes Biol Reprod, September 1, 2009; 81(3): 457 - 464. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. G. Aust, F. Gays, K. M. Mickiewicz, E. Buchanan, and C. G. Brooks The Expression and Function of the NKRP1 Receptor Family in C57BL/6 Mice J. Immunol., July 1, 2009; 183(1): 106 - 116. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Astrakhan, H. D. Ochs, and D. J. Rawlings Wiskott-Aldrich Syndrome Protein Is Required for Homeostasis and Function of Invariant NKT Cells J. Immunol., June 15, 2009; 182(12): 7370 - 7380. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
X. Yan, B. D. Johnson, and R. J. Orentas Induction of a VLA-2 (CD49b)-Expressing Effector T Cell Population by a Cell-Based Neuroblastoma Vaccine Expressing CD137L J. Immunol., October 1, 2008; 181(7): 4621 - 4631. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Walzer, M. Blery, J. Chaix, N. Fuseri, L. Chasson, S. H. Robbins, S. Jaeger, P. Andre, L. Gauthier, L. Daniel, et al. Identification, activation, and selective in vivo ablation of mouse NK cells via NKp46 PNAS, February 27, 2007; 104(9): 3384 - 3389. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
