Cloning of a Novel C-type Lectin Expressed by Murine Macrophages*

We report the cloning of a novel macrophage-restricted C-type lectin by differential display polymerase chain reaction. This molecule, named mouse macrophage C-type lectin, is a 219-amino acid, type II transmembrane protein with a single extracellular C-type lectin domain. Northern blot analysis indicates that it is expressed in cell lines and normal mouse tissues in a macrophage-restricted manner. The cDNA and genomic sequences of mouse macrophage C-type lectin indicate that it is related to the Group II animal C-type lectins. The mcl gene locus has been mapped between the genes for the interleukin-17 receptor and CD4 on mouse chromosome 6, the same chromosome as the mouse natural killer cell gene complex.

Differential Display PCR mRNA-Messenger RNA (mRNA) from 3 ϫ 10 6 cells from the cell lines listed above was harvested using the QuickPrep micro mRNA purification kit (Amersham Pharmacia Biotech).
cDNA-0.5 g of mRNA was resuspended in 12 l of sterile water, 2.5 mM oligo (dT) primer T 12 MA (M represents an equimolar mixture of A, C, G) and incubated at 70°C for 10 min. After snap-cooling on ice, the reaction was continued in 20 l containing 10  cycles of 94°C for 30 s, 40°C for 2 min, and 72°C for 30 s followed by a final extension step of 72°C for 5 min. 5 l of each reaction product were run on a 6% polyacrylamide-8 M urea gel. After drying, the gel was exposed for 16 h to Hyperfilm-HP (Amersham). Selected bands were excised from the gel, eluted in 100 l of sterile water, precipitated, and resuspended in 10 l of sterile water. 4 l of each PCR product were reamplified in a 40 l of reaction volume with conditions similar to the original PCR except for the use of 20 M dNTPs and 10 units of Taq polymerase. Reamplified PCR products were subcloned into the pGEM-T vector (Promega UK, Southampton, UK).

5Ј Rapid Amplification of cDNA Ends (5Ј-RACE)
Adapter-ligated cDNA, synthesized from 1 g of murine spleen poly(A) ϩ RNA (CLONTECH Laboratories UK Ltd., Basingstoke, UK) using the Marathon cDNA amplification kit (CLONTECH), was resuspended in sterile water. A 5Ј-RACE reaction was performed in a 50-l volume containing 50 mM Tris-HCl (pH 9.2 at 25°C), 16 mM (NH 4 ) 2 SO 4 , 2.25 mM MgCl 2 , 0.2 mM dNTP (Amersham), 1 l of DNA polymerase mix (a 20-l stock solution consisted of 14.3 l of Expand Long Template PCR system enzyme mix (Boehringer Mannheim) plus 5.7 l of TaqStart antibody (CLONTECH)), 200 nM Marathon adapter primer AP-1, 200 nM mMCL-specific primer complementary to residues 57-81 of the DD-PCR fragment (residues 856 -880 of the cDNA sequence reported here in Fig. 2A), and 5 l of adapter-ligated cDNA diluted 1:250 in sterile water. Reactions were incubated at 94°C for 3 min, followed by 35 cycles at 94°C for 45 s, 60°C for 45 s, and 72°C for 3 min. The resulting 871-base pair (bp) product was purified using a QIAquick PCR purification kit (Qiagen Ltd., Crawley, UK) subcloned into the pGEM-T vector and sequenced.

Isolation of DNA Clones
A cDNA library was constructed in the ZAPII vector (Stratagene Ltd., Cambridge, UK) using oligo(dT)-primed cDNA from the J774.2 cell line. A Sv/129 mouse liver genomic DNA library in the FixII vector was purchased from Stratagene. Approximately 1 ϫ 10 6 plaques from each library were screened with a [ 32 P]-labeled cDNA probe corresponding to the 871-bp RACE product. Positive plaques from the cDNA library were enriched after a further two rounds of screening, resulting in seven independent cDNA clones that were isolated in pBluescriptII-SK(Ϫ) (Stratagene). All seven cDNA clones were sequenced to obtain unambiguous overlapping readings from both strands. Positive plaques from the genomic library were enriched after three further rounds of screening. DNA from two independent clones was purified using the Wizard prep kit (Promega) and digested with NotI to release the full-length genomic DNA inserts. The inserts were then subcloned into NotI-digested pBluescript SK(Ϫ) and sequenced to obtain unambiguous overlapping readings from both strands.

Sequence Analysis
DNA sequence reactions were performed using the PRISM Ready Reaction DyeDeoxy Terminator sequencing kit (PE Applied Biosystems, Foster City, CA). Samples were subjected to electrophoresis on an ABI 373A DNA sequencer, read automatically, and recorded using ABI Prism Model Version 2.1.1 software (PE Applied Biosystems). Brookhaven Protein Data Bank, GenBank and EMBL data bases were searched for homologous sequences using the BLAST algorithm (12). Protein alignment, alignment consensus sequence, and percent identity were calculated by the Pileup, the Prettybox, and the Gap programs, respectively, included in the EGCG extensions to the Wisconsin Package Version 8.1.0, (13,14). A gap penalty value of 3.0 and a gap length weight of 0.1 were used with the Pileup and Gap programs.

Northern Blot Analysis
15 g (cell lines) or 20 g (tissues) of total RNA were subjected to electrophoresis through a denaturing 1.2% agarose, 6% formaldehyde gel and transferred to a Genescreen Plus nylon membrane (NEN Life Science Products). Equal loading of samples was confirmed by staining the gel with ethidium bromide. The filter was screened with a probe corresponding to the 5Ј-RACE product described above, washed with 1 ϫ SSC (0.15 M NaCl and 0.015 M sodium citrate), 0.1% SDS at 60°C for 1 h and exposed to Hyperfilm-MP (Amersham) at Ϫ70°C for 10 days (cell lines) or 21 days (tissues). For strongly positive tissue samples as shown in Fig. 1C, a shorter 4-h exposure was also performed. The cell lines used were J774.2, RAW 264.7, NS0, NIH-3T3, PG19, BW5147, WEHI-231, CTLL-2, J558L, and MEL 707. All cell lines were obtained from mycoplasma-free strains at the Sir William Dunn School of Pathology. The tissues used were bone marrow, brain, descending colon, heart, kidney, liver, lung, lymph node, muscle, small intestine, spleen, thymus, fetal liver, and resident peritoneal M. These tissues, excluding the fetal liver, were obtained from healthy Balb/c mice, 10 -12 weeks of age and bred and housed at the Sir William Dunn School of Pathology. Fetal tissue was obtained from day 14 Balb/c embryos.

Chromosome Localization
C3H/HeJ-gld and Mus spretus (Spain) mice and ((C3H/HeJ-gld ϫ Mus spretus)F 1 ϫ C3H/HeJ-gld) interspecific back-cross mice were bred and maintained as described previously (15). Mus spretus was chosen as the second parent in this cross because of the relative ease of detection of informative restriction fragment length variants in comparison with crosses using conventional inbred laboratory strains.
DNA isolated from mouse organs by standard techniques was digested with restriction endonucleases, and 10-g samples were electrophoresed in 0.9% agarose gels. DNA was transferred to Nytran membranes (Schleicher & Schull, Inc.), hybridized at 65°C with probes labeled by random-primed method with [ 32 P]dCTP, and washed under stringent conditions, all as described previously (16). Gene linkage was determined by segregation analysis. Gene order was determined by analyzing all haplotypes and minimizing crossover frequency between all genes that were determined to be within a linkage group. This method resulted in determination of the most likely gene order (17).  (52). C, diagrammatic representation of the deduced mMCL protein structure in its monomeric form. C-type lectin (CL) and N-linked glycosylation sites (lollipops schematics) are shown. UK) and precipitated by slowly adding it to an equal volume of 2 ϫ HBS (1.64% w/v NaCl (BDH), 1.18% HEPES (free acid) (Sigma), 0.04% Na 2 HPO 4 (anhydrous) (BDH), pH 7.05). The precipitate was added to 1 ϫ 10 6 CHO.K1 cells cultured in an 80-cm 2 flask in medium and supplements as described above. After 3.5 h, the medium was removed, and the cells were treated with 2 ml of 15% glycerol, 1 ϫ HBS for 2 min at 37°C. Cells were allowed to recover overnight in fresh medium. Geneticin, 2 mg/ml, was added the following morning.

Transfection of CHO.K1 Cells
Enrichment-FLAG-tagged mMCL-expressing cells were enriched 1 week after transfection by selection with the Anti-FLAG M2 (Sigma) monoclonal antibody (mAb) and sheep anti-mouse IgG magnetic Dynabeads (Dynal (UK) Ltd., Wirral Merseyside, UK). Transfected cells were stripped with 0.5 mM EDTA, washed in PBS, 0.5% bovine serum albumin (BSA) (Sigma), and incubated on ice for 1 h in 20 g/ml M2 mAb in PBS/BSA. After further washing in PBS/BSA, the cells were incubated on ice for 40 min in 200 l of PBS/BSA and 3 l of Dynabeads. Cells selected by magnetic attraction during subsequent extensive PBS/BSA washes were allowed to recover overnight in a 25-cm 2 flask in geneticinfree medium. The medium and detached beads were removed the following morning and replaced with fresh medium containing 2 mg/ml geneticin. Two rounds of limiting dilution cloning and subsequent fluorescence-activated cell sorting (FACS) analysis with the anti-FLAG M2 mAb yielded a stable FLAG-tagged, mMCL-expressing clone.

Rabbit Antiserum
200 g of keyhole limpet hemocyanin-conjugated peptide corresponding to amino acids 2-16 of mMCL (H 2 N-WLEESQMKSKGTRHP-COOH) (Multiple Sclerosis Peptide Laboratory, Oxford Brookes University) were diluted in 500 l of PBS, combined with an equal volume of Freund's complete adjuvant (Sigma), and injected subcutaneously into adult New Zealand White rabbits. Booster injections were given as above, except using Freund's incomplete adjuvant (Sigma), at weeks 3, 5, and 10. Pre-immune serum was designated as sample 16P. A test bleed was taken at 8 weeks to check for anti-peptide antibody activity, and a final bleed (50 ml) was collected 3 weeks after the final injection. 5 ml of the serum were absorbed with 4 ϫ 10 7 CHO.K1 cells, which had been fixed with 4% paraformaldehyde (BDH) and permeabilized with 0.2% v/v Triton X-100 (Sigma). This absorbed serum was designated rabbit antiserum 16T.

Immunofluorescence
Transfected and wild-type CHO.K1 cells were grown on glass coverslips, fixed with 4% paraformaldehyde in PBS at 4°C for 1 h, quenched in 10% fetal calf serum for 10 min, washed, blocked in PBS, 15% normal goat serum for 30 min, and then stained for 1 h at 4°C with either 10 g/ml anti-FLAG M2 mAb (Sigma) or a 1:300 dilution of either rabbit serum 16P or 16T in PBS, 15% normal goat serum. After additional washes, cells were incubated at 4°C for 1 h with a 1:300 dilution of fluorescein isothiocyanate-conjugated goat anti-mouse IgG (Chemicon International, Inc., Temecula, CA) or goat anti-rabbit IgG (Sigma), then analyzed by fluorescent microscopy or FACS. Cells analyzed by FACS were detached from the culture flasks with 0.5 mM EDTA before fixation. To permeabilize cells, 0.2% v/v Triton X-100 was added to all PBS, 15% normal goat serum solutions.

Western blot Analysis
Cells were lysed at 4°C in a solution of 1% v/v Nonidet P-40 (Sigma), 150 mM NaCl, 10 mM EDTA, 10 mM NaN 3 , 10 mM Tris-HCl, pH 8, 1 mM phenylmethylsulfonyl fluoride, 5 mM iodoacetamide. Lysates were cleared by centrifugation at 12,000 ϫ g for 30 min at 4°C. Samples were boiled for 3 min in 4% SDS sample buffer and subjected to SDSpolyacrylamide gel electrophoresis. Samples were reduced by boiling FIG. 3. Alignment of the CRDs of five known C-type lectins and mMCL. rMBL-A (accession number: P19999) represents a prototypical C-type lectin CRD. mKupffR (D88577), hML (D50532), hGP120BL (A46274), and cHeplec (chicken hepatic lectin) (P02707) are the four proteins that have the highest protein sequence similarity to mMCL in order of increasing similarity. Each sequence begins at the residue that aligns with the first amino acid of the exon encoding the start of the proposed mMCL C-type lectin domain. Residues are blocked to show similarity to mMCL. Black boxes indicate residues identical to the mMCL residue at that position. Gray boxes indicate residues that frequently occur instead of the mMCL residue as determined by the Prettybox program in the Wisconsin Package (13,14). CL domain shows the C-type lectin domain signature as described (22,26). Invariant residues are represented with single-letter amino codes (Z ϭ E or Q). Residues of a conserved nature are labeled ϭ aromatic, ϭ aliphatic, ⍀ ϭ either aromatic or aliphatic, O ϭ oxygen containing. Secondary structure (2 o ) and calcium binding are shown as determined by Weis for rMBL-A (22). All residues that interact with calcium ions 1 and 2 in rMBL-A are highlighted with arrowheads. Open arrowheads indicate residues that are also crucial for sugar binding in rMBL-A. them in a sample buffer containing a final concentration of 5% 2-␤mercaptoethanol (Sigma) before electrophoresis. Proteins were transferred to Hybond-C extra nitrocellulose membrane (Amersham). The membrane was treated with blocking solution (5% milk protein, 0.1% v/v Tween 20, 15% normal goat serum) and blotted for 1 h in blocking solution containing 10 g/ml anti-FLAG M2 or a 1:500 dilution of either rabbit serum 16P or 16T. After washing with blocking solution, a second 1-h incubation was performed, again in blocking solution, using a 1:1000 dilution of goat anti-mouse IgG (Sigma) or donkey anti-rabbit IgG (Chemicon) coupled to horseradish peroxidase. The bound antibody was detected by enhanced chemiluminescence (Amersham). For competition assays, saturating amounts of either mMCL peptide or FLAG peptide was combined with the primary antibody before its application to the blot.

RESULTS AND DISCUSSION
DD-PCR analysis was performed on cDNA from the following mouse cell lines: J774.2 (M), RAW 264.7 (M), EL-4 (thymoma), NS0 (myeloma), and L929 (fibroblast). This panel of cell lines was chosen to provide a range of immune and nonimmune cell types as well as two independent M and three non-M internal controls to facilitate the selection of M-restricted genes. DD-PCR products amplified from both M cell lines but absent from the non-M cell lines (Fig. 1A) were eluted, reamplified, subcloned into the pGEM-T vector, and sequenced. A 272-bp cDNA fragment amplified exclusively from both M cDNA samples was selected for further study because it contained a putative polyadenylation signal site downstream of a TGA stop codon, indicating that it might be a segment of a functional gene. Subsequently, 5Ј-RACE-PCR was performed on adapter-ligated mouse spleen cDNA using adapter-specific sense and DD-PCR fragment-specific antisense primers. This reaction generated an 871-bp product.
Sequence analysis of the 5Ј-RACE product showed that it included 601 bp of novel sequence and 257 bp of sequence corresponding to the amplified portion of the original DD-PCR fragment. The novel sequence included a second in-frame TAA stop codon 678 bp upstream of the original DD-PCR stop codon, suggesting that the 5Ј-RACE product contained a complete open reading frame. Downstream of the TAA stop codon found in the 5Ј-untranslated region, two potential in-frame start codons were found 20 bp apart. Although neither of these codons lies in a perfect context for translation initiation, the codon at position 161-163 is believed to be the translation start site. It is the first initiation codon downstream of an in-frame stop codon and the only initiation codon with an adenosine in the Ϫ3 position (18,19).
As a means of testing the expression specificity of this putative gene, Northern blot analysis was performed on an expanded range of 11 mouse cell lines including CTLL-2 (interleukin-2-dependent T cell), J558L (myeloma), BW5147 (thymoma), WEHI-231 (B cell lymphoma), PG19 (melanoma), NS0, RAW 264.7, L929, MEL 707 (erythroleukemia), J774.2, and NIH-3T3 (fibroblast). Total RNA from these cell lines probed with the 32 P-labeled 5Ј-RACE product showed a pair of 1.1-kb and 700-bp bands solely in the J774.2 and RAW 264.7 cell lines (Fig. 1B) and confirmed that the gene was expressed in a M-restricted manner. Using gene-specific sense and an- tisense primers, a more sensitive reverse transcription-PCR assay of cDNA from J774.2, NS0, L929, EL-4, RAW 264.7, MEL 707, and P388.1D (M) cell lines was also conducted. Again, specific amplified bands were found only in the three M cell lines (data not shown). To determine whether the same Mrestricted expression pattern exists in vivo, Northern blot analysis was conducted on RNA isolated from 12 different normal mouse tissues. The 871-bp 5Ј-RACE probe recognized a 1.1-kb transcript in resident peritoneal M Ͼ Ͼ bone marrow Ͼ Ͼ spleen ϭ lung Ͼ Ͼ lymph nodes (Fig. 1C). The intensity and distribution pattern of the bands were in accordance with known M populations and supported the evidence of Mrestricted expression as seen in the cell lines. In particular, the strong expression in bone marrow correlated well with previous studies, which showed that mice bone marrow is the richest source of M as determined by the M-restricted marker F4/80 (20). The presence of the 1.1-kb band alone (tissues) or at a higher intensity (cell lines) on both Northern blots strongly suggests that the 1.1-kb transcript is the predominant form of the modified mRNA in tissue M.
The possibility that alternatively spliced transcripts might account for the 700-bp band, which appeared only in the murine M cell lines, was investigated by screening a J774.2 cDNA library with the 5Ј-RACE probe. Sequence analysis of seven independent clones isolated from the library revealed no alternatively spliced variants. Two of the clones were found to lack three consecutive base pairs (390 -392), leading to an in-frame deletion of Gly-77. Three of the seven clones were also found to have matching 3Ј-untranslated regions, which extended an additional 74 bp past the site of the start of the poly(A) tail in the original DD-PCR clone. These three longer clones included a second, rarer AAUAUA polyadenylation signal sequence (21) 25 bp upstream of the start of their poly(A) tails. The physiological abundance and importance of the Gly-77 deletion and the extended 3Ј-untranslated sequence have yet to be determined.
A 918-bp cDNA consensus sequence, including Gly-77, was constructed from the overlapping regions of each cDNA clone using a minimum of two independent cDNA clones to confirm each nucleotide. In total, the consensus sequence consists of 160 bp of 5Ј noncoding sequence, a 660-bp open reading frame, and 98 bp of 3Ј-untranslated sequence ( Fig. 2A). No discrepancies were found between the open reading frame of this consensus sequence and that of the 5Ј-RACE PCR product amplified from mouse spleen, thereby confirming that the gene expressed in the cell lines was the same gene expressed in normal mouse tissue and the same gene that was amplified in he original DD-PCR.
Translation of the open reading frame of the cDNA beginning at the 161-163 start codon yielded a deduced 219-amino acid protein sequence (Fig. 2A). The lack of an identifiable signal peptide and a hydropathy profile displaying a hydrophobic anchor sequence near the amino terminus (Fig. 2B) suggested that the gene encoded a type II integral membrane protein. The complete protein sequence was unique insofar as it displayed no overall sequence homology or identity to any other protein sequence entered in a variety of protein data bases. The final 130 carboxyl-terminal amino acids showed similarity to a wide range of carbohydrate-binding proteins, namely C-type lectins, with the greatest degree of homology to chicken hepatic lectin. The carboxyl terminus of the putative protein sequence was aligned with the carbohydrate recognition domains (CRDs) of four of the lectins with which it had the highest similarity scores: mouse Kupffer cell fucose receptor, human macrophage lectin, human gp120 binding lectin, and chicken hepatic lectin. The sequence of rat mannose binding lectin (rMBL-A), although apparently more distantly related to the novel protein, was included because it has an extensively studied C-type lectin CRD (22)(23)(24)(25). The alignment showed that the putative sequence is 29, 37, 36, 39, and 31 percent identical to the CRDs of these proteins, respectively. In addition it indicated that apart from Gly-158, Pro-173, Gly-191, and Arg-204, the novel protein shares 11 of the 14 invariant and 17 of the 18 highly conserved amino acids used to define C-type lectins (26) (Fig.  3). Sequence data from two independent clones isolated from a FixII mouse liver genomic DNA library screened with the 5Ј-RACE probe also revealed that the protein is encoded within six exons. As is characteristic of Group-II C-type lectins, the region corresponding to its CRD is encoded by three exons, and the amino-terminal cytoplasmic tail and anchor sequence are encoded by two exons (26). The final two introns of this protein precisely match the position of introns found in the CRDs of other Group-II C-type lectins: CD23 (27), the major form of the rat asialoglycoprotein receptor (28), the Kupffer cell fucose receptor (29), and chicken hepatic lectin (30). The number, position, and phasing of all the intron/exon splice sites of the putative protein are analogous to those of chicken hepatic lectin (30). Collectively, this evidence suggested that the novel protein is indeed a C-type lectin. It has a predicted structure of a 20-amino acid amino-terminal cytoplasmic domain, a 20amino acid transmembrane domain attached to an extracellular region comprising a 49-amino acid stalk and a 130-amino acid C-type lectin domain with two potential N-glycosylation sites (Fig. 2C). This protein was named murine macrophagerestricted C-type lectin (mMCL).
Comparative sequence analysis suggests that mMCL has carbohydrate binding capabilities. Among the 11 conserved amino acids mentioned above, mMCL contains the nine amino acid residues that have been identified as essential for calcium cation association and ligand binding within the rMBL-A CRD (Fig. 3) (22-24, 31). However, little can be postulated about the binding ability or specificity of mMCL from its protein sequence alone. Even though much progress has been made in elucidating the three-dimensional structure and binding characteristics of C-type lectin domains (32,33), C-type lectinligand interactions are still not very well understood (34,35). Even within such a relatively small, conserved domain, binding specificity can be altered with the mutation of only one (36) or two amino acids (23). Some molecules containing C-type lectin domains have also been shown to bind peptide sequences (37). Such versatility makes predicting putative ligands for this type of lectin domain difficult at best. For mMCL, this task is even more challenging because a serine rather than the typically conserved proline separates the two critical sugar binding residues corresponding to Glu-185 and Asn-187 in rMBL-A. The proline usually found at this position introduces a tight turn in the loop structure and may play a role in spacing the adjacent amino acids around the binding site. Disruption of that turn by a smaller, more flexible, more hydrophilic amino acid such as the serine could therefore feasibly alter or destroy this site and its binding specificity. mMCL exhibits the highest protein sequence similarity to the members of Group II C-type animal lectins, a diverse set of type-II transmembrane receptors, including rat, mouse, and chicken hepatic lectins (38). Generally these proteins are thought to mediate glycoprotein endocytosis and degradation. mMCL might therefore possess a similar function. However, the characterization of the tumor binding capabilities of the M-specific Group II lectin mouse macrophage galactose/Nacetylgalactosamine-specific lectin (39 -41) might imply that mMCL performs a comparable immune surveillance role. The M-restricted expression and tissue distribution of mMCL, particularly in bone marrow, similarly suggests a hemopoietic function for this protein. Likewise, considering its lectin structure, mMCL may play a role in cell-cell recognition (42).
To determine whether this novel M-restricted gene co-localized with other known lectins or M-specific genes, the chromosomal localization of the mouse mcl gene was investigated. A panel of DNA samples from an interspecific cross that has been characterized for over 1000 genetic markers throughout the genome was analyzed. The genetic markers included in this map span between 50 and 80 centimorgans on each mouse autosome and the X chromosome. 2 Initially, DNA from the two parental mice (C3H/HeJ-gld (C3H/HeJ-gld ϫ Mus spretus)F 1 ) were digested with various restriction endonucleases and hybridized with a mMCL cDNA probe to determine restriction fragment length variants to allow haplotype analyses. Informative MspI restriction fragment length variants were detected: C3H/HeJ-gld, 7.0 kb, 3.8 kb; Mus spretus, 8.6 kb. Comparison of the haplotype distribution of the mcl restriction fragment length variants and those previously defined in this interspecific cross indicated that this locus co-segregated in 112 of the 114 meiotic events examined with the CD4 and CD9 loci on mouse chromosome 6. The haplotype distribution among the other genes localized to mouse chromosome 6 is shown in Fig.  4. The best gene order Ϯ the S.D. indicated the gene order: (centromere) Il-17r-4.4 centimorgan Ϯ 1.9 centimorgan-mcl-1.8 centimorgan Ϯ 1.2 centimorgan-Cd9/Cd4. Intriguingly, these studies place the mcl gene locus just proximal of the nkrp1 cluster of the natural killer cell gene complex (NKC), in close proximity to the NKC linkage group (hcph/cd4/lag3)-a2 m-cd69-prp (43). The NKC contains the FIG. 7. A, Western blot analysis of cell lysates from mMCL-FLAG expressing CHO.K1 cells. Equal amounts of protein from cell lysates of wild-type CHO.K1 cells and pCDNA3/mMCL/FLAG-transfected CHO.K1 cells were run on a 15% nonreducing, polyacrylamide gel and transferred to Hybond C-extra membrane. Portions of the same membrane were incubated with either a mouse anti-FLAG mAb (␣-FLAG), rabbit antiserum (16T), or rabbit preimmune serum (16P) in the absence of peptide or in the presence of saturating amounts of FLAG peptide or saturating amounts of the mMCL peptide fragment used to generate the rabbit antiserum (lanes labeled No peptide, ϩFLAG Peptide, and ϩmMCL Peptide, respectively) followed by incubation with either a peroxidase-conjugated goat anti-mouse IgG or donkey anti-rabbit IgG secondary reagent. Visualization of the bands was by chemiluminescence. Lanes marked W and M identify wild-type and mMCL-transfected cell lysates, respectively. Arrows indicate the position of the monomer as stained by the ␣-FLAG mAb and the 16T antibody. The monomer doublet is visible in the gel stained with the ␣-FLAG mAb just under a much darker nonspecific band that appears to be up-regulated in the transfected cells. The same doublet can be seen using the 16T antibody but only when less protein is loaded on the gel. All unmarked lanes were loaded with molecular mass markers. B, Western blot analysis of the same cell lysates as A run on a 15% polyacrylamide gel under reducing and nonreducing conditions. The proteins were transferred as above and stained with the 16T antibody. Visualization of the bands was by chemiluminescence. All unmarked lanes were empty.
genes encoding numerous C-type lectins, including Ly-49 family members involved in "missing self" recognition and natural killer cell inhibition, CD161 homologues involved in NK cell activation, and CD69, a lymphoid activation marker (44,45). Each of these lectins is a type II transmembrane, disulfidelinked homodimer approximately 200 -300 amino acids in size (46 -49). Together they belong to the animal C-type lectin Group V and are evolutionarily distinct from the Group II lectins that they resemble structurally (26). This gross structural resemblance between Group V and Group II lectins and the propinquity of the NKC and mcl gene loci might suggest a similar self-recognition function for mMCL in M biology. However, excluding the core C-type lectin-conserved residues, mMCL bears very little sequence similarity to any of the lectins found on natural killer cells and other lymphocytes. mMCL lacks, for example, the distinctive cytoplasmic immunoreceptor tyrosine-based inhibitory motifs (50) and activation motifs (51) found in the proteins encoded within the NKC.
Concurrent with the sequence and genomic analysis of this molecule, biological investigations of mMCL were also undertaken in order to study its ligand binding, distribution characteristics, and possible functions. The protein was expressed in the presence of microsomes using the rabbit reticulocyte lysate system to confirm its nature as an integral membrane protein.
A single specific product approximately 30 kDa in size was transcribed (data not shown). Multiple, ultimately unsuccessful attempts were made to generate CHO.K1 and COS-7 cells transiently expressing various soluble and membrane forms of the protein. Eventually a stable line of CHO.K1 cells expressing a carboxyl-terminal FLAG-tagged full-length version of this protein was isolated. Staining of permeabilized and nonpermeabilized transfected cells with the anti-FLAG M2 mAb revealed variable surface expression of mMCL (Fig. 5, A and C). The majority of the overexpressed protein appeared to be trapped within the endoplasmic reticulum of the cells (Fig. 5B). A rabbit antiserum (rabbit 16T) raised against the cytoplasmic tail of mMCL showed weaker but specific staining in mMCL-transfected permeabilized cells but not wild-type permeabilized CHO.K1 cells. mMCL could be detected by FACS analysis of transfected versus wild-type permeabilized CHO.K1 cells using either the M2 mAb or the rabbit antiserum (Fig. 6).
Western blot analysis of cell lysates from mMCL-transfected CHO.K1 cells using either the anti-FLAG M2 mAb or the rabbit 16T serum showed a similar pattern of bands at 30, 60, and 90 kDa and a higher smear running to the top of the gel (Fig. 7A). Since the mMCL polypeptide sequence has a predicted molecular mass of 25.6 kDa, the 30-kDa band probably corresponds to a monomeric form of the protein. The 60-and 90-kDa bands and the higher smear are thought therefore to correspond to the dimer, trimer, and higher multimer forms of mMCL, respectively. Under reducing conditions, the multiple bands collapse to a single band that runs at approximately 35 kDa, indicating the presence of both intra-and intermolecular disulfide bonds (Fig. 7B). After treatment with N-glycanase, the reduced protein runs at approximately 30 kDa again, verifying that at least one of the predicted N-glycosylation sites is utilized (data not shown).
The 30-kDa band can be resolved into a doublet under nonreducing conditions (Fig. 7A). However, cell lysates from mouse spleen, resident peritoneal M, Bio-Gel elicited, and thioglycollate broth-elicited peritoneal M only display a single specific 30-kDa band when stained with the 16T rabbit anti-mMCL polyclonal antibody (data not shown). The absence of the doublet, the 60-kDa dimer, the 90-kDa trimer, and the higher molecular mass species in primary tissue lysates and rabbit reticulocyte system lysates suggests that these forms might be artifacts of overexpression or misfolding of the recombinant protein. The possible formation of higher order oligomers and misfolded monomers may also explain why so much of the expressed protein is retained within the endoplasmic reticulum of the stably transfected CHO.K1 cells rather than being transported to the cell surface.
The low and variable surface expression and possible misfolding of the recombinant form of the mMCL molecule have hitherto prevented the identification and study of its ligand and possible function. Current studies are under way to isolate a mMCL-specific mAb and a functional form of recombinant mMCL protein. Additional investigations into the nature of this lectin have the potential of furthering our understanding of this relatively new and growing field of cell-surface proteins. It may also increase our knowledge of the activities M perform, how these cells are regulated, and how they regulate the cells around them.