Molecular cloning of F4/80, a murine macrophage-restricted cell surface glycoprotein with homology to the G-protein-linked transmembrane 7 hormone receptor family.

F4/80 is a monoclonal antibody that recognizes a murine macrophage-restricted cell surface glycoprotein and has been extensively used to characterize macrophage populations in a wide range of immunological studies. Apart from the tightly regulated pattern of expression of the F4/80 antigen, little is known about its possible role in macrophage differentiation and function. We have sought to characterize the molecule at the molecular level, through the isolation of cDNA clones, and now describe the sequence of the F4/80 protein. The primary amino acid sequence demonstrates homology to two protein superfamilies. The NH2-terminal region consists of seven epidermal growth factor-like domains, separated by approximately 300 amino acids from a COOH-terminal region that shows homology to members of the seven transmembrane-spanning family of hormone receptors. The potential role of these distinct domains is discussed with respect to the possible function of the F4/80 molecule.

F4/80 is a monoclonal antibody that recognizes a murine macrophage-restricted cell surface glycoprotein and has been extensively used to characterize macrophage populations in a wide range of immunological studies. Apart from the tightly regulated pattern of expression of the F4/80 antigen, little is known about its possible role in macrophage differentiation and function. We have sought to characterize the molecule at the molecular level, through the isolation of cDNA clones, and now describe the sequence of the F4/80 protein. The primary amino acid sequence demonstrates homology to two protein superfamilies. The NH 2 -terminal region consists of seven epidermal growth factor-like domains, separated by approximately 300 amino acids from a COOH-terminal region that shows homology to members of the seven transmembrane-spanning family of hormone receptors. The potential role of these distinct domains is discussed with respect to the possible function of the F4/80 molecule.
Macrophages play a crucial role in the initiation and effector stages of both innate and adaptive immune responses. A number of specialized cell surface molecules expressed by macrophages participate in these responses, such as macrophage mannose receptor, macrophage scavenger receptor, complement receptor 3 (CD11b/CD18), and opsonic Fc receptors for antigen-specific immunoglobulins. One of the most highly restricted macrophage membrane molecules is defined by the monoclonal antibody F4/80, which recognizes a 160-kDa glycoprotein on the surface of most mouse macrophage populations (1). The F4/80 mAb 1 has been extensively used in a range of immunohistochemical studies of development in normal mice and in a number of pathologic models as a macrophage-specific marker (2)(3)(4)(5)(6). To date, no defined function has been ascribed to F4/80, although its expression is known to be down-regulated by interferon-␥ and in response to Bacille Calmette-Guérin infection (7,8), as well as being absent from macrophages localized within T cell areas of lymph nodes and spleen. This further suggests that T cells, or a T cell-derived product, may play a role in the regulation of F4/80 expression. The lack of F4/80 expression on migrating veiled cells, derived from F4/80 ϩ Langerhans cells within the skin epidermis, and the low level of expression on blood monocytes (6,9) would suggest that the molecule is in some way involved in cell adhesion within certain tissues. As a means of delineating the physiologic role of F4/80 on mature macrophages, we have successfully isolated F4/80 cDNA clones. Herein we report that the primary amino acid sequence of F4/80 shows a degree of homology to two protein superfamilies: the extracellular NH 2 -terminal region contains seven EGF-like repeats, while the final third of the molecule shows homology to members of the seven transmembrane (Tm7) hormone receptor family.
Peptide Sequencing and cDNA Cloning-F4/80 antigen was purified by Triton X-100 lysis from the J774.2 cell line, grown as a tumor in the peritoneal cavity of BALB/c mice. The pelleted tumor mass was homogenized on a Polytron tissue grinder in a solution consisting of 10 mM Tris-HCl, pH 7.4, 10 mM NaCl, 3 mM MgCl 2 , 3 mM iodoacetamide, 1 mM phenylmethylsulfonyl fluoride, and 5 g/ml pepstatin. The homogenate was centrifuged at 700 ϫ g for 7 min to remove the nuclear pellet, which was discarded, and at 100,000 ϫ g for 30 min to collect the membranerich fraction. The sample was then extracted with lysis buffer containing 1% Triton X-100, 10 mM EDTA, 5 mM NaN 3 , and the protease inhibitors at 4°C for 30 min, before spinning at 100,000 ϫ g for 30 min. The supernatant was collected and stored at Ϫ70°C. Before use, the extracts were thawed, and, after the addition of 1 mM diisopropyl fluorophosphate, they were precleared at 100,000 ϫ g for 30 min before passing sequentially through a PD-10 column (Pharmacia Biotech Inc.) and a bovine IgG-Sepharose CL4B precolumn followed by a 5C1-Sepharose CL4B affinity column equilibrated with lysis buffer. After washing, the 5C1 column was eluted with 50 mM diethylamine, pH 11.5, and fractions containing the majority of the protein, as assessed by absorbance at 280 nm, were neutralized with 2 M glycine, pH 2.0, and pooled. The sample was passed over a second 5C1 affinity column, washed, and eluted as before, and the eluate was concentrated over a 100-kDa cut-off Spectra/Por membrane (Pierce and Warriner, Chester, UK), filtered through a 0.2-m filter, and stored at 4°C. Minor contaminants were removed by preparative SDS-polyacrylamide gel electrophoresis incorporating excision of the antigen band and electroelution. On analytical SDS-polyacrylamide gel electrophoresis, the final antigen preparation * This work was supported by the Medical Research Council, United Kingdom. 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.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EMBL Data Bank with accession number(s) X93328.
‡ To whom correspondence should be addressed. § Present address: Instituto di Patologia Generale, Universita Degli Studi di Trieste, 34127 Trieste, Italy. 1 The abbreviations used are: mAb, monoclonal antibody; PCR, polymerase chain reaction; bp, base pair(s); EGF, epidermal growth factor. displayed a major homogeneous band, and a minor one possibly corresponding to the precursor form of the molecule. The antigen was reduced, alkylated, and digested with trypsin, from which peptides were subjected to amino acid sequence analysis by gas-phase sequencing. A cDNA library was constructed in the ZAPII vector (Stratagene Ltd., Cambridge, UK) using oligo(dT)-primed cDNA from the J774.2 cell line. Approximately 1 ϫ 10 6 plaques were screened with rabbit anti-mouse F4/80 polyclonal antiserum and detected with alkaline phosphataseconjugated goat anti-rabbit IgG. Positive plaques were enriched following a further two rounds of screening, resulting in 10 independent clones that were isolated in pBluescriptII-SK(Ϫ). The clone containing the largest insert, pF4/80(12.2), was sequenced to obtain unambiguous overlapping readings from both strands. 5Ј Rapid amplification of cDNA ends nested PCR using oligonucleotide-anchored spleen cDNA template, in the presence of an anchor-specific primer and (primary PCR) an antisense primer complementary to residues 200 -171 of pF4/80(12.2), followed by (secondary PCR) an antisense primer complementary to residues 146 -120 of pF4/80 (12.2), was performed to characterize the 5Ј region of the F4/80 cDNA. The PCR conditions were as follows: 50-l final reaction volume containing 2 mM MgCl 2 , 200 M dNTP, 10 M each primer, 2.5 units of Taq DNA polymerase (Promega, Madison WI), 2 l of anchored cDNA (primary PCR), or 2 l of 10-fold diluted primary amplified product (secondary PCR), using the following cycling parameters 94°C/45 s, 60°C/45 s, 72°C/2 min for a total of 30 cycles. A 230-bp fragment was amplified, which overlapped clone 12.2 and included novel sequence encoding the initiator methionine and amino acid residues 1-4.
Northern Blot Analysis-15 g of total RNA from a range of mouse cell lines (listed above) was subjected to electrophoresis through a denaturing 1.2% agarose gel and transferred to a Genescreen Plus membrane (DuPont NEN). The filter was probed with a 32 P-labeled ClaI-ApaI cDNA fragment (757 bp) excised from clone pF4/80(12.2), washed with 1 ϫ SSC, 1% SDS at 60°C and exposed to Hyperfilm-MP (Amersham Corp.) at Ϫ70°C.
Transfection of CHO-K1 Cells-A PCR fragment encoding the fulllength F4/80 open reading frame was amplified, with clone pF4/80(12.2) as a template, using the following oligonucleotide primers: 5Ј TAG TAG AAG CTT AGT ACG ATG TGG GGC TTT TGG CTG CTC CTC TTC TGG GGC TTC AG 3Ј, which contains sequence (in boldface) corresponding to the first 13 amino acids of the F4/80 precursor protein as well as a HindIII restriction enzyme site (underlined), and 5Ј TAG TAG TCT AGA GAA AGG ATG TTA ACC CAT CTT GGA AGT GG 3Ј, which contains sequence (in boldface/antisense) located at the end of the F4/80 cDNA open reading frame, including the translation termination codon, as well as an XbaI restriction enzyme site (underlined). PCR conditions were as follows: 100-l reaction volume containing 1 mM MgSO 4 , 200 M dNTP, 50 M each primer, 3 units of Vent® DNA polymerase (New England Biolabs, Beverly MA), 100 ng of pF4/80(12.2) using the following cycling parameters 94°C/1 min, 55°C/1 min, 75°C/3 min for a total of 35 cycles. The amplified fragment was subcloned into HindIII-XbaI-digested pcDNA3 (Invitrogen, San Diego, CA), and the construct was stably transfected into CHO-K1 cells by the calcium phosphate precipitation technique (11). Following transfection, cells were dispensed into 96-well flat bottomed microtitre plates, and clones selected with 250 g/ml Geneticin (Life Technologies, Inc.). A control construct, pcDNA3/␤-galactosidase, was kindly provided by Dr. David Greaves, Sir William Dunn School of Pathology.

RESULTS AND DISCUSSION
The macrophage cell line J774.2 has been shown previously to express high levels of the F4/80 antigen (1,10), and, as such, an unamplified cDNA library was constructed with poly(A) ϩ RNA from this source in the ZAPII vector and screened using rabbit polyclonal antisera raised against purified F4/80 antigen. Following a further two rounds of screening, 10 positive clones were isolated and characterized by restriction enzyme analysis and sequencing. None of the clones appeared to contain a full-length cDNA coding sequence, as witnessed by the lack of an in-frame ATG at their 5Ј end. A 5Ј rapid amplification of cDNA ends PCR strategy was therefore applied, utilizing antisense oligonucleotide primers specific for internal sequence within clone pF4/80(12.2) to amplify a cDNA fragment that overlapped pF4/80(12.2) and encoded 46 bp of novel 5Ј sequence including the ATG initiation codon. The overall composite sequence consists of 3286 bp, including a stretch of 49 adenosines corresponding to the poly(A) tail and an upstream AATAAA motif. The open reading frame encodes a precursor of 931 amino acids with a predicted signal peptide of 27 residues (12), resulting in a mature protein of 904 amino acids with a predicted mass of 98.9 kDa (Fig. 1A). On SDS-polyacrylamide gel electrophoresis analysis, F4/80 appears as a smear of approximately 160 kDa, which, together with the mass differential from the deduced sequence, suggests extensive glycosylation of the molecule. The presence of 10 potential N-glycosylation sites (Asn-X-Ser/Thr) and a region highly rich in Ser and Thr residues (Fig. 1A, amino acids 399 -642) suggests that the protein is heavily N-and O-glycosylated in agreement with earlier studies. The predicted amino acid sequence also contained four tryptic peptides obtained from the purified protein, ranging from 64 to 96% similarity (Table I). To confirm that the cloned cDNA encoded the F4/80 protein, Northern blot analysis was performed using RNA from a range of mouse cell lines and screened with a 757-bp probe from clone pF4/80(12.2). Fig. 2 demonstrates that the transcript recognized by this probe is expressed exclusively in cells of the macrophage lineage, in accordance with the well documented macrophage-specific ex-pression pattern of F4/80. The level of mRNA expression in J774.2 cells is also higher than in RAW 264.7 cells, which correlates with the increased level of F4/80 surface expression on the J774.2 line. The approximate size of the mRNA species (3.2 kb) also correlated well with the size of the cloned cDNA. As further evidence of the identity of the cDNA, CHO-K1 cells were transfected with a full-length expression construct in the pcDNA3 vector. Following selection with Geneticin, cells were analyzed by fluorescence-activated cell sorting for F4/80 antigen expression using the noncompetitive mAbs F4/80 and 5C1 (Fig. 3). The 5C1 mAb was raised against purified F4/80 antigen and has not been described previously. Compared with cells transfected with a control pcDNA3/␤-galactosidase construct, the pcDNA3/F4/80 transfectants demonstrated detectable surface expression of the F4/80 antigen.
The primary amino acid sequence of F4/80 demonstrates a high degree of homology to members of two independent protein superfamilies. First, the NH 2 -terminal region of the protein contains seven tandem EGF-like domains (13). These repeats of approximately 50 amino acids are characterized by the spatial arrangement of six cysteine residues that form three disulfide bonds within each domain, thereby generating a tightly folded structure. A search of the SwissProt and NBRF-PIR data bases with the F4/80 sequence identified a high degree of homology to a number of proteins containing EGF repeats, with the highest scores between F4/80 and connective tissue components such as fibrillin 1 (14) and fibulin 2 (15). As well as the six invariant cysteines, five of the EGF-like domains contain consensus motifs implicated in Ca 2ϩ binding (16), which may play a role in stabilizing the conformation required for ligand interaction. Preliminary analysis of the mouse F4/80 gene demonstrates that each separate EGF-like domain is encoded by a single exon as described for the genomic organization of other EGF superfamily members. 2 A role for EGF-like domains in numerous protein-protein interactions has been proposed, such as the critical requirement of two EGF repeats in the neurogenic Drosophila protein Notch for its interaction with the Delta and Serrate proteins (17).
The second region of homology identified between F4/80 and members of a protein superfamily is located at the COOHterminal region of the F4/80 protein. A hydropathy profile of the F4/80 sequence (18) demonstrated an abundance of hydrophobic residues within a region of approximately 250 COOHterminal amino acids (Fig. 4), suggesting that the molecule may span the cell membrane a number of times. Protein data base searches identified significant homology scores between F4/80 and members of the Tm7 hormone receptor family, including the receptors for peptidic hormones such as parathyroid hormone, calcitonin, vasoactive intestinal peptide, glucagon, and secretin (19). This recently described receptor family shares a common overall topology with an extracellular NH 2 terminus, an intracellular COOH terminus, and a central region consisting of seven transmembrane segments, which re-  sults in three external loops and three internal loops. Characteristic residues were found to be conserved between F4/80 and the members of this particular subset of Tm7 receptors in the transmembrane segments (with the exception of Tm6), the three intracellular loops and the cytoplasmic tail immediately following Tm7. In addition, F4/80, in common with the other members of the family, contains a cysteine residue in each of the first and second extracellular loops. The formation of a disulfide bridge between these cysteines is believed to be crucial to the overall tertiary structure of the receptors. The Tm7 receptors interact with heterotrimeric ␣-␤-␥ G-proteins, on the cytosolic surface of the membrane, which are involved in signal transduction following ligand binding to the extracellular loops of the Tm7 molecule (20). The F4/80 ligand, and function, remains unknown, and its identification will determine whether F4/80 serves as a receptor for a hormone involved in macrophage differentiation and function. The presence of seven EGF-like repeats in the NH 2 -terminal region of F4/80 is an unusual divergence for Tm7 molecules which, with a limited number of exceptions (20,21), have relatively short NH 2 termini with no defined protein superfamily domains. We suggest that F4/80 possibly interacts with two separate ligands, via the EGF-like domains and an extracellular portion of the Tm7 multispan region, respectively. This notion is strengthened by the presence of an Arg-Gly-Asp motif ( Fig. 1A; amino acids 506 -508), often found in matrix proteins with multiple EGF repeats, which could play a role in cell adhesion following recognition by an integrin molecule (22). Based upon the structural elements and macrophage-restricted expression pattern of the molecule, we propose that F4/80 is involved in macrophage adhesion within tissues combined with receptor signaling following its interaction with a peptide ligand, possibly resulting in the activation of adenylate cyclase and increased intracellular cAMP levels (19). This receptor activity may therefore influence macrophage responses within a defined tissue microenvironment. A recent report has described the cloning of a cell surface molecule (designated EMR1) from a human neuroectodermal cDNA library, which shows an extremely high degree of simi-larity to the F4/80 sequence described here (23). This sequence shows 68% overall identity to F4/80 and contains six EGFrepeats and seven postulated transmembrane segments. Reverse transcriptase PCR analysis suggests that expression of the EMR1 molecule is not as tightly regulated as F4/80 is in mice, although increased levels of EMR1 transcripts appear to be expressed in peripheral blood mononuclear cells. The wide distribution of EMR1 expression, in comparison with the restricted pattern of F4/80 expression, is intriguing. The development of mAbs directed to the human protein will aid greatly in defining the localization of cells expressing this molecule, as will in situ hybridization studies on normal human tissue sections. The deduced structure of the F4/80 and EMR1 proteins suggests a role in the cellular response to an undefined hormone or an interaction, possibly through the EGF-like repeats, with an alternative protein ligand. The identification of the F4/80 ligand(s) will help to elucidate the function of this specialized molecule in macrophage physiology.