Cleavage of Ig-Hepta at a “SEA” Module and at a Conserved G Protein-coupled Receptor Proteolytic Site*

Ig-Hepta is a member of a new subfamily of the heptahelical receptors and has an unusually long N terminus extending toward the extracellular side of the plasma membrane. Pulse-chase experiments in 293T cells using antisera specifically recognizing its N- and C-terminal regions demonstrated that Ig-Hepta is core-glycosylated cotranslationally and proteolytically processed into a two-chain form in the endoplasmic reticulum, followed by maturation of oligosaccharide chains and dimerization. The cleavage occurs at two highly conserved sites: one in a “SEA” module (a module first identified in sperm protein, enterokinase, andagrin) near the N terminus and the other in the stalk region preceding the first transmembrane span, generating ∼20-, 130-, and 32-kDa fragments. The latter two remain tightly associated non-covalently even after cleavage as revealed by immunoprecipitation of native and myc-tagged Ig-Hepta constructs that were transiently expressed in 293T cells. The dimer consisting of four chains, (130 kDa + 32 kDa)2, is linked by disulfide bonds. A fusion protein of the extracellular domain of Ig-Hepta and the Fc domain of immunoglobulin was found to be a good substrate of the processing enzymes and used for determining the exact cleavage sites in the SEA module and juxtamembrane stalk region.

G protein-coupled receptors (GPCRs) 1 comprise a large superfamily of proteins in the body and are involved in the recognition and transduction of a variety of extracellular signals. They share a common basic structure of seven transmembrane spans (TM7) with extracellular N and intracellular C termini. Mammalian GPCRs have been classified into three major groups on the basis of their sequence similarity to rhodopsin (class I or type A), secretin receptor (class II or type B), and metabotropic receptor (class III or type C) (1). During the past few years, a subgroup of the class II GPCRs has emerged whose members have unusually large N-terminal extracellular domains that contain a number of well-defined protein modules (2,36). Identification of cell types expressing these molecules and their potential roles in cell adhesion and signaling have become a focus of research in immunology, neuroscience, and developmental biology. This subgroup is referred to as LNB-TM7, where LNB means long N-terminal and type B (2). The members of the LNB-TM7 family include (i) EMR1 (3), CD97 (4), F4/80 (5), and ETL (6), which contain EGF modules; (ii) HE6 (7) and GPR56 (8) with mucin-like regions; (iii) ␣latrotoxin receptors CL1, CL2, and CL3, which have a galactose-binding lectin homologous region and an olfactomedin homologous region (9); (iv) BAI1, BAI2, and BAI3 with thrombospondin type-1 repeats (10,11); (v) Celsr1-3 (12,13) and Flamingo (14) with cadherin repeats; and (vi) Ig-Hepta with immunoglobulin repeats (15). Despite these variations in the membrane-distal region of the extracellular domain, their membrane-proximal regions or stalks are highly conserved. Namely, they contain a characteristic Cys-box motif close to the extracellular face of the membrane, suggesting a common role. For the ␣-latrotoxin receptors, it has been shown that proteolytic cleavage takes place immediately downstream to the conserved Cys-box (9), and similar processing has been suggested for other members (2).
Ig-Hepta cDNA was cloned from a rat lung cDNA library in our laboratory. It was shown to contain two C2-type immunoglobulin-like domains and to be highly glycosylated (15). Northern blot analysis subsequently demonstrated that mRNA transcripts are expressed abundantly in the lung and significantly in the kidney (15). Immunohistochemistry demonstrated alveolar wall and intercalated cell localizations in the rat lung and kidney, respectively (15). One striking and unusual feature of the Ig-Hepta molecule is that it exists as a disulfide-linked dimer; although there are a growing number of GPCRs that can be detected as homo-or heterodimers, many of them are noncovalently associated except the disulfide-linked dimers between calcium-sensing and metabotropic glutamate receptors (16,17) and between and ␦ opioid receptors (18) (for review, see Refs. 19 and 20). While performing Western blot analysis of mature Ig-Hepta to characterize the nature of the dimer, we noticed that N-and C-terminal-directed antisera stained distinct bands of ϳ130 and ϳ32 kDa, respectively, suggesting that Ig-Hepta undergoes proteolytic processing during the biosynthetic process. In the present study, therefore, we performed pulse-chase experiments to confirm this possibility. We further determined the site of cleavage and found it is located in the highly conserved stalk region mentioned above, implying that similar processing also occurs in other members of the LNB-TM7 family.
N-terminal sequencing of Ig-Hepta also revealed another proteolytic processing occurring at a "SEA" module that is located close to the N terminus. The SEA module was first described as a motif present in an ectodomain of a number of mucin-like membrane proteins and so named after the first three proteins in which it was identified (sperm protein, enterokinase, and agrin). The number of SEA module-containing proteins now exceeds 73 (available at dylan.embl-heidelberg.de/). The function of the SEA module is not clear, but it serves as a site for proteolytic cleavage (21,22). Based on this fact, Wreschner et al. (22) have proposed a mechanism whereby a combination of ligand and receptor is generated from a single precursor molecule by its SEA module-mediated cleavage. The significance of the cleavage of Ig-Hepta at its SEA module is discussed in relation to this hypothesis.
Cell Culture-293T cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 5% fetal bovine serum (FBS), 100 units/ml penicillin, and 100 g/ml streptomycin. All the cells were maintained in a humidified atmosphere at 37°C under 5% CO 2 .
Transient Transfections-Transient transfections were performed with the LipofectAMINE Plus system according to the manufacturer's protocol. In brief, 293T cells at 80% confluence in 35-mm dishes were transfected with 1 g of DNA, 6 l of Plus reagent, and 4 l of Lipo-fectAMINE in Opti-MEM medium for 3 h. Cells were harvested 48 h after transfection.
Pulse-Chase Analysis and Immunoprecipitation-For pulse-chase analysis, transiently transfected 293T cells were starved in cysteine/ methionine-free DMEM with dialyzed 5% FBS for 30 min, labeled for 15 min in the same culture medium supplemented with 0.2 mCi/ml [ 35 S]cysteine/methionine. After the pulse, the radiolabeling medium was removed, the cell surface was washed three times with 5 ml of DMEM, and cells were chased with DMEM supplemented with 5 mM cysteine/methionine and 5% FBS. Cells were washed three times with ice-cold phosphate-buffered saline (PBS) and lysed in radioimmune precipitation buffer (0.1% SDS, 1% Nonidet P-40, 1% sodium deoxycholate, 150 mM NaCl, 10 mM sodium phosphate, pH 7.0) with Complete protease inhibitor mixture. The cell lysates were centrifuged at 15,000 rpm for 15 min to remove cellular debris and precleared with protein G-Sepharose beads. Proteins were immunoprecipitated by incubation with antibodies on ice for 2 h and then with protein G-Sepharose beads for another 2 h at 4°C with rotation. Immunoprecipitates were washed five times with radioimmune precipitation buffer, boiled for 5 min at 95°C in Laemmli sample buffer, and analyzed by SDS-PAGE and autoradiography.
Brefeldin A Treatment-293T cells transiently transfected with Ig-Hepta cDNA were pretreated with BFA (5 g/ml) for 1 h and metabolically labeled for 15 min with [ 35 S]methionine/cysteine and chased for 1 h with cold media in the continued presence of BFA. Cells were then washed with BFA-free medium three times and chased for additional 6 h. The cells were lysed at each time point and immunoprecipitated with a mixture of anti-N Ig-Hepta and anti-C Ig-Hepta polyclonal antibody. Immune complexes were analyzed by SDS-PAGE in the presence of 2-mercaptoethanol, and the radioactive bands were visualized.
Deglycosylation Experiments-Membrane proteins from 293T cells transiently transfected with Ig-Hepta were solubilized with 0.5% SDS (w/v), and the extracts (20 g of protein) were boiled for 3 min in 20 l of 0.1 M 2-mercaptoethanol, and 50 mM sodium phosphate buffer, pH 7.2. After cooling to room temperature, Nonidet P-40 was added to give a final concentration of 1%. To the extracts, 2 milliunits of PNGase F (EC 3.5.1.52, Roche Molecular Biochemicals) were added, and the reaction mixture was incubated at 37°C for 3 h. A control incubation was carried out in which 50 mM sodium phosphate buffer was added in place of the enzyme. The proteins were separated by SDS-PAGE, transferred to a polyvinylidene difluoride membrane, and detected by immunoblotting as described above.
Enzymatic Deglycosylation of Immunoprecipitants-Following immunoprecipitation, samples were washed with radioimmune precipitation buffer without proteinase inhibitors for three times and eluted from the beads by incubation in 30 l of denaturing buffer (0.5% SDS, 1% 2-mercaptoethanol). Treatment with PNGase F was performed for 3 h at 37°C as described above. Samples were electrophoresed on polyacrylamide.
Antibody Production-For antibody production, DNA fragments encoding the C-terminal cytosolic domain of Ig-Hepta (residues 1268 -1349) and the N-terminal extracellular domain of Ig-Hepta (residues 571-1016) were amplified by PCR and cloned into the vector pRSET, and the constructs were transformed into Escherichia coli BL21 (DE3)pLysS and used for fusion protein production. Fusion proteins were insoluble and purified under denaturing conditions on nickelnitrilotriacetic acid resin columns. Rabbit polyclonal antibodies using the recombinant proteins were produced as previously described (15).
Stable Transfection of 293T Cells with sIg-Hepta-Fc cDNA-To establish cell lines that express sIg-Hepta-Fc, the expression vector encoding sIg-Hepta-Fc was linearized by ScaI restriction enzyme and transfected into exponentially growing 293T cells using the Lipo-fectAMINE Plus method. Selection of stable transfectants was carried out by adding 500 g/ml Zeocin into the medium. After 3 weeks, Zeocinresistant cells were cloned by limiting dilution and expanded to obtain stable cell lines. Expression of sIg-Hepta-Fc was determined by Western blotting of cell supernatants.
Purification of sIg-Hepta-Fc-Stably transfected cells were grown for 7 days in serum-free Opti-MEM medium, and the supernatants were harvested, centrifuged (10,000 ϫ g, 1 h), filtered through a 0.22-m filter membrane to remove cell debris and membranes, and loaded as 40-ml aliquots onto 1-ml columns of protein G-Sepharose equilibrated in PBS. The columns were washed with 10 ml of PBS and eluted with 2.5 ml of 50 mM citric acid. The eluates were neutralized with 1 M Tris base and concentrated using Centriprep-10 concentrators. Protein concentration was determined by the bicinchoninic acid (BCA) method using bovine serum albumin as the standard, and the purity of the samples was assessed by SDS-PAGE and Coomassie Blue staining.
N-terminal Sequence Analysis-For N-terminal sequencing, 5 g of purified sIg-Hepta-Fc was electrophoresed on 10% SDS-polyacrylamide gel and electrotransferred onto polyvinylidene difluoride membranes. The protein band of interest was excised from the membrane, and N-terminal sequence was determined by automated Edman degradation using a PPSQ-21 sequencer (Shimadzu, Kyoto, Japan).
Site-directed Mutagenesis of Ig-Hepta-GFP-A single Thr (ACG) to Ala (GCG) mutation at proteolytic cleavage site (P1Ј, residue 994) was introduced into pEGFP-N3-Ig-Hepta by PCR-based mutagenesis using mutated synthetic oligonucleotides. Briefly, a DNA fragment was amplified using the forward primer T994A-Xho-F shown below and the antisense mutant primer T994A-Nru-R, and an overlapping fragment was amplified using the sense mutant primer T994A-Nru-F and the downstream reverse primer T994A-Apa-R. Both fragments were gelpurified, cut with NruI, ligated, and re-amplified with the Pfu DNA polymerase with the T994A-Xho-F and T994A-Apa-R primers. The product was digested using Aor51HI-Bpu1102I restriction enzymes, gel-purified, and subcloned into pEGFP-N3-Ig-Hepta to replace the target segment. DNA sequencing was used to verify the sequence. The following primers were used for constructing the above mentioned point mutations (the underlined sequences indicate the mutated bases): Detection of GFP-Cell extract was equilibrated with Laemmli sample buffer and kept at room temperature for 5 min. Ten g were subjected to SDS-PAGE analysis at 4°C, and fluorescent bands were detected by exposing the gel to Bioimage Analyzer FLA2000 (Fujifilm, Tokyo, Japan) at 473-nm excitation and 510-nm emission.

RESULTS
Evidence for Proteolytic Processing of Ig-Hepta-While characterizing antisera, anti-N Ig-Hepta and anti-C Ig-Hepta , raised against the N-and C-terminal domains of rat Ig-Hepta, respectively, we noticed that they recognize distinct bands on Western blot analysis of Ig-Hepta (Fig. 1A). When extracts of 293T cells transiently expressing Ig-Hepta were analyzed, anti-N Ig-Hepta antiserum reacted with broad bands of 30 -160 kDa and their dimers (lanes 1 and 2), whereas anti-C Ig-Hepta antiserum detected a band of 32-kDa and its dimer (lanes 3 and 4). In both cases, the dimer bands disappeared upon reduction (data not shown). The broad bands of 130 -160 kDa can be explained by glycosylation; treatment with glycosidase PNGase F, which cleaves both high mannose and complex N-linked oligosaccharides, reduced their sizes to ϳ95 kDa (1B, lane 3). These results suggest that Ig-Hepta is proteolytically processed into a mature two-chain form composed of the fully glycosylated 150-kDa N-terminal fragment and the 32-kDa C-terminal fragment.
To see whether this proteolytic processing is a general property of Ig-Hepta, we next performed Western blot analysis using whole detergent extracts of the rat lung. Although anti-N Ig-Hepta detected 130-to 160-kDa bands corresponding to the N-terminal fragment, anti-C Ig-Hepta yielded only a faint band of 32 kDa that was difficult to distinguish from nonspecific staining (data not shown). We therefore enriched the C-terminal fragment by immunoprecipitation and subjected it to Western blot analysis (Fig. 1C), which demonstrated the presence of the processed C-terminal fragment of 32 kDa in the lung (lane 5) as well as in 293T cells (lane 4).
Processing of Ig-Hepta Monitored by Pulse-Chase Experiments-To determine the time course of the post-translational modification of Ig-Hepta, we performed pulse-chase experiments. 293T cells expressing rat Ig-Hepta were pulse-labeled with [ 35 S]methionine/cysteine for 15 min and chased for 0 -120 min (Fig. 2). Immunoprecipitation of labeled products with either N-or C-terminal domain-directed antiserum revealed that rat Ig-Hepta is synthesized as a 170-kDa precursor and is rapidly cleaved into two chains: a 130-kDa N-terminal fragment ( Fig. 2A, upper panel, arrowhead) and a 32-kDa C-terminal fragment ( Fig. 2A, lower panel). The size of the C-terminal fragment remained unchanged during the 2-h chase time. However, the size of the N-terminal fragment increased from 130 to 150 kDa after 1 h of chase ( Fig. 2A, arrow) because of modification of oligosaccharide chains as demonstrated by deglycosylation of the pulse-chased products (Fig. 2B). Deglycosylation experiments also showed that the initial glycosylation (attachment of a common N-linked oligosaccharides or core-glycosylation) occurs almost cotranslationally as seen by a marked reduction (ϳ20 kDa) in size of the 170-kDa band as well as the band of 130 kDa (Fig. 2, A and B). The time courses of process-ing indicate that the cleavage occurs in the endoplasmic reticulum (ER) before the late steps in the processing of oligosaccharide chains is completed in the Golgi complex. To confirm this speculation, we performed the following experiment using BFA, an inhibitor that prevents the ER-to-Golgi vesicular trafficking. The identity of an ϳ150-kDa band maximally seen at 20 min of chase ( Fig. 2A, double arrowhead on the left) will be addressed later.
Proteolytic Cleavage of Ig-Hepta in ER-293T cells were transiently transfected with an Ig-Hepta expression vector, preincubated for 1 h with BFA to achieve complete inhibition, pulse labeled for 15 min in the presence of BFA, and chased for 1 h in a medium containing BFA. BFA was then removed, the chase was further continued for 6 h, and cell lysates were assayed for the molecular species of Ig-Hepta accumulated during the chase by immunoprecipitation, SDS-PAGE, and autoradiography. As shown in Fig. 3, Ig-Hepta that accumulated in ER in the presence of BFA was the proteolytically processed form consisting of the 130-kDa N-terminal fragment (Fig. 3, lanes 2-4, upper panel) and 32-kDa C-terminal fragment (Fig. 3, lower panel). However, the possibility cannot be excluded that the cleavage could be occurring due to the presence of enzymes normally found in the cis-Golgi, because some of the enzymes normally found in the early part of the Golgi are known to be present in ER through retrograde transport. On removal of BFA, the 130-kDa N-terminal extracellular domain was gradually converted to the fully glycosylated higher molecular weight species (lanes 5 and 6), suggesting that the complete maturation of the sugar chains occurs at a later stage in the Golgi. From the sizes of the cleaved fragments, the site of cleavage is predicted to be located in a juxtamembrane region of the N-terminal extracellular domain.
Efficient Cleavage of Ig-Hepta-The degree of cleavage was monitored by constructing a fluorescent derivative of Ig-Hepta termed Ig-Hepta-GFP that has a GFP tag at its C terminus. Because GFP is stable, if not heated, in the Laemmli sample buffer for SDS-PAGE, this construct allowed us to detect the GFP-tagged C-terminal fragment by using a fluorescence gel scanner (Fig. 4). A single band of ϳ50 kDa was seen when extracts of 293T cells were transiently transfected with the Ig-Hepta-GFP construct (lanes 2 and 5, WT-GFP). The absence of higher molecular weight unprocessed species indicates that the processing occurs highly efficiently.
Non-covalent Association of N-and C-terminal Fragments-Despite the efficient cleavage of the N-terminal extracellular domain of Ig-Hepta during its post-translational modification, the cleaved extracellular domain was not released into the culture medium (Fig. 5A, lane 1) and was rather recovered from membrane fractions (lane 3). As expected, when a truncated form covering only the extracellular domain of Ig-Hepta (Ig-Hepta-ECD) was expressed in the same expression system, it was secreted and recovered from the culture medium (Fig. 5A,  lane 2). These pieces of experimental evidence suggest that the N-terminal fragment is tightly associated with certain membrane component(s), most likely with its C-terminal fragment. To explore this possibility, we prepared Ig-Hepta that has an myc tag at its C terminus (Ig-Hepta-myc) and performed immunoprecipitation analysis using a commercially available anti-myc monoclonal antibody. The anti-myc antibody precipitated the myc-tagged C-terminal domain together with the N-terminal fragment (Fig. 5B). These results clearly demonstrate non-covalent association of the cleaved N-and C-terminal fragments.
Determination of the Cleavage Site in Membrane Proximal Region-In an attempt to determine the proteolytic cleavage site of Ig-Hepta, we first tried immunoaffinity purification of the 32-kDa C-terminal fragment. Such studies, however, have been hampered because of its membrane-bound nature, and hence, difficulties result in obtaining sufficient amounts for sequencing. As an alternative approach, we constructed a chimeric protein that contains a candidate cleavage site sequence and is easy to purify. The chimera was composed of the Nterminal extracellular domain of rat Ig-Hepta and the Fc domain of human IgG1 and named sIg-Hepta-Fc (s for soluble). When the chimeric construct was expressed in 293T cells, sim- ilar proteolytic processing occurred, yielding a 38-kDa fragment corresponding to the C-terminal Fc portion (Fig. 6A). The C-terminal fragment was purified from the culture medium by affinity chromatography on protein G-Sepharose and found to have the following N-terminal sequence by amino acid sequencing: TSFSILMSPD (Fig. 6B, upper sequence). This result strongly indicates that the cleavage site is Leu 993 -Thr 994 , which is located in a juxtamembrane region of the N-terminal extracellular domain (23 amino acid residues N-terminal to the first transmembrane span; Figs. 6B and 7).
To confirm the site of cleavage, we constructed a mutant Ig-Hepta-GFP molecule whose cleavage site sequence is changed by site-directed mutagenesis. Mutation of Thr 994 to Ala (T994A-GFP) at the P1Ј site completely abolished the cleavage (Fig. 4, lane 3). This result strongly supports the above The arrows designate the points of cleavage in the sIg-Hepta-Fc. The N-terminal 10 amino acid residues of the 150-kDa N-terminal and 38-kDa Fc fragments, determined by automated Edman degradation, are shaded. C2, C2-type immunoglobulin-like domain; the dark box at the N terminus represents a presequence of about 24 residues that was removed cotranslationally; vertical bars indicate potential N-linked glycosylation sites; Fc, Fc domain of human immunoglobulin. C, schematic model of mature Ig-Hepta that is a disulfide-linked homodimer whose monomeric unit is composed of two non-covalently associated chains generated by proteolytic processing of the precursor at the SEA module (site 1) and the GPCR proteolytic site (GPS, site 2) near the juxtamembrane region. The N-terminal fragment generated by site 1 cleavage appears to be released. Dark boxes represent transmembrane spans.
conclusion that the peptide bond between Leu 993 and Thr 994 in the juxtamembrane region is the authentic cleavage site. Although the juxtamembrane nature of the cleavage site is reminiscent of that of membrane protein secretases, the relatively strict sequence specificity and the ability to cleave even a soluble substrate, sIg-Hepta-Fc, indicate that the proteinase involved in the processing of Ig-Hepta may be different from the secretases.
Another Cleavage within a "SEA" Module Near the N Terminus-A surprising result was obtained when the N-terminal sequence of sIg-Hepta-Fc (Fig. 6A, lane 3, upper band) was determined. Our expectation was that it begins just after the signal sequence, but it yielded a sequence beginning from a residue far beyond the signal sequence cleavage site, namely from Ser 224 : SVVVDYIVEV (Fig. 6B, lower sequence). A data base search (available at dylan.embl-heidelberg.de/) indicated that the sequence around this cleavage site conforms well to the consensus sequence of the SEA module (23), which has been established as a proteolytic cleavage site of membraneassociated mucin proteins, including MUC1 (22). After cleavage, the relatively short N-terminal fragment appears to be released from sIg-Hepta-Fc, because no corresponding band of 20 -30 kDa was observed on SDS-PAGE of the affinity-purified sIg-Hepta-Fc (Fig. 6A, lane 3).
This cleavage between Gly 223 and Ser 224 explains the intermediate band of ϳ150 kDa seen in an early phase of the pulse-chase experiment ( Fig. 2A, double arrow). The N-terminal fragment cleaved off at the SEA module, however, could not be chased, because the anti-N Ig-Hepta , used for immunoprecipitation of the labeled products, are directed to the Ig-like repeat region and therefore its epitope does not cover the most Nterminal short fragment.

DISCUSSION
The LNB-TM7 subfamily has recently emerged in the class II or type B GPCR family (2,36). Typical members include (i) leukocyte activation antigen CD97, which is associated with inflammation (4); (ii) epithelial cell-restricted cell-surface glycoprotein HE6 of the epididymis (7); (iii) receptors for ␣-latrotoxin, a potent excitatory neurotoxin in the black widow spider venom (9); (iv) brain-specific angiogenesis inhibitors BAI1-3 (10, 11); (v) cadherin/EGF/laminin A G-type repeat/seven-pass receptor Celsr1 (12) and its close relative, Flamingo, whose Drosophila relative is localized at cell-cell boundaries in the wing and involved in the regulation of planar cell polarity (14); (vi) ETL with EGF-like repeats whose expression is developmentally regulated (6); and (vii) Ig-Hepta studied here. These members are considered to be putative GPCRs based on their structural features; in the case of the ␣-latrotoxin receptors, there is direct evidence for their interaction with G o , a G protein (24). Attention is now focused on their roles in cell-cell recognition, cell adhesion, and signaling. Concerning the processing during maturation, only limited information is available. For example, although most LNB-TM7 family members have been suggested to be proteolytically cleaved during biosynthesis (6,9,14,25), their processing events have not been studied in detail except the determination of the cleavage site in the calcium-independent receptor of ␣-latrotoxin (CIRL or CL1) (9). In the present study, we monitored the biosynthesis of Ig-Hepta by pulse-chase experiments and clarified its proteolytic processing and glycosylation.
Interestingly, two cleavage sites were identified by using a chimeric construct of Ig-Hepta containing Fc as a tag for purification. One site is located close to the extracellular surface of the membrane (Figs. 6B and 7, site 2). The amino acid sequences surrounding this cleavage site are highly conserved among the family members (Fig. 7), suggesting that similar processing occurs in other members; in fact, the processing site of CL1 determined by Krasnoperov et al. (9) is indeed identical to that of Ig-Hepta determined here. This explains why the stalk regions are highly conserved among the family members despite striking differences in the membrane-distal regions. It is also noteworthy that, in addition to the family members mentioned above, there are many protein sequences deposited in the DDBJ/GenBank TM /EBI data base that contain the conserved Cys-box and cleavage site sequences ( Fig. 7 and accession numbers listed in the legend), many of which are members of the LNB-TM7 family, but there are examples of the following: (i) Type 1 membrane proteins with a single transmembrane span, including suREJ1, a sea urchin sperm glycoprotein considered to be acting as a receptor for egg jelly (26,27) and FIG. 7. Protein sequence alignment of the juxtamembrane region of LNB-TM7 family members. Amino acid sequences obtained from the NCBI data base for LNB-TM7 members were aligned with the ClustalW program followed by manual adjustment. The lower two lines of sequences represent similar juxtamembrane sequences found in other family members of transmembrane proteins. The conserved juxtamembrane cysteine residues (Cys-box) are boxed, and the proteolytic cleavage site is indicated by an arrow. TM1, the first transmembrane span; 7 TM, protein with seven transmembrane spans; 1 TM, protein with a single transmembrane span; and 11 TM, protein with 11 transmembrane spans. The DDBJ/GenBank TM /EBI accession numbers for the proteins aligned are as follows: rat Ig-Hepta (AB019120), human HE6 (X81892), rat CIRL (U72487), rat ETL (AF192401), mouse Flamingo (AB028499), human CD97 (XM_031278), human BAI1 (AB005297), mouse Celsr1 (NM_009886), human EMR1 (NM_001974), Strongylocentrotus purpuratus sperm receptor for egg jelly (suREJ) (AAB08448), and S. purpuratus sperm receptor for egg jelly 3 (suREJ3) (AAL26499). Other sequences deposited in the DDBJ/GenBank TM /EBI data base containing the conserved Cys-box motif present in the juxtamembrane region are as follows: human Ig-Hepta (AL096772), human latrophilin-2 (AJ131581), human latrophilin-3 (AF307080), Drosophila melanogaster Flamingo (AB028498), human BAI2 (AB005298), human BAI3 (AB005299), human EMR2 (XM_009412), human EMR3 (XM_030847), Caenorhabditis elegans cosmid B0286 (U39848), C. elegans cosmid B0457 (Z54306), and Hemonchus contortus 110-R (AJ272270).
(ii) channel-like proteins with 11 putative transmembrane spans, including suREJ3 (28) and its human homolog (29), a member of the polycystin-1 family whose last six transmembrane spans are similar to those of voltage-activated calcium channels (30 -32). Until recently, the cleavage has been thought to be a phenomenon specifically seen in the GPCR processing, and the cleavage site has been called "GPCR proteolytic site" (GPS). But the motif name GPS becomes inappropriate, because similar proteolytic cleavages have been demonstrated in suREJ3 (28) and here in soluble, secreted sIg-Hepta-Fc (Fig. 6A). The recognition of the cleavage site, therefore, appears to be solely dependent on the sequence surrounding the processing site. The Cys-box motif present immediately before the cleavage site is characterized by four invariant Cys residues (Fig. 7). The presence of a large number of proteins containing the Cys box followed by the cleavage site sequence suggests that the proteolytic processing seen in Ig-Hepta is a general event in the maturation of precursors of a certain group of membrane proteins. The conserved juxtamembrane location of the cleavage site suggests that the processing enzyme responsible is a membrane protein, and such a juxtamembrane location allows an efficient encounter between the substrate and the active site of the enzyme. Indeed, efficiency of the processing was significantly reduced in a soluble secreted form (Ig-Hepta-Fc) compared with the membrane-bound form (Figs. 4 versus 6A). The nature of the non-covalent association of the cleaved fragments is also interesting but remains to be clarified.
The other cleavage site identified by N-terminal sequencing is located in the SEA module (Fig. 8A). The SEA module is an extracellular domain found in a number of highly O-glycosylated membrane proteins (for a list, see the Single Modular Architecture Research Tool (SMART) site, dylan.embl-heidelberg.de/) and serves as a proteolytic cleavage site (22,23). The C-terminal sequence of the module present in Ig-Hepta ( 223 GS-VVV 227 ), and its cleavage site (Gly 223 -Ser 224 ) match perfectly those found in MUC1, a mucin-like transmembrane protein (21). Wreschner et al. (22) have recently proposed a mechanism whereby one and the same gene can encode both a receptor protein and its specific ligand. It is postulated that generation of such receptor-ligand partnership is effected by SEA modulemediated proteolytic cleavage of transmembrane proteins such as MUC1. They further suggested that Ig-Hepta might be a typical example of the ligand-receptor alliances based on the presence of a perfect SEA module in Ig-Hepta. In the present study, we established that the module actually functions as a site for proteolytic cleavage. Furthermore, in the case of Ig-Hepta, the cleaved N-terminal fragment contains a potential prohormone processing site ( 48 RPKR 51 ; Fig. 8, A and C) and a repeat of two variant forms of the EGF-like domain signature 2 ( Fig. 8D) (33), making the hypothesis and hence the functional characterization of the N-terminal fragment very attractive.
Several groups of proteinases, including the following two major groups, have been identified that are involved in posttranslational proteolysis of membrane proteins and secreted proteins: (i) membrane protein secretases that are involved in the processing of the Alzheimer's amyloid precursor protein, angiotensin-converting enzyme, certain cytokine receptors, and others (for review, see Ref. 34) and (ii) subtilisin/Kex2plike endoproteinases such as furin and prohormone convertases (35). The majority of secretases are metalloproteinases, located at the cell surface or in vesicles close to the plasma membrane, and cleave precursors at a site located a fixed distance from the membrane; the cleavage is not strictly dependent on the sequence. Members of the prohormone convertase family are mainly localized in the trans-Golgi network and cleave prohor-mones at sites marked by paired or multiple basic amino acid residues. Although the identity of the processing enzyme for Ig-Hepta remains to be clarified, there is a possibility that it may represent a new family of processing proteinases that are present in the endoplasmic reticulum and recognize the highly conserved HLTXF(S/A)(I/V)L(M/L) sequence following the Cysbox motif (Fig. 7). The facts that proper processing of Ig-Hepta occurs in cultured 293T cells and COS-7 cells and that the potential recognition sequences are found in a variety of proteins in the Expressed Sequence Tag data base suggest that one or more of the processing enzymes are ubiquitously distributed and involved in a novel mechanism of proteolytic processing common to the Cys box-containing proteins. Our finding that sIg-Hepta-Fc, a soluble chimeric construct of Ig-Hepta, can serve as a good substrate for the processing enzyme may stimulate the purification of the enzyme, because availability of soluble substrates is essential for purification.