Activation and Functional Characterization of the Mosaic Receptor SorLA/LR11*

We previously isolated and sequenced the ; 250-kDa type 1 receptor sorLA/LR11, a mosaic protein with ele-ments characterizing the Vps10p domain receptor family as well as the low density lipoprotein receptor family. The N terminus of the Vps10p domain comprises a consensus sequence for cleavage by furin ( 50 RRKR 53 ) that precedes a truncation found in sorLA isolated from human brain. Here we show that sorLA, like sortilin-1/ neurotensin receptor-3, whose lumenal domain consists of a Vps10p domain only, is synthesized as a proreceptor that is cleaved by furin in late Golgi compartments. We show that the truncation conditions the Vps10p domain for propeptide inhibitable binding of neuropeptides and the receptor-associated protein. We further demonstrate that avid binding of the receptor-associated protein, apolipoprotein E, and lipoprotein lipase not inhibited by propeptide occurs to sites located in other lumenal domains. In transfected cells, about 10% of full-length sorLA were expressed on the cell surface capable of mediating endocytosis. However, the major pool of receptors was found in late Golgi compartments, suggesting possible interaction with newly synthesized ligands. The results show that sorLA, following activation by truncation, binds multiple ligands and may mediate both endocytosis and sorting. new, ; 10.000 cpm 125 I-GST-sorLA propeptide, and an increase in the acid-soluble fraction was taken as a measure of degradation. Internalization of IL-2R/sorLA Chimera— Transfectants or CHO-K1 control cells were incubated fo r 2 h at 4 °Cwith 125 I-labeled antibody (3 3 10 4 cpm/ml) directed against the lumenal IL-2R domain (anti-Tac, Roche). Following washings in ice-cold buffer, the cells were reincu-bated in 37 °C warm medium for 0–60 min. The incubations were stopped at various times by the addition of ice-cold Tris-HCl buffer, pH 2.5. The supernatant was recovered after 5 min, and the cells were lysed in 1 M NaOH. The radioactivity was determined in both fractions and defined as cell surface-associated ( i.e. acid-releasable) and internalized ligand, respectively.


From the Department of Medical Biochemistry, University of Aarhus, DK-8000, Aarhus C, Denmark
We previously isolated and sequenced the ϳ250-kDa type 1 receptor sorLA/LR11, a mosaic protein with elements characterizing the Vps10p domain receptor family as well as the low density lipoprotein receptor family. The N terminus of the Vps10p domain comprises a consensus sequence for cleavage by furin ( 50 RRKR 53 ) that precedes a truncation found in sorLA isolated from human brain. Here we show that sorLA, like sortilin-1/ neurotensin receptor-3, whose lumenal domain consists of a Vps10p domain only, is synthesized as a proreceptor that is cleaved by furin in late Golgi compartments. We show that the truncation conditions the Vps10p domain for propeptide inhibitable binding of neuropeptides and the receptor-associated protein. We further demonstrate that avid binding of the receptor-associated protein, apolipoprotein E, and lipoprotein lipase not inhibited by propeptide occurs to sites located in other lumenal domains. In transfected cells, about 10% of fulllength sorLA were expressed on the cell surface capable of mediating endocytosis. However, the major pool of receptors was found in late Golgi compartments, suggesting possible interaction with newly synthesized ligands. The results show that sorLA, following activation by truncation, binds multiple ligands and may mediate both endocytosis and sorting.
A new, highly conserved type 1 receptor, termed sorLA/ LR11, was recently identified in man, rodents, chicken, and hydra (1)(2)(3)(4)(5)(6). SorLA is a mosaic protein that, based on its domain structure, may be regarded as a member of both the vacuolar protein sorting 10 protein (Vps10p) 1 domain receptor family and the low density lipoprotein receptor (LDLR) family. The lumenal part of the ϳ250-kDa mammalian receptor (see Fig. 1) comprises a domain with homology to the yeast sorting protein Vps10p and the mammalian Vps10p domain receptors sortilin-1 (7) and sorCS (8), a cluster of five F/YWTD/␤-propeller modules, an epidermal growth factor precursor-like module, and a cluster of 11 LDLR class A (LA) repeats, all found in LDLR family members, and a domain of six structural elements (fibronectin type III repeats) found in neural cell adhesion molecules. SorLA has a single transmembrane domain and a 54-residue cytoplasmic tail comprising features typical of both endocytosis-and sorting-competent receptors (1,2).
SorLA is mainly expressed in the nervous system but is also found in non-neuronal tissues such as testis, ovary, and lymph nodes (1)(2)(3)(4). In situ hybridization analysis of adult murine brain showed sorLA transcripts in specific neuronal populations, including Purkinje cells of cerebellum and neurons in hippocampal formations and cerebral cortex (4,5). Similar results were obtained by immunohistochemistry, which also suggested a predominating intracellular localization (9). In addition, sorLA is expressed in cell bodies and proximal axons of cultured rat sympathetic neurons (10). SorLA expression is developmentally regulated, and high transcript levels are found during periods of active morphogenesis (4,5). Interestingly, sorLA is up-regulated during proliferation and downregulated following differentiation of neuroblastoma cells (11). Concerning non-neuronal cells, recent results have shown low expression of sorLA in normal rabbit aortas but a comparatively marked expression in the intimal smooth muscle cells of aortas displaying atheromatous lesions in rabbits fed with a high cholesterol diet (12). These observations suggest that sorLA may play a role in developmental cellular proliferation and in pathological events.
Little is known about the molecular function of sorLA. Until now, two ligands have been identified qualitatively by blotting to electrophoretically resolved sorLA. One is the 39-kDa receptor-associated protein (RAP) used for affinity purification of sorLA from brain extracts (1). RAP is an endoplasmatic reticulum resident protein that binds to clusters of LA repeats in all members of the LDL receptor family (for review, see Refs. 13 and 14) and to sortilin-1 (7,15). The other is rabbit ␤-very low density lipoprotein (2), a chylomicron remnant surrogate rich in apolipoprotein E (apoE) that also binds to the LA repeat clusters of LDLR family members. In addition, the neuropeptide head activator binds to a component in solubilized human neuronal NT2 cells that most likely represents sorLA (16). The head activator, a conserved 11-residue peptide of unknown function in mammals, stimulates head-specific growth in hydra and is reported to stimulate mitosis in NT2 cells (6,16). These results, together with the structural features, suggest that sorLA is a multifunctional receptor that may be involved in ligand transport and sorting and perhaps in the propagation of ligand-induced signaling.
We recently showed that sortilin-1 (ϳ95 kDa), whose lumenal part consists of a Vps10p domain only, is conditioned for binding of ligands such as RAP, neurotensin, and lipoprotein lipase (LpL) by furin-mediated cleavage and removal of the propeptide (15,17). SorLA purified from brain tissue is Nterminally truncated, and the preceding sequence 50 RRKR 53 abides by the consensus for cleavage by furin (1). It therefore seemed possible that sorLA, like sortilin-1, is synthesized as a proreceptor and activated by truncation.
The aim of the present work was to analyze whether sorLA is activated by cleavage, to characterize binding of candidate ligands, and to establish whether the receptor is endocytosiscompetent. We show by analysis of wild type and mutated minireceptors secreted from transfected Chinese hamster ovary (CHO) cells that sorLA is activated by furin-mediated cleavage and removal of the propeptide. The truncated Vps10p domain was purified by affinity chromatography, and ligand binding to this domain was compared with binding to purified soluble sortilin-1 (i.e. the Vps10p domain) and to purified soluble sorLA comprising the entire lumenal part (L-sorLA). We show that the Vps10p domain of sorLA, like that of sortilin-1, binds propeptides, neuropeptides, and RAP, apparently to overlapping sites. In addition, we show avid binding of RAP, apoE, and LpL to different domains in L-sorLA, presumably to the LA repeat cluster. In cells transfected with full-length sorLA, about 10% of the receptors were expressed on the cell surface, capable of mediating endocytosis and degradation of bound ligand, whereas most receptors were found in perinuclear compartments colocalizing with the Golgi protein golgin-97.

EXPERIMENTAL PROCEDURES
Antibodies and Ligands-Rabbit anti-Vps10p domain peptide antiserum was raised (Neosystem, Strasburg, France) against the synthetic peptide Gly 370 -Arg 388 and used for initial detection of sorLA-(54 -731). Purified Vps10p domain (see below) was used to generate a rabbit anti-sorLA Vps10p domain antibody, and purified propeptide (residues 1-53) was used to generate anti-propeptide antibody (DAKO, Glostrup, Denmark). To produce rabbit antibodies against the LA repeat cluster for Western blotting of full-length sorLA, the segment (Val 1044 -Asn 1532 ) was expressed in Escherichia coli as a thioredoxin fusion protein and partially purified by heat treatment as described for the LA repeats of the LDL receptor (18). The Ig fractions were purified from rabbit serum using protein A-Sepharose and kept at a stock concentration of ϳ5 mg/ml.
Recombinant RAP and GST-sortilin-1 propeptide, as well as the soluble lumenal domain of sortilin-1 (i.e. the Vps10p domain), were produced as described previously (15,19). Coupling of peptides to CNBr-activated Sepharose 4B (AP Biotech) followed instructions from the manufacturer. The head activator purchased from Bachem (Heidelberg) was monomerized by boiling 1 mM peptide in 0.1 M HCl and neutralizing to pH 7.0 (20). Neurotensin was from Sigma, and recombinant apoE, isoform 3, was from Calbiochem. LpL purified from bovine milk as active dimeric enzyme (ϳ600 units/mg) was a generous gift from Dr. G. Olivecrona (Umeå University, Umeå, Sweden) (21).
Expression of SorLA Constructs-The sorLA cDNA (GenBank TM accession number U60975) was obtained from two overlapping clones, both in pBluescript SK: one from a human brain ZAP cDNA library (Stratagene, La Jolla, CA) covering the C-terminal 1925 amino acid residues and one overlapping clone in pBK-CMV from a Jurkat ZAP Express library covering the N terminus and a 5Ј-untranslated sequence (1). A fragment covering the 5Ј end was generated by cleaving the Jurkat clone with NarI and blunt-ended with T4 polymerase followed by NheI cleavage. The brain-derived clone was cleaved with XbaI (blunt-ended) and NheI. The full-length sorLA sequence was generated by inserting the NarI/NheI Jurkat 5Ј fragment into the predigested brain-derived clone, and a NotI/XhoI fragment covering full-length sorLA was then transferred from pBluescript to the pcDNA 3.1/Zeo(ϩ) vector (Invitrogen, Carlsbad, CA). A fragment covering residues 1-731 was amplified by PCR using a T7 5Ј primer and a 3Ј primer containing an XhoI cleavage site and the C terminus of the fragment (GACCTC-GAGTTACTCGTTCTCTTCGAGGGG) and cloned into the same vector to produce the soluble minireceptor. The consensus site for cleavage by furin Arg-Arg-Lys-Arg 53 was mutated to Gly-Arg-Lys-Gly by PCR-mediated overlap extension of two PCR fragments produced with (i) forward mutation primer (5Ј-AGAAGCCGCTCGGGAGGAAAGGGAG-CGCTGCCCT) and primer downstream in the sorLA sequence and (ii) reverse mutation primer (5Ј-AGGGCAGCGCTCCCT TTCCTCCCGA-GCGGCTTCT) and primer upstream in the vector. The NotI/PpuMIcleaved PCR fragment was cloned into NotI/PpuMI-cleaved sorLA-(1-731) pcDNA 3.1/Zeo(ϩ). To produce a construct covering the entire lumenal part (L-sorLA), full-length sorLA in pBluescript was digested with NsiI and XhoI. A linker with compatible ends encoding Ser 2098 to Asp 2107 followed by two stopcodons was constructed and, following annealing of the 5Ј-phosphorylated oligonucleotides (TCTGCAACG-CAGGCTGCCAGATCTACGGATTGATGAC and TCGAGTCATCAA-TCCGTAGATCTGGCAGCCTGCGTTGCAGATGCA), the linker was inserted into the NsiI/XhoI-predigested pBluescript vector. Finally, the resulting NotI/XhoI fragment covering sorLA-(1-2107) was transferred to the pcDNA 3.1/Zeo(ϩ) vector. To produce a construct covering the lumenal part of the interleukin 2 receptor (Tac/CD25) and the transmembrane and cytoplasmic domains of sorLA (IL-2R/sorLA chimera), IL-2R was transferred from pCMVIL-2R (22) by NheI-XbaI digestion and ligation to pcDNA 3.1/Zeo(ϩ), and the IL-2R/sorLA chimera was obtained by PCR-mediated overlap extension of two fragments. One was produced by PCR with pCMVIL-2R as a template using primer containing a partial sorLA transmembrane sequence and an IL-2R lumenal domain sequence (5Ј-CACCACAGCAGCAACCTGGTACTCT-GTTGT) in combination with an upstream IL-2R primer. The other was produced by PCR with sorLA as template using the reversed primer (5Ј-ACAACAGAGTACCAGGTTGCTGCTGTGGTG) in combination with a downstream primer in pcDNA 3.1/Zeo(ϩ). The extended fragment was digested with Bsu36I and XhoI and cloned into predigested IL-2R/pcDNA 3.1/Zeo(ϩ). All constructs were verified by DNA sequencing.
Expression of SorLA Propeptide-The propeptide sequence was amplified from sorLA cDNA using a 5Ј primer containing a BamHI site and the propeptide N-terminal sequence (5Ј-CTTGGATCCGAAGTCTG-GACGCAGAGG) and a 3Ј primer containing a stop codon and a XhoI site C-terminal to the furin cleavage site (5Ј-GTTCTCGAGTCACCGTT-TCCTCCGGAG). Amplification of the GC-rich stretch was facilitated by 5% Me 2 SO in the PCR reaction (23). The purified PCR product was digested with BamHI/XhoI and cloned into the pGEX-4T-1 vector (AP Biotech). The construct was expressed in bacterial strain BL21(DE3) and purified by affinity chromatography on glutathione-Sepharose. The fusion product was cleaved with thrombin (Sigma), and the propeptide was separated from GST and residual fusion product using Centricon-10 (Amicon, Beverly, MA) and concentrated using Centricon-3.
Metabolic Labeling, Affinity Precipitation, and Analysis of Glycosylation-The procedures essentially followed those published for analysis of sortilin-1 (15). In brief, transfected cells were biolabeled for 3-5 h in modified Eagle's medium without cysteine and methionine using 200 Ci/ml Pro-mix (Amersham Pharmacia Biotech) in the absence or the presence of 15 g/ml brefeldin A (Roche Molecular Biochemicals). Labeled sorLA protein in cell lysates or medium was immunoprecipitated using anti-sorLA Vps10p domain antibody and GammaBind G beads (AP Biotech) or precipitated using Sepharose beads coupled with RAP or GST-sorLA propeptide. For treatment with PNGase-F (Roche Molecular Biochemicals), washed GammaBind beads with antibody-bound sorLA protein were heated (3 min, 95°C) in 10 l of 1% SDS, heated again after addition of 90 l 20 mM sodium phosphate, 10 mM EDTA, 10 mM sodium azide, 0.5% Triton X-10, pH 7.2, and cooled before the addition of 0.5 unit PNGase-F. Alternatively, washed beads were suspended in 100 l of 50 mM sodium phosphate, 0.01% SDS, 1 M mercaptoethanol, pH 5.5, containing 4 milliunits of endoglycosidase H (Endo-H; Roche Molecular Biochemicals). The reactions were stopped after incubation for 16 h at 30°C followed by analysis by PAGE and autoradiography. The cleavage of cellular sorLA-(1-731) minireceptor by furin was performed as described previously for soluble sortilin-1 (15).
Surface Plasmon Resonance Analysis-All of the measurements were performed on a BIAcore 2000 instrument (Biacore, Uppsala, Sweden) equipped with CM5 sensor chips. The carboxylated dextran matrix of the chip was activated as described (15), and purified sorLA-(54-731) or sorLA-(54-2107) was immobilized to an estimated density of ϳ60 fmol/mm 2 (flow cell 1). Samples for binding (40 l) were injected at 5 l/min at 4°C in 10 mM Hepes, 150 mM NaCl, 1.5 mM CaCl 2 , 1 mM EGTA, 0.005% Tween 20, pH 7.4. Binding was expressed in relative response units as the response obtained from the flow cell with immobilized minireceptor minus the response obtained using an activated but uncoupled chip (flow cell 2). The chip was regenerated by injecting 20 l of 10 mM glycine, 20 mM EDTA, 500 mM NaCl, 0.005% Tween 20, pH 4.0. Kinetic parameters were determined using BI-Aevaluation 3.0 software. The amount of ligand bound/mol immobilized minireceptor was estimated by dividing the ratio RU ligand /mass ligand by RU minireceptor /mass minireceptor .
Immunocytochemistry-For confocal microscopy, cells transfected with full-length sorLA or mock-transfected cells were washed, fixed with 4% paraformaldehyde (15 min), washed in Tris/balanced salt solution, and, when appropriate, permeabilized in the same buffer with 0.5% Triton X-100 (10 min) followed by incubations with primary and secondary antibodies. Primary antibodies were rabbit anti-Vps10 Ig and anti-golgin-97 (Molecular Probes, Eugene OR). Secondary antibodies were anti-rabbit Ig conjugated with Alexa 488 (Molecular probes) and Cy-5 (Zymed Laboratories Inc., San Francisco, CA), respectively. The microscopy was performed using a Zeiss LSM-5 instrument.
Quantification of Cell Surface-expressed SorLA-The cells were washed in ice-cold phosphate-buffered saline, pH 8.0, and incubated with 0.5 mg/ml membrane-impermeable reagent sulfo-N-hydroxysuccinimidobiotin (Pierce) for 90 min at 4°C. After washes in buffer with 50 mM Tris to quench unreacted reagent, the cells were lysed (10 min, 4°C) in 1% Triton X-100, 20 mM Tris, 10 mM EDTA, 150 mM NaCl, pH 8.0, with proteinase inhibitors, and the biotinylated cell surface proteins were precipitated with streptavidin-coupled Sepharose 4B beads (Zymed Laboratories Inc. Lab, CA). The fractions of streptavidin-bound and unbound sorLA were detected by Western blotting using horseradish peroxidase-conjugated swine anti-rabbit Ig (DAKO) as a secondary antibody and ECL (Amersham Pharmacia Biotech). Quantification was performed using a FUJIFILM LAS-1000 plus luminescence image analyzer.
SorLA-mediated Degradation of GST-propeptide-GST-sorLA propeptide was iodinated to about 0.2 mol 125 I/mol protein using chloramine-T as the oxidizing agent and separated from unreacted iodine using Sephadex G25F. The tracer was at least 98% precipitable in trichloroacetic acid. The cells were incubated at 37°C in 250 l of medium with ϳ10.000 cpm 125 I-GST-sorLA propeptide, and an increase in the acid-soluble fraction was taken as a measure of degradation.
Internalization of IL-2R/sorLA Chimera-Transfectants or CHO-K1 control cells were incubated for 2 h at 4°C with 125 I-labeled antibody (3 ϫ 10 4 cpm/ml) directed against the lumenal IL-2R domain (anti-Tac, Roche). Following washings in ice-cold buffer, the cells were reincubated in 37°C warm medium for 0 -60 min. The incubations were stopped at various times by the addition of ice-cold Tris-HCl buffer, pH 2.5. The supernatant was recovered after 5 min, and the cells were lysed in 1 M NaOH. The radioactivity was determined in both fractions and defined as cell surface-associated (i.e. acid-releasable) and internalized ligand, respectively.

RESULTS
SorLA Propeptide Is Removed by Furin-mediated Cleavage-To determine whether sorLA is subject to functional propeptide cleavage, CHO cells were stably transfected with a construct, sorLA-(1-731), comprising the putative propeptide, the Vps10p domain, and the first seven residues of the adjacent YWTD repeat cluster (Fig. 1). The transfectants were biolabeled, and newly synthesized minireceptor secreted into the medium or present in cell lysates was immunoprecipitated. The relatively small product (compared with full-length sorLA) facilitated analysis of molecular sizes. Fig. 2A shows that minireceptor obtained from the medium (lane 1) had a slightly higher apparent molecular weight than the species in the cell lysate (lane 3). To eliminate differences in size caused by dif-ferential glycosylation, N-linked sugars were removed by treatment with PNGase-F. After deglycosylation, the cellular form was larger than the secreted form (lane 4 versus lane 2) signifying cleavage of the minireceptor. This occurred within cells or at the cell surface because labeled minireceptor isolated from cell lysates remained uncleaved upon incubation in conditioned CHO cell medium (not shown). After deglycosylation and treatment with furin, minireceptor from cell lysates achieved the same size as the species in the medium (lane 5 versus lane 2), suggesting that furin is the responsible reagent. These results were confirmed in transfected cells biolabeled in the presence of brefeldin A, which causes retention of newly synthesized proteins in endoplasmatic reticulum (24). Thus, treatment with brefeldin A blocked secretion of the minireceptor (not shown), and following deglycosylation with PNGase-F, treatment with furin caused a truncation of the minireceptor retained within the cells (lane 8 versus lane 7).
Treatment with Endo-H was performed to clarify whether cleavage took part in the distal part of the synthetic pathway. To confirm the importance of furin, we expressed a minireceptor in which the consensus cleavage site was disrupted by point mutations (Fig. 1). Fig. 3A (lanes 2 and 4) shows that mutated minireceptor recovered from medium and lysates of biolabeled transfectants achieved the same size after treatment with PNGase-F in contrast to nonmutated minireceptor treated in parallel (Fig. 3B, lanes 2 and 4). In addition, immunoblotting (Fig. 3C) demonstrated that the mutated but not the wild type product reacted with anti-propeptide antibody, whereas both products were recognized by anti-sorLA Vps10p antibody. When taken together, the results establish that the sorLA minireceptor is synthesized as a proprotein, which is converted to its mature form by furin-mediated cleavage in TGN.
Purification of SorLA Vps10p Domain Minireceptor and L-sorLA-We reasoned that the Vps10p domain of sorLA, like that of sortilin-1 (15), might bind RAP as well as its own propeptide. Minireceptors secreted from biolabeled transfectants were therefore subjected to affinity precipitation using Sepharose beads coupled with RAP or GST-sorLA propeptide. Fig. 4A shows that the wild type minireceptor was precipitated by both RAP (lane 1) and propeptide (lane 2) beads, whereas minireceptor mutated in the furin cleavage site was left in the supernatant to be precipitated with antibodies (Fig. 4B). Analogous results were obtained when using biolabeled L-sorLA comprising the entire lumenal domain (not shown). Wild type sorLA minireceptor comprising the Vps10p domain and L-sorLA were purified by affinity chromatography from the medium of transfectants (Figs. 5A and 7A, insets), and Nterminal sequencing yielded the sequence 54 SAALQ, confirming that mature receptor had been cleaved immediately after the consensus motif RRKR 53 . To elucidate the domain specificity of ligands previously reported to bind sortilin-1 (Fig. 1) or sorLA, we immobilized the Vps10p domains sorLA-(54 -731) and soluble sortilin-1-(45-725) (15) as well as L-sorLA-(54 -2107) onto chips for surface plasmon resonance analysis.
Comparison of Ligand Binding to SorLA and Sortilin-1 Vps10p Domains-The first aim was to characterize propeptide binding to the sorLA Vps10p domain. Fig. 5A shows that the signal (response units) obtained with 100 nM GST-sorLA propeptide (pH 7.4) was about 6-fold higher than that obtained with 100 nM 53-residue sorLA propeptide. This is in accordance with the molecular weight ratio of about 32/6 when regarding the GST-propeptide as a monomer. Binding of GST-propeptide was abolished at pH 5.0 (not shown), and 1 M GST alone did not bind. The K d value for propeptide binding with or without GST was calculated at 5-15 nM from a series of curves obtained at pH 7.4 with concentrations ranging from 5 nM to 10 M (not shown). When using a saturating concentration (10 M), 0.6 mol of GST-sorLA propeptide was bound per mol immobilized minireceptor. This suggests a 1:1 binding stoichiometry because some immobilized molecules are likely to be unavailable for binding.
Prompted by the structural similarities between the sortilin-1 and sorLA Vps10p domains, we next probed for binding of the GST-sortilin-1 propeptide to the chip with the sorLA Vps10p domain. As shown in Fig. 5B, the signal obtained at 100 nM GST-sortilin-1 propeptide was similar to that obtained at the same concentration of GST-sorLA propeptide, and analysis of a series of curves (not shown) confirmed similar binding affinities. Moreover, 100 nM sorLA propeptide reduced the signal obtained with GST-sortilin propeptide almost to the level obtained with sorLA propeptide alone (Fig. 5B). The reverse experiment was also performed, and we found that GST-sorLA propeptide binds to soluble sortilin-1 with a K d of about 10 nM (not shown). The results demonstrate that sorLA and sortilin propeptides bind to the Vps10p domains of sorLA and sortilin with about equal affinities. 10 M RAP, thereby confirming competition for binding to the sorLA Vps10p domain. These results are similar to those obtained previously for binding of RAP and sortilin-1 propeptide to the sortilin-1 Vps10p domain, except that RAP shows a lower affinity to the sorLA Vps10p domain than to the sortilin-1 Vps10p domain (15).
Neurotensin was tested because sortilin-1 is an established receptor for this 13-residue neuropeptide (15,25). Fig. 6B shows that binding of 1 M RAP was inhibited about 60% by 20 M neurotensin. This result and the size of the signal obtained with 20 M neurotensin alone are in broad agreement with previous results obtained with soluble sortilin-1 (15). The binding of GST-sorLA and GST-sortilin-1 propeptides were inhibited 20 and 50%, respectively, by 20 M neurotensin, and in control experiments neither RAP nor propeptide binding was inhibited by 20 M bacitracin (not shown). The K d value for binding of neurotensin to sorLA minireceptor was estimated at ϳ30 nM from separate experiments with concentrations ranging from 5 nM to 20 M, suggesting an affinity slightly lower than that for binding to sortilin-1.
The head activator peptide was tested because a homobipeptide binds with high affinity to a component in lysates of NT2 cells compatible with sorLA (16). The experiments were performed using the monomerized head activator reported to stimulate mitosis of NT2 cells at 2 nM concentration (16), as well as  1a and 2a). The precipitates were analyzed by reducing SDS-PAGE and autoradiography of 2,5diphenyloxazole-impregnated gels. peptide dissolved directly in buffer. Binding of 100 nM GSTpropeptide or 1 M RAP was 50% inhibited by 20 M head activator, and K d for direct binding of the peptide was ϳ500 nM, although no accurate value could be measured because of the small size of the signal (not shown). Binding to the sortilin-1 Vps10p domain was indistinguishable from that to the sorLA Vps10p domain, and we observed no difference between binding of the monomerized peptide and the peptide dissolved directly in buffer (not shown).
apoE was tested because previous results have shown binding to full-length sorLA (2), and LpL was tested because it binds to sortilin-1 (17). However, binding of these ligands could not be demonstrated to the sorLA Vps10p domain (not shown).
The results show that the sorLA and sortilin Vps10p domains bind propeptides equally well and with high affinity. RAP and neurotensin also bind to both Vps10p domains, although with lower affinity to the sorLA Vps10p domain. The head activator binds to both domains with low affinity, whereas apoE and LpL exhibit insignificant binding to the sorLA Vps10p domain.
Comparison of Ligand Binding to SorLA Vps10p Domain and to L-sorLA-This was performed because sorLA, in contrast to sortilin-1, comprises alternative domains including a cluster of LA repeats similar to the ligand-binding repeats in LDLR family members. The results (not shown) demonstrated identical binding of propeptides, neurotensin, and the head activator to the two receptor constructs, signifying that binding of these ligands occurs only to the Vps10p domain. On the other hand, RAP displayed much higher binding to L-sorLA than to the isolated Vps10p domain, and apoE and LpL also bound avidly to L-sorLA. Fig. 7A shows two components of RAP binding to L-sorLA. The minor rapidly dissociating component is compatible with binding to the Vps10p domain (K d ϳ100 nM; Fig. 6), whereas the slowly dissociating component must be due to binding to other sorLA domains, most likely to the LA repeat cluster. The overall K d for high affinity RAP binding was estimated at ϳ0.1 nM from a series of curves, and this binding was not inhibited by the sorLA propeptide (not shown). Recombinant apoE (Fig. 7B) and LpL (not shown) bound avidly to L-sorLA, and the K d value for apoE binding was calculated to be Ͻ1 nM. When taken together, the results suggest that sorLA is a multifunctional receptor capable of binding several ligands to distinct sites located in at least two domains.
SorLA Is Mainly in the Golgi Compartment and ϳ10% on the Cell Surface- Fig. 8A shows the localization of full-length sorLA in permeabilized CHO cell transfectants stained with anti-Vps10p antibody. It appears that sorLA is mainly intracellular and colocalized with the Golgi autoantigen golgin-97 in paranuclear compartments (Fig. 8, A versus B), signifying predominant Golgi localization. Fig. 8C shows a minor expression on the surface of nonpermeabilized transfectants. No staining was observed in mock-transfected cells when using the anti-Vps10p antibody (not shown). To estimate the level of cell surface expression, the transfectants were treated with the nonpermeable reagent sulfo-N-hydroxysuccinylimidobiotin at 4°C and lysed, and biotinylated proteins were recovered on streptavidin-Sepharose beads. Subsequently, samples of precipitated and of nonprecipitated protein were subjected to SDS-PAGE, and their relative content of sorLA was determined by Western blotting. Fig. 8D shows that biotinylated (lane 1) and nonbiotinylated (lane 2) sorLA was readily detected, and scanning densitometry (see legend) revealed that about 10% of sorLA had been accessible to biotinylation and therefore represent the fraction of receptors expressed on the cell surface. As a control, we show that biotinylated sorLA did not bind to Sepharose beads without streptavidin but remained in the unbound fraction (lanes 3 and 4). A double band was observed on the Western blots in agreement with the pattern originally observed in sorLA isolated from brain (1). Because this might be due to differential glycosylation, we immunoprecipitated sorLA from metabolically labeled cells followed by treatment with PNGase-F and analysis by SDS-PAGE. As shown in Fig.  8D (lane 5 versus lane 6), deglycosylation resulted in a single band, signifying that the antibodies reacted with a single but differentially glycosylated protein.
SorLA Mediates Endocytosis-GST-sorLA propeptide was used to assess whether cell surface expressed sorLA can mediate degradation of extracellular ligand. As shown in Fig. 9A, the cells transfected with full-length sorLA degraded 125 I-GSTpropeptide faster than control cells, and degradation was not observed in conditioned medium. Unlabeled GST-propeptide, but not GST alone (not shown), inhibited degradation of labeled GST-propeptide half-maximally at about 300 nM (Fig. 9A), and RAP and neurotensin at high concentrations (20 M) inhibited propeptide degradation by about 50 and 30%, respectively (not shown). In addition, degradation was inhibited by the weak base chloroquine that raises pH in intracellular compartments (Fig. 9A).
To measure internalization of cell surface bound ligand, we initially incubated sorLA transfectants with 125 I-GST-propeptide at 4°C followed by incubation at 37°C and assessment, at various times, of acid-releasable and nonreleasable (i.e. internalized) radioactivity. However, the results were difficult to assess because of binding of GST-propeptide to surface structures other than sorLA (not shown). We therefore used cells transfected with the IL-2R/sorLA chimera consisting of the transmembrane and cytoplasmic domains of sorLA and the lumenal domain of IL-2R (Tac/CD25), thus taking advantage of available highly specific anti-Tac antibodies (22). 125 I-Labeled anti-Tac bound to the IL-2R/sorLA transfectants at 4°C but not to mock transfected cells (not shown). As shown in Fig. 9B, about 60% of the 125 I-anti-Tac, initially bound to the transfectants at 4°C, was internalized after incubation for 15 min at 37°C. These results show that sorLA can mediate ligand uptake and degradation, presumably in lysosomes. DISCUSSION The results show that sorLA is synthesized as an inactive proreceptor that is converted to the mature ligand-binding form by cleavage of a 53-residue N-terminal propeptide in the late Golgi compartment/TGN. This is in accordance with the recent observation that sorLA immunoprecipitated from metabolically labeled NT2 cells, when deglycosylated, is initially in a slightly larger form than that obtained after a chase (16). Furin is likely to be the main enzyme responsible for cleavage because it is abundant in the Golgi compartment/TGN. However, it is possible that other enzymes may also participate in the conversion in vivo because proconvertases of the subtilisin/ Kex-2-like family have overlapping substrate specificities. We also show that the isolated propeptide binds to mature sorLA in a pH-dependent way and inhibits binding of other ligands to the Vps10p domain. To achieve a fully active sorLA, the propeptide must therefore be removed, most likely in an acidic compartment. The mechanism of activation is analogous to that described for sortilin-1 (15) and may characterize all members of the mammalian Vps10p receptor family. In addition to sorLA and sortilin-1, the family comprises sorCS/sortilin-2 (8) and two as yet uncharacterized family members (26,27), tentatively designated sortilin-3 and -4 (accession numbers AB028982 and AB037750), both of which have N-terminally located consensus sites for cleavage by furin.
Putative Role of Propeptide and RAP-We show that mature sorLA has at least two lumenal domains capable of binding multiple ligands. One is the Vps10p domain, and the other is most likely the LA repeat cluster similar to ligand-binding domains in LDLR family members. The function of both propeptide and RAP may be to assist in protein folding and/or to prevent aggregation provoked by premature ligand binding in the early synthetic pathway.
The propeptide, when still attached to sorLA, prevents binding to the Vps10p domain. Surprisingly, we find that isolated sorLA and sortilin-1 propeptides, which share no obvious structural similarities, bind to mature sorLA and sortilin-1 Vps10p domains with equal affinity. This raises the possibility that propeptides may cross-bind on several Vps10p receptors and modulate binding of external ligands.
Mature sorLA Vps10p domain also binds RAP, although with moderate affinity, as previously shown for sortilin-1 (15). This is surprising because RAP is an endoplasmatic reticulum resident protein thought to be essentially absent in furin containing compartments (for review, see Ref. 28). Although the functional significance is not understood, the finding that RAP and propeptides cross-compete for binding to sorLA and sortilin-1 Vps10p domains reflects a common theme in recognition that may be shared among Vps10p family receptors. However, RAP binds with a much higher affinity to the cluster of 11 LA repeats than to the Vps10p domain. Recent studies on LRP have shown that high affinity RAP binding depends on pairs of LA repeats each harboring surface exposed acidic residues between Cys IV and Cys V of the ϳ40 residue repeats (29). SorLA has five LA repeats with exposed Asp/Glu residues (four in consecutive repeats) that may explain the RAP binding. In contrast to the Vps10p domain, the LA repeat cluster is exposed for premature ligand binding in the early synthetic pathway, and RAP may protect the LA repeat cluster and thereby prevent receptor aggregation as previously reported for LRP (28). In addition, RAP may function as a folding chaperone for sorLA as suggested for members of the LDLR family (30,31).
Vps10p Domain Ligands-Our results show overlapping, but not identical, specificities for binding to mature sortilin-1 and sorLA Vps10p domains. Thus, LpL (17) and apoE 2 bind to sortilin-1 but not to the sorLA Vps10p domain (present results), whereas neurotensin binds to both domains. Sortilin-1, also designated neurotensin receptor-3 (15,25), exhibits basically low expression on the surface of neuronal cells. Upon secretion, neurotensin binds to the G-protein-coupled neurotensin receptors-1 and -2, and as a result, sortilin-1 is translocated to the cell surface and appears to participate in the scavenging of neurotensin (25). Future studies should show whether sorLA has a similar role in termination of neurotensin signaling or may help in presenting neurotensin on the cell surface.
The result that sorLA binds the monomeric head activator peptide with only a low affinity is surprising because the head activator homobipeptide binds to a component in lysates of NT2 cells with a K d value of 2-3 nM (16). One possible explanation is that the affinity of the homobipeptide, which is an artificial ligand, is much higher than that of monomeric peptide. However, this seems unlikely because 2 nM monomeric head activator causes a marked stimulation in mitosis of NT2 cells (16). Although the question remains unresolved, it seems most likely that the head activator peptide, to achieve high affinity for sorLA, has to interact with a component not present in the purified system. In addition, the finding that sortilin-1 also displays low affinity binding of the head activator opens the possibility that this property may be shared among several Vps10p receptors.
LA Repeat Cluster Ligands-The results show high affinity binding of the three ligands previously shown to interact with all members of the mammalian LDLR family: RAP, apoE, and LpL (13,14). The main point is that sorLA has a domain with overall binding activity comparable with that of classical LDLR family receptors. The large members LRP and megalin, as well as the very low density lipoprotein receptor, bind more than 30 structurally distinct ligands. These receptors have overlapping specificities as well as ligands of their own (13,14), and future studies should specify the pattern for the binding of ligands to sorLA.
Putative Function in Cells-The distribution of sorLA with ϳ10% on the cell surface and ϳ90% in perinuclear compartments is similar to that of sortilin-1 (7,17) and of mannose 6-phosphate receptors, which mediate both sorting of newly synthesized ligands and uptake from the cell surface. This is in accordance with the presence of acidic clusters in the cytoplasmic domains of these receptors (Asp 2162 -Asp 2170 in sorLA) that may function as determinants for TGN localization (32). We show that sorLA on the cell surface mediates uptake and degradation of bound ligand similar to sortilin-1 (17), mannose 6-phosphate receptors, and LDLR family receptors. In sorLA, internalization may depend on Phe 2144 -Tyr 2149 , which is in accordance with the overall internalization motif (F/Y)XXXX(F/Y), or on the acidic cluster. In view of its predominant cellular localization, it may be proposed that sorLA can also function as a sorting receptor following interaction with ligands in the Golgi compartment/TGN where furin mediates cleavage and activation of the Vps10p domain and where RAP has been removed and retrieved to the endoplasmatic reticulum.
In conclusion, we show that the Vps10p domain of sorLA is activated by cleavage in the late Golgi compartment/TGN and that cell surface-expressed sorLA can mediate uptake and degradation of bound ligand. We demonstrate the overlapping specificities of the sorLA and sortilin-1 Vps10p domains for binding of propeptides and neuropeptides and show that sorLA, in addition, has an overall binding activity comparable with that of the LDLR family receptors.