Molecular Identification of a Novel Candidate Sorting Receptor Purified from Human Brain by Receptor-associated Protein Affinity Chromatography*

(Received for publication, July 16, 1996, and in revised form, November 11, 1996)

Claus M. Petersen Dagger §, Morten S. Nielsen Dagger , Anders Nykjær Dagger , Linda Jacobsen Dagger , Niels Tommerup , Hanne H. Rasmussen Dagger , Hans Røigaard Dagger , Jørgen Gliemann Dagger , Peder Madsen Dagger and Søren K. Moestrup Dagger

From the Dagger  Department of Medical Biochemistry, University of Aarhus, 8000 Aarhus C, Denmark and  The John F. Kennedy Institute, DK-2600 Glostrup, Denmark

ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
Acknowledgments
REFERENCES


ABSTRACT

Receptor-associated protein (RAP) is an endoplasmic reticulum/Golgi protein involved in the processing of receptors of the low density lipoprotein receptor family. A ~95-kDa membrane glycoprotein, designated gp95/sortilin, was purified from human brain extracts by RAP affinity chromatography and cloned in a human cDNA library. The gene maps to chromosome 1p and encodes an 833-amino acid type I receptor containing an N-terminal furin cleavage site immediately preceding the N terminus determined in the purified protein. Gp95/sortilin is expressed in several tissues including brain, spinal cord, and testis. Gp95/sortilin is not related to the low density lipoprotein receptor family but shows intriguing homologies to established sorting receptors: a 140-amino acid lumenal segment of sortilin representing a hitherto unrecognized type of extracellular module shows extensive homology to corresponding segments in each of the two lumenal domains of yeast Vps10p, and the extreme C terminus of the cytoplasmic tail of sortilin contains the casein kinase phosphorylation consensus site and an adjacent dileucine sorting motif that mediate assembly protein-1 binding and lysosomal sorting of the mannose-6-phosphate receptors. Expression of a chimeric receptor containing the cytoplasmic tail of gp95/sortilin demonstrates evidence that the tail conveys colocalization with the cation-independent mannose6-phosphate receptor in endosomes and the Golgi compartment.


INTRODUCTION

Sorting of newly synthesized lysosomal enzymes from the Golgi compartment to late endosomes in eukaryotic cells is a sophisticated transport process involving specific sorting receptors in the trans-Golgi network. In mammals, the 46- and 275-kDa mannose-6-phosphate (M6P)1 receptors are the known sorting receptors that bind to phosphorylated mannose residues in lysosomal hydrolases (1). In yeast, a M6P-independent sorting pathway has been demonstrated by identification of the vacuolar protein-sorting 10 protein (Vps10p) (2) and a highly homologous protein encoded by the yeast VTH2 gene (3). Both are capable of targeting yeast carboxypeptidase Y to lysosomes (2, 3). Mammalian counterparts to these sorting receptors have so far not been identified. However, studies of I-cell disease patients suggest that mammals may sort lysosomal enzymes by alternative mechanisms (4-9).

The 40-kDa endoplasmic reticulum/Golgi receptor-associated protein (RAP) assists folding and processing of the cysteine-rich low density lipoprotein (LDL) receptor class A repeats in receptors of the LDL receptor family (10-13). In addition to the high affinity binding to the LDL receptor family proteins and the newly identified LDL receptor type A repeat containing receptor sorLA/LR11 (14, 15), RAP binds calmodulin and is phosphorylated by calmodulin-dependent kinase II and casein kinase II (16). Recently, independent observations have shown the binding of RAP to an approximately 100-kDa protein expressed in osteosarcoma (17) and Chinese hamster ovary cells (18).

In the present study we have identified, purified, and characterized a ~95-kDa protein by RAP affinity chromatography of membrane protein extracts from human brain. Cloning of the encoding gene by cDNA library screening disclosed a type I receptor structure and striking sequence homologies with yeast Vps10p and the cation-dependent (CD) and cation-independent (CI) M6P receptors. Transfection analyses of the novel receptor, which represents the first known RAP binding receptor not related to the LDL receptor family, disclosed that the cytoplasmic tail conveys localization to the Golgi and CI-M6P receptor-rich vesicles.


EXPERIMENTAL PROCEDURES

Preparation of Brain Membranes and RAP Affinity Chromatography

Human cadaver brain tissue samples were, with permission, obtained from autopsy cases 12-24 h postmortem, and crude membrane fractions were generated, solubilized in 1% CHAPS (Boehringer, Mannheim, Germany) dissolved in 2 mM CaCl2, 1 mM MgCl2, 10 mM Hepes, and 140 mM NaCl, pH 7.8 (buffer A) and applied to RAP affinity chromatography (16). After washing, protein retained on the column was eluted at pH 4.0 in phosphate-buffered saline containing 10 mM EDTA and 0.25% CHAPS. The eluted fractions (1 ml) were neutralized by Tris-base, analyzed by SDS-PAGE, and finally pooled and concentrated using a centricon-10 from Amicon (Beverly, MA).

SDS-PAGE, Two-dimensional Gel Electrophoresis, Deglycosylation, and Blotting

SDS-PAGE was performed in a Laemmli system using a 4% stacking gel and a 4-16% gradient separation gel. Two-dimensional gel electrophoresis was done as described (16). Deglycosylation was performed using peptide-N-glycosidase-F (19). Molecular weight markers were from Sigma or Novex (San Diego, CA). For ligand/immunoblotting cells, membrane preparations and whole tissue samples were lysed for 10 min at 4 °C in 1% Triton X-100 or Triton X-114, 10 µg/ml aprotinin, 2 mM phenylmethylsulfonyl fluoride, and 0.1 M Tris/HCl (pH 8). After centrifugation (5 min, 104 × g, 4 °C), the pellet was discarded and supernatants were stored at -20 °C. Ligand blotting was performed as described using 5 × 104 cpm/ml of 125I-RAP or 125I-urokinase in complex with plasminogen activator inhibitor-1. Iodinated ligands were prepared as described previously (20).

For immunoblotting, Immobilon membranes were blocked in 2% Tween 20 buffer (pH 7.4) and incubated with antibodies in 0.2% defatted milk powder, 0.05% Tween 20, and buffer A (pH 7.4).

Amino Acid Sequencing

Pooled fractions of RAP affinity-purified proteins were resolved by SDS-PAGE and transferred to Immobilon filters by electroblotting in 50 mM Tris-base, 50 mM boric acid buffer. Membrane-bound proteins were stained using 0.1% amido black in 45% methanol, 9% acetic acid, and the respective bands were cut from the membrane, incubated (15 min) in water containing 0.2% polyvinylpyrrolidone, washed twice in water, and cleaved in situ by overnight incubation at 37 °C in 200 µl digestion buffer (ammonium bicarbonate, 0.01% polyvinylpyrrolidone, 0.1 M Ca2+) supplemented with 1 µg of trypsin (modified trypsin, sequencing grade; Promega). The digestion mixture (supernatant) was recovered and acidified by addition of 5-10 µl of trifluoroacetic acid, and the membrane was finally washed twice in distilled water. The washing solutions were added to the digestion mixture, and the peptide separation was performed on an applied Biosystems microbore high pressure liquid chromatography (model 172) with a 500-µl injection loop. The digests were fractionated on a 2 × 100-mm, 5 µm Nucleosil 18 reverse phase column. Peptides were eluted in a linear gradient of acetonitrile and 0.1% trifluoroacetic acid at a flow rate of 100 µl/min and detected at 214 nm. Amino acid sequencing of selected peptides and N-terminal sequencing of undigested protein was performed in a 477 gas-phase sequenator connected on line to a 120A PTH-analyzer. The sequenator was operated according to the manufacturer's instructions with previously described modifications (21).

cDNA Cloning and Sequencing

A degenerated oligodeoxyribonucleotide probe (5'-GA(A/S)TT(C/T)GGIATGGCIAT(A/C/T)GG-3'), derived from the peptide sequence EFGMAIG and containing deoxyinosine at the most ambiguous positions, was used to screen a Jurkat cDNA library established in the ZAP Express vector (Stratagene). Following 5'-end labeling with [gamma -32P]ATP, the probe was used to screen four sets of replicate nylon filters (22 × 22 cm, Amersham), each containing 1.25 × 105 plaque-forming units. The filters were hybridized at 50 °C using the tetramethylammonium chloride technique (22, 23). Three positive clones were identified, purified, and rescued into the pBK-CMV vector according to the manual from Stratagene. Further analysis of the purified clones was carried out by dot and Southern blotting using the original cloning probe and three additional, fully degenerated, oligodeoxyribonucleotide probes derived from peptides PGEDEE, DFVAKL, and EEHLTT. Two of the clones reacted with all four probes, and one of these, containing an insert of 4.5 kb, was selected for sequence analysis. The insert was recovered by EcoRI cleavage yielding two fragments. For enzyme restriction analysis, the fragments were subcloned into the M13BM20/21 vector, and single-stranded DNA was sequenced by dideoxy sequencing using a T7 sequencing kit (Pharmacia). The partial sequences were used to construct 17mer oligonucleotides allowing consecutive sequencing of both strands. As the 500-base pair upstream G/C-rich sequence contained several compressions, this stretch was sequenced using a T7 inosine sequencing kit (U. S. Biochemical Corp.).

Transient Expression and Immunocytochemistry

The 5' fragment of the selected gp95/sortilin cDNA, spanning from the XbaI site of the pBK-CMV vector to the SmaI site at position 3590, was cloned into the eukaryotic expression vector pMT21 (24). A cDNA encoding a chimeric receptor was constructed by generating a 3' HindIII site and 5' XhoI site in the gp95/sortilin cDNA encoding the cytoplasmic tail (aa 748-800) and inserting this sequence into the previously described plasmid vector (25) (kindly donated by Dr. Susan E. LaFlamme) encoding the extracellular and the transmembrane domain of the interleukin 2 (IL-2) receptor (Tac/CD25). Transfection of the pMT21-gp95/sortilin, the chimeric receptor construct, and control plasmid into COS-1 cells was carried out using LipofectAMINE (Life Technologies, Inc.) as transfection agent (26). 48 h after transfection, the cells cultured in Petri dishes were washed and lysed in 1% Triton X-100 for ligand and Western blotting. For confocal and fluorescence microscopy, transfected cells were washed in 10 mM phosphate, 150 mM NaCl, pH 7.3 (buffer B), fixed in buffer B containing 3.7% formaldehyde (10 min), and permeabilized in the same solution with 1% Triton X-100 (1 min). Fixed cells were washed in 0.05% Tween 20 buffer B, and incubations with primary and secondary antibodies were carried out in antibody diluent (DAKO, Denmark). Nuclei were stained by incubation with 4',6-diamino-2-phenylindole dihydrochloride hydrate (DAPI; Sigma). For surface labeling, the cells were incubated at 4 °C with the primary antibody prior to fixation in formaldehyde. Horseradish peroxidase-, fluorescein isothiocyanate-, and rhodamine-conjugated anti-rabbit/anti-mouse antibodies were from DAKO. Rabbit anti-gp95/sortilin peptide anti-serum was generated (Neosystem, Strasburg, France) using a synthetic peptide (SAPGEDEEGGRVRDFVAKLA) based on the N-terminal sequence obtained from purified gp95/sortilin. Mouse anti-CD25 was from Boehringer. Metabolic labeling of transfected cells was done overnight in cysteine- and methionine-free medium supplemented with L-[35S]methionine and L-[35S]cysteine (Pro-mix; Amersham) and 10% full medium. Labeled cells were lysed in 1% Triton X-114, Tris/HCl (pH 8.0), and following phase separation, proteins were precipitated from the fluid phase (in 0.1% CHAPS buffer A) using RAP immobilized on Sepharose beads and uncoated beads for control. Precipitated protein was analyzed by reducing SDS-PAGE, and diphenyloxazole-fluorographed gels were exposed at -70 °C.

Northern Blotting

Northern blotting was performed using tissue blots 7760-1, 7759-1, and 7767-1 (Clontech) containing 2 µg of poly(A)+ purified RNA from different human tissues. Hybridizations were carried out with a random-primed probe including 1581 nucleotides (an EcoRI fragment) of the sortilin clone. Washing was done at high stringency conditions according to standard procedure, and results were analyzed by autoradiography.

Chromosome Mapping

Fluorescence in situ hybridization using the biotin-labeled 4.5-kb clone with corresponding DAPI banding and measurement of the relative distance from the short arm telomere to the signals (FLpter value) was performed as described previously (27).


RESULTS

Purification of RAP Binding Proteins from Human Brain Membranes

A ~95-kDa RAP binding protein (designated gp95/sortilin) was identified by ligand blotting of human brain tissue lysates (Fig. 1A) and purified by RAP affinity chromatography. Fig. 1B presents a non-reducing SDS-PAGE profile of the proteins eluting from the RAP column and shows that in addition to high molecular weight proteins (>200,000), including the 515-kDa alpha -chain of LDL receptor-related protein, a single protein of the expected molecular mass of ~95 kDa was released from the column. As determined by ligand blotting, both the high molecular weight proteins and gp95/sortilin bound RAP (not demonstrated). Fig. 2A shows that 125I-RAP binding to gp95/sortilin was inhibited by heparin and completely abolished in the presence of 100 nM unlabeled RAP or (not shown) 10 mM EDTA. Furthermore, binding of 125I-RAP was not inhibited by urokinase in complex with plasminogen activator inhibitor type-1 (200 nM), and 125I-labeled urokinase-inhibitor complex did not bind, verifying that the RAP binding protein was different from the very low density lipoprotein receptor. To ensure that the observed RAP binding was accounted for by a single protein, electroeluted gp95/sortilin was subjected to reducing SDS-PAGE and two-dimensional gel electrophoresis. As demonstrated in Fig. 2B, reduction did not significantly alter the mobility of gp95/sortilin and did not give rise to additional protein bands. Results where similar when using two-dimensional gels (not shown), which presented gp95/sortilin as a streak of different pI species (pI 4.5-5.0) in accordance with differences in glycosylation. peptide-N-glycosidase-F treatment indicated that glycosyl groups account for about 10% of the mass of the native protein (Fig. 2B).


Fig. 1. Purification of RAP binding proteins by RAP affinity chromatography. A, human brain membranes were lysed in buffer A with 1% CHAPS. Following nonreducing SDS-PAGE (4-16%) and electrotransfer to an Immobilon membrane, the membrane was blocked and incubated overnight at 4 °C in buffer A supplemented with 1% bovine serum albumin and 0.1 nM (105 cpm/ml) 125I-RAP, washed, and subjected to autoradiography. B, brain membrane lysates were diluted to 0.2% CHAPS in buffer A and recycled overnight on a column of Sepharose-immobilized RAP. Following wash, proteins retained on the column were eluted by lowering pH to 4.0 and adding 10 mM EDTA to the buffer. 15 µl of the 1-ml fractions were subjected to nonreducing SDS-PAGE (4-16%, silver staining). The asterisk indicates the ~250-kDa 125I-RAP binding protein, SorLA (14).
[View Larger Version of this Image (67K GIF file)]



Fig. 2. RAP binding and deglycosylation of purified gp95/sortilin. A, affinity-purified gp95/sortilin was cut from Coomassie-stained gels and electroeluted. The electroeluted protein was subjected to 125I-RAP blotting as described in the legend to Fig. 1 in the absence (lane 1) or presence of 100 units heparin (lane 2) and 100 nM unlabeled RAP (lane 3). B, SDS-PAGE of electroeluted gp95/sortilin unreduced (lane 1), reduced (lane 2), or reduced and peptide-N-glycosidase-F-treated (lane 3). A silver-stained 4-16% gradient gel is shown.
[View Larger Version of this Image (53K GIF file)]


N-terminal Sequencing and Antibody Generation

N-terminal sequencing of purified gp95/sortilin produced a single 20-amino acid sequence (Ser45-Ala64, Fig. 4) that showed no homology to any known protein sequence in the EMBL or Genbank data bases. A synthetic peptide, derived from the sequence obtained by the N-terminal sequencing, was used for generation of rabbit antibodies. As demonstrated in Fig. 3, the anti-peptide antibody reacted with gp95/sortilin but not with any of the other proteins eluting from the RAP affinity column (panel B). Moreover, the antibody bound to electroeluted gp95/sortilin (Fig. 3A, shown to bind RAP in Fig. 2A) and, like 125I-RAP, to a ~95-kDa protein in crude lysates of brain membranes (Fig. 3C). Thus, the combined findings identify gp95/sortilin as a hitherto undescribed RAP-binding membrane glycoprotein.


Fig. 4. Primary sequence of the 95-kDa protein as deduced by cDNA cloning. The probable endoplasmic reticulum import signal sequence is shown in italic. The hydrophobic transmembrane segment is in bold. A cysteine-rich region with sequence similarity and similar spacing of cysteines as the two similar regions in yeast Vps10p is shaded. Putative sorting motifs in the cytoplasmic region are boxed. Potential glycosylation sites are indicated by asterisks. The sequences verified by protein sequencing are underlined. Arrow indicates the predicted (28) signal cleavage site.
[View Larger Version of this Image (59K GIF file)]



Fig. 3. Western and ligand blotting of brain gp95/sortilin. Panel A shows Western blotting of purified and electroeluted gp95/sortilin (same as in Fig. 2). Panel B shows a silver stain (lane 1) and a Western blot of a same sample of RAP affinity-purified protein (lane 2), Panel C shows 125I-RAP binding (lane 1) and a Western blot (lane 2) of crude lysates of human brain membranes. Blotting with 125I-RAP was performed as described in the legend to Fig. 1. For Western blotting, blocked membranes were incubated with anti-gp95/sortilin peptide serum (1:200), and peroxidase-conjugated swine anti-rabbit antibody was used as secondary antibody.
[View Larger Version of this Image (60K GIF file)]


Sequence Analysis of gp95/Sortilin

To obtain information of the primary sequence of gp95/sortilin, N-terminal sequences of tryptic fragments isolated by reverse phase chromatography were determined. 90 residues dispersed over 11 fragments were identified, and four sequences were selected for construction of oligonucleotides. A cDNA clone containing a 4.5-kb insert hybridizing with all four oligonucleotides was isolated from a human Jurkat cDNA library and sequenced. The sequence presents a single open reading frame encoding a 33-aa signal peptide and a 800-aa protein (Fig. 4) with a predicted mass of 88.8 kDa and a pI of 5.7. The amino acid sequences obtained from the tryptic fragments were all identified in the translated sequence (underlined residues in Fig. 4). As seen in Fig. 4, the translated protein has a typical type I receptor structure with two hydrophobic regions. The N terminus (aa -33 to -1) constitutes a potential endoplasmic reticulum import signal sequence with a hydrophobic core and a predicted cleavage site at Gln1 (28). The second hydrophobic stretch (aa 759-780) in the C-terminal region constitutes the putative transmembrane segment. These predictions suggest that the synthesized protein consists of a lumenal domain of 726 amino acids with 6 potential N-linked glycosylation sites and a short cytoplasmic tail of 53 residues. Interestingly, the peptide sequence Ser45-Ala64, which was determined as the N terminus of the affinity-purified protein, is not the N terminus of the protein sequence deduced from the cDNA sequence. Instead, this sequence is C-terminal to the RWRR (aa 41-44) sequence that abides by the basic consensus motif Arg-4-Xaa-3-Arg-2/Lys-2-Arg-1-Xaa+1 for cleavage (between positions -1 and +1) by the trans-Golgi proteinase furin (29). Consequently, these data suggest that gp95/sortilin is synthesized as a precursor, which is processed into a 5-kDa peptide and a 95-kDa membrane protein.

The encoding gp95/sortilin gene (designated sort1) was mapped to the proximal part of the short arm of chromosome 1 as determined by the fluorescence in situ hybridization technique (not shown). At least one signal was displayed in 30 of 50 analyzed methaphases with 46 of the 200 chromatids (23%) being labeled. The FLpter value was 0.41 ± 0.04, corresponding to a localization at 1p13.1-p21.3.

Sequence Alignments

Comparison of the protein sequence with known sequences of the Genbank and EMBL data bases revealed no similarity to the LDL receptor but striking sequence homologies with the 210-kDa yeast receptor for carboxypeptidase Y sorting and with the M6P receptors, in particular the CI-M6P receptor.

The 1371-aa lumenal domain of Vps10p consists of two approximately 685-aa domains of 20% identity. Alignment of the entire lumenal domain of the 95-kDa protein with Vps10p shows a similar degree of identity to both Vps10p domains. Moreover, a particular high degree of homology is seen between the three cysteine-rich regions comprised by Vps10p aa 562-693, Vps10p aa 1225-1355, and gp95/sortilin aa 567-709. Alignment of the three sequences (Fig. 5A) shows a pronounced similarity of these regions and an almost identical spacing of ten cysteines. The structural homology between these "ten cysteine consensus" (10CC) regions strongly indicates that they represent a hitherto unrecognized type of extracellular module.


Fig. 5. gp95/sortilin homologies. A, alignment of 10 cysteine consensus (10CC) regions in gp95/sortilin and Vps10p. B, alignment of the C terminus of gp95/sortilin and the M6P receptors. Identical residues (closed) and conservative changes (stippled) are indicated around the corresponding amino acids.
[View Larger Version of this Image (48K GIF file)]


The homology with the 275-kDa CI-M6P receptor (Fig. 5B) is seen in the cytoplasmic tail of gp95/sortilin, which in the C terminus contains a sequence (HDDSDEDLL) identical with the phoshorylation site and sorting signal in the C terminus of the CI-M6P receptor and very similar (Fig. 5B) to the corresponding stretch in the CD-M6P receptor. Fig. 6 is a schematic presentation of the structural similarities between gp95/sortilin, yeast Vps10p, and the CI-M6P receptor.


Fig. 6. Schematic presentation of similarity between gp95/sortilin, yeast Vps10p, and the human CI-M6P receptor.
[View Larger Version of this Image (33K GIF file)]


Transient Expression of gp95/Sortilin and Chimeric Receptors

The sequence encoding gp95/sortilin was cleaved from the cloned cDNA and inserted into a pMT21 vector for transfection of COS-1 cells. Fig. 7A shows an anti-gp95/sortilin peptide antibody staining demonstrating a high level of gp95/sortilin expression in transfected COS-1 cells. The expression of gp95/sortilin was confirmed by Western blotting performed on lysates of transfected cells (Fig. 7B) and RAP affinity precipitation of a ~95-kDa expression product from transfected biolabeled cells (Fig. 7C). Confocal lasermicroscopy of transiently expressed full-length protein in COS-1 cells demonstrated localization of gp95/sortilin in endoplasmic reticulum and, in particular, in the Golgi compartment. Staining was not detectable on the cell surface (not shown).


Fig. 7. Transient expression of gp95/sortilin in COS-1 cells. A, COS-1 cells were transfected using pMT21 inserted with the gp95/sortilin-encoding sequence. Following 48 h of incubation, the cells were fixed, permeabilized, and incubated with anti-gp95/sortilin peptide antibody. Staining was visualized with fluorescein isothiocyanate-conjugated anti-rabbit antibody and fluorescence microscopy. Nuclei were stained by DAPI. Dots indicate positions of unstained cells. B, Western blot of human gp95/sortilin transfected COS-1 cells (lane 1) and mock-transfected controls (lane 2). Cell lysates were resolved by reducing SDS-PAGE, electroblotted, and stained using anti-gp95/sortilin peptide antibody and peroxidase-conjugated anti-rabbit antibody. C, COS-1 cells were biolabeled overnight and lysed in 1% Triton X-114. Following phase separation, RAP binding proteins were precipitated using RAP-Sepharose. The precipitates were analyzed by SDS-PAGE and autoradiography. Results in transfected (lane 1) and mock-transfected (lanes 2) cells are shown. The position of the alpha 2-macroglobulin receptor/low density lipoprotein receptor (LRP/alpha 2MR) is indicated.
[View Larger Version of this Image (76K GIF file)]


To further determine the localization conveyed by the C-terminal g995/sortilin tail, a chimeric receptor comprising the lumenal and transmembrane domains of the IL-2 receptor (Tac/CD25) and the cytoplasmic tail of gp95/sortilin was expressed in COS-1 cells. Fig. 8 demonstrates a predominant localization of the chimeric receptor in the Golgi compartment (panel a) and demonstrates in addition that it exhibits a pronounced colocalization with the CI-M6P in Golgi as well in endosome-like vesicles (panels b-d).


Fig. 8. Confocal laser microscopy of COS-1 cells showing localization of CI-M6P receptors and transiently expressed chimeric receptors, containing the lumenal and the transmembrane domain of the IL-2 receptor and the cytoplasmic tail of gp95/sortilin. a, cluster of cells stained for chimeric receptors with anti-IL-2 receptor antibodies. b, single cell stained with anti-IL-2 receptor antibody. c, the same cell as in panel b stained with anti-CI-M6P receptor antibody. d, computerized imaging (black stain) showing colocalization of staining in b and c.
[View Larger Version of this Image (122K GIF file)]


Gp95/Sortilin mRNA Transcripts in Human Tissues

The occurrence of gp95/sortilin mRNA in vivo was estimated by Northern blotting using total mRNA isolated from various human tissues. As demonstrated in Fig. 9, the probe hybridized to two species of mRNA, e.g. one transcript of about 8 kb and another of about 3.5 kb. The 8-kb transcript appeared particularly abundant, but both species were expressed at high levels in brain, spinal cord, heart and skeletal muscle, thyroid, placenta, and testis. In contrast, samples from lymphoid organs and cells and from kidney, colon, and liver contained only minor amounts of specific mRNA.


Fig. 9. Northern blot analysis of gp95/sortilin transcripts in human tissues. Nylon membranes carrying poly(A)+ RNA (2 µg/lane) from 23 human tissues were incubated with a 35P-labeled 1581-base pair probe, partially encoding gp95/sortilin and subsequently washed under high stringency conditions. Hybridization was visualized by autoradiography. 1, heart; 2, brain; 3, placenta; 4, lung; 5, liver; 6, skeletal muscle; 7, kidney; 8, pancreas; 9, spleen; 10, thymus; 11, prostate; 12, testis; 13, ovary; 14, small intestine; 15, colon (mucosal lining); 16, peripheral blood leukocyte; 17, stomach; 18, thyroid; 19, spinal cord; 20, lymph node; 21, trachea; 22, adrenal gland; 23, bone marrow.
[View Larger Version of this Image (30K GIF file)]



DISCUSSION

The ~95-kDa protein, designated gp95/sortilin, was purified by RAP affinity chromatography and represents the first RAP binding receptor not related to the LDL receptor family. gp95/sortilin shows no homology to the LDL receptor family receptors but displays some striking structural similarities to the recognized sorting receptors, yeast Vps10p and the CI-M6P receptor, suggesting it to be a novel candidate sorting receptor. Moreover, the homology between gp95/sortilin and Vps10p defines a previously unrecognized cysteine-rich structural protein module.

gp95/sortilin purified from human brain represents a truncated form cleaved at the furin cleavage site after the RWRR sequence in the N-terminal part (aa 41-44) of the receptor. The furin-specific Arg-Xaa-Arg/Lys-Arg cleavage site is found in a number of growth factor precursors and in some growth factor proreceptors including those for the insulin receptor, insulin-like growth factor 1 receptor, and hepatocyte growth factor beta  (29). Also, several reports indicate that furin is responsible for the generation of bioactive proteins such as transforming growth factor beta  (30) and the insulin receptor (31). Future studies should determine if full-length gp95/sortilin represents an inactive precursor form and if the 5-kDa peptide thought to be generated by the furin-mediated cleavage has any biological function. However, the functional relevance of the furin cleavage site seems supported by the fact that SorLA, another RAP binding receptor, which we recently identified (asterisk in Fig. 1), also carries a Vps10p homologous domain truncated at an N-terminal furin site (14).

Homology to Yeast Vps10p Reveals a Novel Extracellular Protein Module

gp95/sortilin exhibits two interesting homologies. The most extensive concerns the lumenal region of sortilin, which aligns with each of the two lumenal domains in yeast Vps10p and with similar domains in the yeast VTH2 product (3) and in Sorla-1 (14)/LR11 (15). The homology between the domains is particularly strong within a ~140-aa stretch (Fig. 5A), and the conserved positions of 10 cyteine residues in these segments strongly indicate that they represent a hitherto unrecognized type of lumenal module.

Homology to Functional Cytoplasmic Motifs in the M6P Receptors

The second type of homology is found in the cytoplasmic tail of gp95/sortilin, which in the extreme C terminus contains a nine-residue stretch (HDDSDEDLL) also found in the extreme C terminus of the CI-M6P receptor and similar to the EESEERDDHLL segment in the CD-M6P receptor. This nanopeptide is potentially important in gp95/sortilin as it comprises two functional sites, e.g. a casein kinase II phosphorylation site and an adjacent dileucine motif, that are known to be major functional sites for sorting of the M6P receptors between the trans-Golgi network (TGN) and late endosomes (32, 33). In fact, it is conceivable that the potential sorting of gp95/sortilin is much similar to that of the CI-M6P receptor as suggested by the confocal microscopy of the intracellular localization of the transiently expressed full-length protein and of chimeric receptors containing the gp95/sortilin tail (Fig. 8).

M6P receptor-mediated sorting involves trafficking by clathrin-coated vesicles derived from both Golgi and the plasma membrane. In contrast to plasma membrane vesicles, which are associated with type 2 assembly proteins (AP-2), formation of Golgi vesicles depends on the recruitment of AP-1 assembly proteins. In the CD-M6P receptor, the casein kinase site is essential to the formation of clathrin-coated vesicles, the initial step in TGN sorting, while the dileucine as such relates to downstream sorting events (32, 34). The di-leucine motif, in some cases adjacent to a potential kinase site, is a well known sorting motif in many membrane receptors, but in contrast to similar phosphorylation sites in the cytoplasmic tails of other receptors (35-37), the casein kinase II site in the CD-M6P receptor, and very likely in the CI-M6P receptor, mediates membrane binding of type-1 assembly proteins (AP-1) and conditions TGN vesicle formation (34). It therefore seems that the AP-1 binding of the M6P receptors relates not just to the kinase site per se but also to its unique position at the C terminus in M6P receptors. Thus, the structural identity/homology between gp95/sortilin and the M6P receptors and in particular the identical C-terminal localization of the stretch in all three receptors may reflect that gp95/sortilin is engaged in protein sorting from TGN and possibly even engaged in AP assembly.

Other Sorting Motifs in the Cytoplasmic Tail

In addition to the dileucine motif and the potential casein kinase phosphorylation site, gp95/sortilin contains at least two candidate sorting signals. (i) The sequence YSVL, e.g. aa 14-17 in the cytoplasmic tail, abides by the consensus motif YXXZ, where Z stands for either a bulky hydrophobic or an aromatic residue (35, 38). This type of motif mediates rapid internalization and sorting of several membrane proteins including the CD-M6P receptor and the CD3gamma (38, 39), and in the CI-M6P receptor, the sequence YSKV (aa 26-29 of the cytoplasmic domain) is known to be decisive for rapid internalization (40-42). Interestingly, a similar aromatic residue-based sorting signal, the tetrapeptide YSSL, conveys sorting in yeast Vps10p (3). (ii) The sequence FLVHRY (aa 9-14 in the cytoplasmic tail of gp95/sortilin), immediately adjacent to the transmembrane domain and partly overlapping the tetrapeptide YSVL, constitutes another candidate internalization/sorting signal in accordance with the overall motif (F/Y)XXXX(F/Y), which has been deduced from active sites in both M6P receptor and LDL receptor family receptors (1, 35, 38, 39).

In summary, we have cloned, sequenced, and expressed a previously unknown receptor purified from human brain by RAP affinity chromatography. Strong homologies to known eukaryotic lysosomal sorting receptors and the putative sorting signals in its C-terminal tail strongly suggest gp95/sortilin as a receptor involved in endosomal protein sorting. In agreement with this suggestion, expression experiments demonstrated that the tail conveys colocalization with the CI-M6P receptor in both Golgi and endosome-like vesicles. Finally, the identification of this protein is of structural interest, since the established homology between gp95/sortilin, sorLA/LR11, and two separate domains in the yeast Vps10p defines a novel mammalian acidic protein module with conserved positioning of 10 cysteines. Further studies are in progress to investigate the ligand binding properties of this domain.


FOOTNOTES

*   This work was supported by The Danish Medical Research Council, The Novo Nordisk Foundation, Danish Biotechnology Research and Development program 1995-1998, The Carlsberg Foundation, the Aage Bangs Foundation, and the Leo and Karen M. Nielsens Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. The 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 GenBankTM/EMBL Data Bank with accession number(s) X98248[GenBank].


§   To whom correspondence should be addressed: Dept. of Medical Biochemistry, University of Aarhus, Ole Worms Alle, Bldg. 170, DK-8000, Denmark. Tel.: 45-89-422880; Fax: 45-86-131160.
1    The abbreviations used are: M6P, mannose-6-phosphate; CD, cation-dependent; CHAPS, 3-((cholamidopropyl)D-imethylammoniol)1-propanesulfonate; CI, cation-independent; LDL, low density lipoprotein; TGN, trans-Golgi network; Vps10p, vacuolar protein sorting 10 protein; PAGE, polyacrylamide gel electrophoresis; kb, kilobase(s); aa, amino acid(s); IL-2, interleukin 2.

Acknowledgments

The expert technical help by Nina Jørgensen, Helle Nielsen, Kate Nielsen, and Winnie Pedersen is greatly acknowledged.


REFERENCES

  1. Kornfeld, S. (1992) Annu. Rev. Biochem. 61, 307-330 [CrossRef][Medline] [Order article via Infotrieve]
  2. Marcusson, E. G., Horazdovsky, B. F., Cereghino, J. L., Gharakhanian, E., and Emr, S. D. (1994) Cell 77, 579-586 [Medline] [Order article via Infotrieve]
  3. Cooper, A. A., and Stevens, T. H. (1996) J. Cell Biol. 133, 529-541 [Abstract]
  4. Reitman, M. L., Varki, A. P., and Kornfeld, S. (1981) J. Clin. Invest. 67, 1574-1579 [Medline] [Order article via Infotrieve]
  5. Hasilik, A., Waheed, A., and von Figura, K. (1981) Biochem. Biophys. Res. Commun. 98, 761-767 [Medline] [Order article via Infotrieve]
  6. Waheed, A., Polhlmann, A., Hasilik, K., von Figura, A., van Elsen, A., and Leroy, J. G. (1982) Biochem. Biophys. Res. Commun. 105, 1052-1058 [Medline] [Order article via Infotrieve]
  7. Owada, M., and Neufeld, E. F. (1982) Biochem. Biophys. Res. Commun. 105, 814-820 [Medline] [Order article via Infotrieve]
  8. Tsuji, L., Omura, K., and Suzuki, Y. (1988) Clin. Chim. Acta 176, 115-121 [Medline] [Order article via Infotrieve]
  9. Glickman, J. N., and Kornfeld, S. (1993) J. Cell Biol. 123, 99-108 [Abstract]
  10. Bu, G., Geuze, H., Strous, G. J., and Schwartz, A. L. (1995) EMBO J. 14, 2269-2280 [Abstract]
  11. Willnow, T. E., Armstrong, S. A., Hammer, R. E., and Herz, J. (1995) Proc. Natl. Acad. Sci. U. S. A. 92, 4537-4541 [Abstract]
  12. Willnow, T. E., Rohlmann, A., Horton, J., Otani, H., Braun, J. R., Hammer, R. E., and Herz, J. (1996) EMBO J. 15, 2632-2639 [Abstract]
  13. Bu, G., and Rennke, S. (1996) J. Biol. Chem. 271, 22218-22224 [Abstract/Free Full Text]
  14. Jacobsen, L., Madsen, P., Moestrup, Lund, A. H., Tommerup, N., Nykjær, A., Sottrup-Jensen, L., Gliemann, J., and Petersen, C. M. (1996) J. Biol. Chem. 271, 31379-31383 [Abstract/Free Full Text]
  15. Yamazaki, H., Bujo, H., Kusunoki, J., Seimiya, K., Kanaki, T., Morisaki, N., Schneider, W. J., and Saito, Y. (1996) J. Biol. Chem. 271, 24761-24768 [Abstract/Free Full Text]
  16. Petersen, C. M., Ellgaard, L., Nykjær, A., Vilhardt, F., Vorum, H., Thøgersen, H. C., Nielsen, M. N., Jacobsen, L., Moestrup, S. K., and Gliemann, J. (1996) EMBO J. 15, 4165-4173 [Abstract]
  17. Gliemann, J., Nykjær, A., Petersen, C. M., Jørgensen, K. E., Nielsen, M., Andreasen, P. A., Christensen, E. I., Lookene, A., Olivecrona, G., and Moestrup, S. K. (1994) Ann. N. Y. Acad. Sci. 737, 20-38 [Medline] [Order article via Infotrieve]
  18. Battey, F. D., Gåfvels, M. E., FitzGerald, D. J., Argraves, W. S., Chappell, D. A., Strauss, J. F., III, and Strickland, D. K. (1994) J. Biol. Chem. 269, 23268-23273 [Abstract/Free Full Text]
  19. Jensen, P. H., Gliemann, J., and Ørntoft, T. (1992) FEBS Lett. 305, 129-132 [CrossRef][Medline] [Order article via Infotrieve]
  20. Nykjær, A., Petersen, C. M., Møller, B., Jensen, P. H., Moestrup, S. K., Holtet, T. L., Etzerodt, M., Thøgersen, H. C., Munch, M., Andreasen, P. A., and Gliemann, J. (1992) J. Biol. Chem. 267, 14543-14546 [Abstract/Free Full Text]
  21. Sottrup-Jensen, L. (1995) Anal. Biochem. 225, 185-188 [CrossRef][Medline] [Order article via Infotrieve]
  22. Jacobs, K. A., Rudersdorf, R., Neill, S. D., Dougherty, J. P., Brown, E. L., and Fritsch, E. F. (1988) Nucleic Acids Res. 16, 4637-4650 [Abstract]
  23. Honoré, B., Madsen, P., and Leffers, H. (1993) J. Biochem. Methods 27, 39-48
  24. Kaufman, R. J., Davies, M. V., Wasky, L. C., and Michnick, D. (1991) Nucleic Acids Res. 19, 4485-4490 [Abstract]
  25. LaFlamme, S. E., Thomas, L. A., Yamada, S. S., and Yamada, K. M. (1994) J. Cell Biol. 126, 1287-1298 [Abstract]
  26. Düzgünes, N., and Felgner, P. L. (1993) Methods Enzymol. 221, 303-306 [Medline] [Order article via Infotrieve]
  27. Tommerup, N., and Vissing, H. (1995) Genomics 27, 259-264 [CrossRef][Medline] [Order article via Infotrieve]
  28. Von Heijne, G. (1983) Eur. J. Biochem. 133, 17-21 [Abstract]
  29. Hosaka, M., Nagahama, M., Kim, W.-S., Watanabe, T., Hatsuzawa, K., Ikemizu, J., Murakami, K., and Nakayama, K. (1991) J. Biol. Chem. 266, 12127-12130 [Abstract/Free Full Text]
  30. Dubois, C. M., Laprise, M.-H., Blanchette, F., Gentry, L. E., and Leduc, R. (1995) J. Biol. Chem. 270, 10618-10624 [Abstract/Free Full Text]
  31. Bravo, D. A., Gleason, J. B., Sanchez, R. I., Roth, R. A., and Fuller, R. S. (1994) J. Biol. Chem. 269, 25830-25837 [Abstract/Free Full Text]
  32. Johnson, K. F., and Kornfeld, S. (1992) J. Biol. Chem. 267, 17110-17115 [Abstract/Free Full Text]
  33. Johnson, K. F., and Kornfeld, S. (1992) J. Cell Biol. 119, 249-257 [Abstract]
  34. Mauxion, F., Le Borgne, R., Munier-Lehmann, H., and Hoflack, B. (1996) J. Biol. Chem. 271, 2171-2178 [Abstract/Free Full Text]
  35. Sandoval, I. V., and Bakke, O. (1994) Trends Cell Biol. 4, 292-297 [CrossRef]
  36. Haft, C. R., Klausner, R. D., and Taylor, S. I. (1994) J. Biol. Chem. 269, 26286-26294 [Abstract/Free Full Text]
  37. Hunziker, W., and Fumey, C. (1994) EMBO J. 13, 2963-2969 [Abstract]
  38. Trowbridge, I. S., Collawn, J. F., and Hopkins, C. R. (1993) Annu. Rev. Cell Biol. 9, 129-161 [CrossRef]
  39. Johnson, K. F., Chan, W., and Kornfeld, S. (1990) Proc. Natl. Acad. Sci. U. S. A. 87, 10010-10014 [Abstract]
  40. Lobel, P., Fujimoto, K., Ye, R. D., Griffith, G., and Kornfeld, S. (1989) Cell 57, 787-796 [Medline] [Order article via Infotrieve]
  41. Canfield, W. M., Johnson, K. F., Ye, R. D., Gregory, W., and Kornfeld, S. (1991) J. Biol. Chem. 266, 5682-5688 [Abstract/Free Full Text]
  42. Jadot, M., Canfield, W. M., Gregory, W., and Kornfeld, S. (1992) J. Biol. Chem. 267, 11069-11077 [Abstract/Free Full Text]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.