The Lysosomal Targeting of Acid Sphingomyelinase

Xiaoyan Ni Department of Anatomy and Cell Biology McGiII University Montreal, Canada

A thesis submitted to McGiII University in partial fulfillment of the requirements of the degree of Master in Science

395 Wellington Street 395, rue Wellington Ottawa ON K1A ON4 Ottawa ON K1A ON4 Canada Canada

Your file Votre référence ISBN: 978-0-494-22756-5 Our file Notre référence ISBN: 978-0-494-22756-5

NOTICE: AVIS: The author has granted a non L'auteur a accordé une licence non exclusive exclusive license allowing Library permettant à la Bibliothèque et Archives and Archives Canada to reproduce, Canada de reproduire, publier, archiver, publish, archive, preserve, conserve, sauvegarder, conserver, transmettre au public communicate to the public by par télécommunication ou par l'Internet, prêter, telecommunication or on the Internet, distribuer et vendre des thèses partout dans loan, distribute and sell theses le monde, à des fins commerciales ou autres, worldwide, for commercial or non sur support microforme, papier, électronique commercial purposes, in microform, et/ou autres formats. paper, electronic and/or any other formats.

The author retains copyright L'auteur conserve la propriété du droit d'auteur ownership and moral rights in et des droits moraux qui protège cette thèse. this thesis. Neither the thesis Ni la thèse ni des extraits substantiels de nor substantial extracts from it celle-ci ne doivent être imprimés ou autrement may be printed or otherwise reproduits sans son autorisation. reproduced without the author's permission.

ln compliance with the Canadian Conformément à la loi canadienne Privacy Act some supporting sur la protection de la vie privée, forms may have been removed quelques formulaires secondaires from this thesis. ont été enlevés de cette thèse.

While these forms may be included Bien que ces formulaires in the document page cou nt, aient inclus dans la pagination, their removal does not represent il n'y aura aucun contenu manquant. any loss of content from the thesis. ••• Canada ABSTRACT

Acid sphingomyelinase (ASM), a member of the saposin like protein (SAPLIP) family, is a lysosomal hydrolase that converts sphingomyelin to ceramide. The deficient activity of ASM causes a variant form (i.e., type AIB) of the inherited disorder Niemann-Pick disease. The lysosomal targeting mechanism of ASM has not been conclusively identified.

Previous studies suggested that ASM could use another membrane-associated receptor as weIl as M6P receptor to target lysosomes. Sortilin, a type 1 transmembrane glycoprotein, belongs to a novel family of receptor proteins. Both the luminal domain and the cytoplasmic domain of sortilin show structural features typical of receptors involved in lysosomal or vacuolar targeting. Using a dominant-negative sortilin construct lacking the cytoplasmic taïl,

1 proved that sortilin was involved in the lysosomal targeting of ASM. Confocal microscopy revealed that truncated sortilin partially inhibited the lysosomal targeting of ASM in COS 7 cells and completely abolished the lysosomal targeting of ASM in I-cells. Pulse-chase experiments also suggested that sortilin was involved in normal sorting of newly synthesized

ASM. Over-expression of truncated sortilin accelerated and enhanced the secretion of ASM from COS 7 cells and I-cells. Co-immunoprecipitation assays further confmned the interaction between sortilin and ASM. 1 also observed that the lysosomal transport of ASM was reduced to sorne extent in I-cell disease fibroblasts as compared to normal fibroblasts. In conclusion, both the M6P receptor and sortilin mediate lysosomal targeting of ASM.

1 RÉSUMÉ

L'acide sphingomyelinase (ASM), un membre de la famille des protéines saposine

(SAPLIP), est un hydrolase lysosomal qui convertit le sphingomyeline en ceramide.

L'activité déficiente de l'ASM cause une forme variante (c'est-à-dire le type AIB) de la maladie héréditaire Niemann-Pick. Le transport de l'ASM aux lysosomes n'a pas été déterminé jusqu'à présent. Plusieurs études précédentes ont suggéré que l'ASM pourrait être transporté aux lysosomes par deux mécanismes dont un qui implique le récepteur M6P. Le récepteur sortilin, un glycoprotéine de type I appartient à une nouvelle famille de récepteurs.

Le domaine luminal ainsi que le domaine cytoplasmique du récepteur sortilin montre des caractéristiques typiques des récepteurs impliqués dans le transport aux lysosomes ou aux vacuoles. Utilisant un agencement du récepteur sortilin qui ne possédait pas de portion cytoplasmique, nous avons démontré que le sortilin est impliquée dans le transport de l' ASM aux lysosomes. La microscopie confocale a révélé que le sortilin sans portion cytoplasmique a diminué le transport de l' ASM dans des cellules COS-7 et a totalement interrompu ce même transport dans les cellules ICD. Les résultats des expériences pulse-chase suggèrent aussi que le sortilin est impliqué dans le transport normal de l' ASM. La surexpression du sortilin sans portion cytoplasmique a accéléré la sécretion de l' ASM dans des cellules COS-7 et ICD. Des tests de coimmunoprécipitation ont comfirmé l'existence d'interaction entre le sortilin et l'ASM. Nous avons aussi observé que le transport de l'ASM aux lysosomes était réduit dans les fibroblasts de patients atteints de l'ICD comparées à des fibroblasts normaux.

Finalement, le récepteur M6P ainsi que le sortilin sont impliqués dans le transport de l' ASM aux lysosomes.

ii To my parents and husband who have been a source of inspiration and love.

111 ACKNOWLEDGEMENTS

1 would first like to appreciate my supervisor Dr. Carlos Morales for his expert guidance, continuaI encouragement and support throughout my master studies. And 1 would like to extend gratitude to Dr. Yves Clermont for his constant caring and encouragement during my studies.

Jibin Zeng and Jacob Hassan are to be infinitely thanked for their helpful technical assistance on my molecular biology experiments and immunofluorescence staining. Thanks to Archana Srivastava for help trouble shoot with the confocal microscope.

1 thank my committee meeting advisors: Dr. Louis Hermo, Dr. John Presley and Dr.

Michael Greenwood for their professional advice on my research project. Especially, 1 would like to appreciate Dr. Louis Hermo for his constructive criticism on my seminars.

1 also thank Dr. Volkan Seyrantepe for providing me normal fibroblast cell line and thank Dr. C.M .. Peterson for giving me anti-sortilin antibody.

Special thanks to Maryssa Canuel and Myriam Block for their excellent French translation of the abstracto

Finally, 1 would like to express my most sincere thanks to my husband for his always support and to people in Morales's lab: Jibin Zeng, Maryssa Canuel, Libin Yuan and Ji-Hae

Kim, for their enthusiasm, emotional support and friendship.

IV TABLE OF CONTENTS

Abstract 1

Resume II

Dedication III

Acknowledgements iv

Table of contents v

List of figures Vlll

Introduction 1

Literature review 4

1 Saposin-like protein (SAPLIP) family 5

1 Members of SAPLIP family 5

2 Structural features of SAPLIPs 6

3 Lipid-binding ability of SAPLIPs 8

4 Role of saposin-like domain in the intracellu1ar targeting 9

ofSAPLIPs

5 Acid sphingomyelinase belongs to the SAPLIP family Il

II Targeting of Iysosomai proteins 13

1 Mannose 6-phosphate dependent transport of lysosomal proteins. 13

1.1 Biosynthesis and phosphorylation 13

1.2 Phosphotransferase and I-cell disease 14

1.3 The mannose 6-phosphate (M6P) receptors 16

v 1.5 Final destination of the mannose 6- phosphate receptors 19

2 Mannose 6-phosphate independent transport of lysosomal 19

proteins

2.1 Mannose 6-phosphate independent vacuolar targeting in yeast 19

2.2 Mannose 6-phosphate independent lysosomal targeting in 20

mammalian ceUs

III The VpslOp superfamily 22

1 A novel receptor family 22

2 The genes ofhuman Vpsl0p family 22

3 Two common structural features ofVpsl0p receptors 23

4 Vps 1Op family receptors bind unrelated ligands 24

5 Structure and functions of cytoplasmic tails of V ps 1Op receptors 25

6 Sortilin: a multifunction protein 26

IV GGAs are required for the sorting soluble lysosomal proteins 28

1 A novel family proteins interacting with ADP-ribosylation 28

proteins (ARFs)

2 Structure of GGA proteins 29

3 Functions of GGAs in the yeast 29

4 GGAs interacting proteins 30

5 SortiliniGGA-mediated pathways and the sorting of ASM 31

Materials and methods 32

Reagents and antibodies 32

VI DNA constructs 32

Celllines and cell cultures 33

Transfections of COS-7 cells 33

Electroporation 34

Immunofluorescence and confocal microscopy 34

Co-immunoprecipitation assays 35

Metabolic labeling and pulse-chase experiments 35

Results 37

Lysosomal targeting of ASM 37

Sortilin interacts with GGA through its cytoplasmic domain 38

Dominant-negative sortilin altered the lysosomal targeting 39

ofASM

ASM targeting is abolished when the M6P and sortilin pathways 40

are blocked

Dominant-negative sortilin enhances the secretion of ASM 41

Sortilin binds and interacts with ASM 42

Discussion 44

Conclusion 52

Abbreviation 53

References 57

Appendix - Research Compliance Certificate 66

vu LIST OF FIGURES

Figure 1 - Intracellular localization of ASM in I-Cells.

Figure 2 - Lysosomal targeting of ASM in nonnal fibroblasts and in I-cells.

Figure 3 - Expression of sortilin-EGFP.

Figure 4 - Intracellular localization of full-length sortilin and truncated sorti lin.

Figure 5 - Effects of truncated GGA on intracellular trafficking of sortilin.

Figure 6 - Co-immunoprecipitation assay showing association of sortilin and GGA.

Figure 7 - Effects offull-length and truncated sortilins on lysosomal targeting of ASM.

Figure 8 - Effect of truncated GGA on lysosomal trafficking of ASM in COS 7 Cells.

Figure 9 - Effect oftruncated sortilin on lysosomal targeting of ASM in I-cells.

FigurelO- Influence oftruncated sortilin on the sorting of ASM in COS7 cells and I-cells.

Figurell- Co-immunoprecipitation showing interaction between sortilin and ASM.

Vlll INTRODUCTION

In mammalian cells the targeting of newly synthesized lysosomal proteins to the lysosomes is dependent on their recognition by two specific mannose 6-phosphate receptors: the 46 KDa cation-dependent mannose 6-phosphate receptor (CD-M6P-Rc) and the 300 KDa cation-independent mannose 6-phosphate receptor (CI-M6P-Rc).(1) Studies with cells deficient in CD-M6P-Rc or in CI-M6P-Rc indicate that both receptors participate in lysosomal enzyme sorting, and that the CI-M6P-Rc becomes the dominant receptor as it is able to compensate for the loss of the CD-M6P-Rc.(2)

Although the M6P-Rcs play an important role in the intracellular transport of lysosomal enzymes, some considerations suggest that there are alternative mechanisms for lysosomal enzyme targeting. For example, lysosomes from patients with mucolipidosis II (3) contain a number of soluble lysosomal proteins such as prosaposin and GM2 activator protein

(GM2AP).(4) (5)

Recently, the lysosomal trafficking of prosaposin was proved to be mediated by sortilin, a 95 KDa glycoprotein purified from human brain extract.(6) Sortilin is a type 1 transmembrane protein, which contains a VpslOp domain in its luminal region and belongs to the VpslOP superfamily.(6) The VpslOP superfamily is a novel sorting receptor family, named after a yeast vacuolar sorting protein (VpslOp) that transports carboxypeptidase Y

(CPY) from the trans-Golgi network (TGN) to the vacuole.(7) Interestingly, the sequence of the cytoplasmic tail of sortilin Îs similar to that ofthe CI-M6P-Rc.(6) Sortilin was also shown to bind and to co-Iocalize with the monomeric adaptor protein GGA, a novel ubiquitous coat protein mediating the formation of intracellular transport-intermediates and selection of

1 cargo.(8) AIl these findings suggest that sortilin is potentially an alternative receptor capable of sorting soluble lysosomal enzymes from Golgi apparatus to the lysosomes.

Prosaposin, a precursor of four lysosomal saposins (A-D), belongs to the saposin-like protein (SAPLIP) family. SAPLIPs also include acyloxyacyl hydrolase (AOAH), surfactant protein B (SPB) and acid sphingomyelinase (ASM). The SAPLIP family share a common saposin-like domain made up of 5 alpha helices enclosing a hydrophobie cavity.(9) The saposin-like domain is believed to play a role in the intracellular trafficking of SAPLIPs from trans-Golgi network to the lysosomes.(lO,ll) It has been demonstrated that the interaction between prosaposin and sortilin is mediated by the C-terminus of prosaposin. The C-terminus of prosaposin contains a saposin-like domain that has been implicated in the transport of prosaposin to the lysosomes.(12) Similarly, studies on AOAH lacking the N-terminal saposin-like domain showed that truncated AOAH does not reach the lysosomes and is not fully processed into mature protein.(10)

Acid sphingomyelinase (ASM), a soluble lysosomal enzyme that hydrolyses sphingomyelin into ceramide and phosphocholine, is synthesized as a precursor protein that consists of a signal peptide followed by a saposin-like domain and a phosphodiesterase domain.(13) ASM activity is reduced in the lysosomes of cells from ICD patients, which suggests that ASM may be transported to lysosomes via the M6P-Rc. Although ASM is endocytosed from the extracellular space, its internalization is not inhibited by free mannose

6-phosphate, indicating that this hydrolase could also be sorted by an alternative receptor.(14)

2 Based on the literature, we postulated that sortilin interacts with soluble lysosomal proteins that contain saposin-like motifs and mediates the transport of these proteins from

Golgi apparatus to the lysosomes. Given the fact that ASM belongs to the SAPLIP family, we have tested the hypothesis that ASM is alternatively sorted to the lysosomes by sortilin.

To this effect we investigated the subcellular localization and the sorting mechanism of ASM in COS-7 cells and of cells from I-cell disease (ICD) patients using a dominant-negative sortilin construct, immunofluorescence, and pulse-chase analysis. Our results showed that both M6P-Rc and sortilin are involved in the sorting and transport ASM from Golgi apparatus to the lysosomes.

3 LITERATURE REVIEW

The objective of this review is to discuss the most relevant publications in the field of sorting and trafficking of lysosomal resident pro teins. Although sorting and trafficking are two different processes, they are temporally and spatially related. Both are composed of several steps and both involve the participation of luminal, transmembrane and cytoplasmic proteins. During the first step of sorting a soluble lysosomal protein or "cargo" must be recognized by a specific sorting receptor. This process takes place in the lumen of the Golgi apparatus and sometimes on the external surface of the plasma membrane(15) Most commonly a lysosomal protein must be tagged first with a molecule such as mannose 6- phosphate, to allow its recognition by a sorting receptor which binds to this tag. However, protein sorting in most eukaryotic cells may also involve protein-protein interactions between the cargo and the receptor. Consequently, eukaryotic cells may have an additional repertoire of receptors that recognize amino acid sequences and/or motifs within the lysosomal cargo.

Therefore, such motifs have the property to specify the sorting and final destination of the cargo. This possibility will be addressed in the present review.

For sorting and trafficking, a receptor must also bind a repertoire of cytoplasmic coat proteins (adaptor proteins, clathrin, etc) that cause vesicles to bud from donor membranes

(Le., the trans-Golgi network or TGN) and to traffic to acceptor membranes (late endosomes and lysosomes). In addition to the Golgi apparatus, other major sites ofvesicle formation and budding inc1ude the endoplasmic reticulum and the plasma membrane. The Golgi apparatus is known to be the major site of sorting for newly synthesized proteins destined to the lysosomal compartment. Two different types of proteins are sorted to the lysosomes:

4 transmembrane and soluble lysosomal proteins. Due to their specific characteristics these two types of proteins use different sorting mechanisms. This will be also discussed in the present review.

Finally, this review will present a succinct description of cargo molecules with

special attention to putative motifs within their primary structure, which may specify sorting and trafficking.

1. Saposin-Iike protein (SAPLIP) family

1. Members of SAPLIP family

The saposin-like protein family (SAPLIP) is a large and diverse group of proteins found in a variety of eukaryotic cells from plants and animais. The SAPLIP members share cysteine-rich saposin-like sequences. These conserved cysteines, form three intra-domain disulfide bonds that create a common structural framework upon which other conserved amino acids form five amphipathic a-helices implicated in diverse functions. (9,16)

Members of the SAPLIP family include saposins A, B, C, D, NK-Iysin, surfactant protein B (SPB), acid sphingomyelinase (ASM), acyloxyacyl hydrolase (AOAH) and plant aspartic protease. Most SAPLIPs appear to bind to or interact with one or more membrane lipids, yet their properties and presumed functions are quite different. (17-19)

Saposins A-D are derived from a precursor protein, prosaposin, by the proteolytic cleavage of the latter in the lysosomes. Therefore, the saposins are found mainly in this compartment, where they facilitate the catabolism of glycosphingolipids with short oligosaccharide groups. The deficiency of prosaposin and/or at least two saposins has been associated with lipid storage disorders in humans. (20,21)

5 AOAH is a lipase found in phagocytic cells, which c1eaves fatty acyl chains from bacteriallipopolysaccharides (LPS). Thus, this protein plays a major role in the elimination of microorganisms phagocytosed by monocytes and macrophages. (22)

SPB is a 9 KDa hydrophobic protein, produced in alveolar type II cells, that enhances the rate of spread of surfactant along the water-air interface in the pulmonary alveolus.

Deficiency of SPB has been found in infants with congenital alveolar proteinosis.(23)

ASM is a soluble lysosomal hydrolase that cleave the phosphodiester bond of sphingomyelin to ceramide and phosphocholine. It is encoded by sphingomyelin phosphodiesterase-l (SMPDl) gene. Mutations of this gene cause inherited lysosomal storage disorders Niemann-Pick disease type A and B.(24,25)

NK-Iysin is a basic polypeptide consisting of 78 amino acids with an antibacterial activity and the capacity of lysing tumor cells. It was originally found in the porcine small intestine and synthesized by lymphocytes. (26)

Phytepsin, a plant aspartic protease, resides in barley grain, roots, stems, leaves and flowers. It is considered to be a plant homologue of mammalian lysosomal cathepsin D and yeast vacuolar protease A. Although the exact function of phytepsin is unclear, it may participate in metabolic turnover and in protein processing events in barley tissues.(27,28)

2. Structural features of SAPLIPs

SAPLIPs contain saposin-like motifs, which are composed, on average, of 80 amino acids in length and have a characteristic pattern of six conserved cysteines that form three disulfide bonds. The arrangement of these cysteines was first determined by Schuette et al,

(29) who found that the three disulphide bonds are formed between the first and last

6 cysteines, between the second and the second last cysteines and between the third and the fourth cysteines. Disulfide cross-linking renders the backbone structure of SAPLIPs remarkably stable to heat, low pH and proteolytic degradation.

NK-Iysin was the first protein of this family whose three dimensional (3-D) structure was determined by nuclear magnetic resonance (NMR). (30) Similarly, the plant-specific domain ofphytepsin was the first member of the SAPLIP family for which the 3-D structure has been determined by x-ray crystallography. (9) Comparison of the pro-phytepsin and NK lysin showed high similarity in their 3-D structures. Both ofthem contain five a-helices that form a helical cage. The internal surface of the cage is lined with side chains from hydrophobic residues while the outer surface is mainly hydrophilic. The five a-helices are oriented in an up-down-up-down manner. The helices interact with each other through the hydrophobic residues, which are stabilized by the inter-helical disulfide bridges. Study of the structures of saposins A, C and D showed that they all contain four or five a-helices forming similar "helical cages" as in the NK-Iysin and plant aspartic protease. The crystal structure of human saposin B revealed an unusual shell-like dimmer consisting of a monolayer of a helices enclosing a large hydrophobic cavity. (31) This characteristic a-helix cage or

"bundle" was referred to as "saposin fold" or "saposin-like domain", and suggested as a representative structure of all SAPLIPs. In the sequence alignment, the N- and C- terminal parts of the saposin-like domain ofphytepsin are swapped comparing with those ofNK-lysin, while the swapping does not impact the orientation of the helices. (9)

Although the GM2AP is not considered to be a saposin-like protein, it contains 8 cysteines that form four disulfide bonds. The secondary structure of the GM2AP differs from

7 that of the SAPLIPs. CD (Circular Dichroism) spectroscopy predicts that the GM2AP has a unique hydrophobic ~-cup structure that results in the formation of a spacious cavity, largely, due to the presence of extensive ~-sheet structures.

Saposin B is considered a special SAPLIP since its structure is similar to that of the known monomeric members of the saposin-like family. However the Cl-helices of saposin B differ from the rest of the SAPLIP members in the sense that they are repacked into a different tertiary conformation that form a homodimer.

3. Lipid-binding ability of SAPLIPs

AlI SAPLlPs share both lipid-binding and membrane-perturbing properties, which is a characteristic feature of this protein family. In fact, the surfactant protein B (SPB) interacts preferentially with anionic lipids forming monolayer capable of lowering the normal surface tension at the alveolar interface.(32) NK-lysin, a tumorolytic and antibacterial peptide ofNK cells, interacts with and destabilizes lipid bi-layers. Saposin C stimulates glucosylceramidase by interacting with phophatidyl-serine-containing membranes. (33)

Inhibition of sphingolipid biosynthesis, with organic and synthetic compounds such as fumonisin BI (FBI) and tricyclodecan-9-yl xanthate potassium salt (D609); can interfere with the lysosomal targeting of prosaposin. Since both FB 1 and D609 inhibit the biosynthesis of sphingomyelin it has been postulated that this sphingolipid interact with prosaposin. (34)

Indirectly, these results indicate that lipid binding is important for intracellular trafficking of

SAPLIPs.

Interestingly, the removal of sugar moiety did not influence the association of saposin

D with phospholipid membranes, suggesting that deglycosylation does not affect the lipid-

8 binding property of SAPLIPs.(35) Similarly, tunicamycin treatment did not interfere with the lipid interaction of non-glycosylated prosaposin and increased its transport to the lysosomes, suggesting that an intrinsic sequence may specify the lipid binding activity. On the other hand, the maintenance of disulfide bonds appears to be essential for the interaction of

SAPLIPs with membrane lipids in lysosome and vacuole targeting. In fact, mutations that disrupt disulfide bridges in saposin B and C are the causes of metachromatic leukodystrophy and Gaucher-like disease, respectively.(36,37) In conclusion, disulfide bonds are critical for the three dimensional structure of saposin-like domains of SAPLIP and for the formation of

"helical cages". Therefore, it is not difficult to presume that the hydrophobic internal surface of the helical cages provides an interface for binding between membrane lipid and SAPLIPs.

Thus, the presence of saposin-like domain is the structural base of the lipid-binding property ofall members of the SAPLIP family. (38,39)

4. Role of saposin-like domain in the intracellular targeting of SAPLIPs

The association of SAPLIPs with membrane lipids is weIl demonstrated, in addition,

SAPLIPs also seem to interact with membrane-associated receptor proteins during Golgi mediated intracellular sorting of SAPLIPs. There are some direct and indirect evidences suggesting that saposin-like domain may play an important role in the intracellular targeting ofSAPLIPs.

The AOAH is a soluble lysosomal enzyme that is synthesized as a single-chain preCursor in the endoplasmic reticulum (ER), which in tum is proteolytically processed into mature protein in the lysosomes. The mature protein consists of a small and a large subunit.

Since the small subunit bears a saposin-like domain, AOAH is also considered a member of

9 the SAPLIP family.(40) Immunofluorescence staining and pulse-chase experiments showed that both a recombinant AOAH large subunit and an AOAH variant that lacked a 33-amino

acid region containing the saposin-like domain within the small subunit could not be found in the lysosomes. Both proteins were secreted to the extracellular space and in the latter case the mutant AOAH was poorly processed. On the other hand, cells transfected with wild-type recombinant AOAH, the enzyme reached the lysosomes. (1 0) Data from another lab showed that an AOAH variant that lacked both Cys2 and Cys3 retained 40% of its enzymatic activity but did not concentrate in lysosomes. These results indicate that the absence or disruption of the saposin-like domain apparently prevents normallysosomal targeting of AOAH. (10)

Studies carried out in plant cells showed that the saposin-like domain of SAPLIPs plays a role in vacuolar transport. Barley aspartic protease (phytepsin) was shown previously to reach the vacuole via trafficking through the Golgi apparatus. However, deletion of the plant-specifie insert (PSI) that contains a saposin-like domain led to efficient secretion of truncated phytepsin instead of vacuolar targeting. (41)

In a recent study, the selective deletion of prosaposin functional domains proved that both the D domain (corresponding to saposin D) and the C-terminus were required for lysosomal targeting of prosaposin.(34) Furthermore, when a chimeric protein composed of albumin followed by the D domain and C-terminus of prosaposin, the fusion protein was routed to the lysosomes.(12) In fact, the C-terminus of prosaposin contains a saposin-like domain that is significantly similar to the N-terminus of surfactant protein B (SP-B). SP-B also requires the presence of the N-terminus for its transient routing to multivesicular and lamellar bodies.(11)

10 Structural and functional studies of pro-phytepsin revealed that a putative membrane receptor-binding region might be located on the outer surface of the interdomain of pro phytepsin. This region is composed of amino acid residues found in the plant-specific insert and also in the mature phytepsin. Therefore~ it is likely that the function of the saposin-like domains is to bring prophytepsin into contact with membrane microdomains, such as lipid rafts, containing the sorting receptor. Thus, the resulting complex could be then packed into sorting vesicles destined to the vacuoles.

5. Acid sphingomyelinase belongs to the SAPLIP family

Acid sphingomyelinase (ASM) is a soluble lysosomal enzyme found in all mammalian cells, possessing optimum pH between 4.5-5.5. ASM hydrolyzes sphingomyelin into ceramide and phosphocholine. Individuals who have mutations in the ASM gene develop type A and B Niemann-Pick disease (NPD).( 42,43) NPD is characterized by accumulation of undigested sphingomyelin in the lysosomes. Similarly to the type A NPD patients, the ASM knock out mice accumulate sphingomyelin in the lysosomes of cells belonging to the reticuloendothelial system, predominantly in the liver, spleen, lung, bone marrow and brain, leading to death in early childhood.(43) Sphingomyelin storage also leads to unbalanced cholesterol-sphingolipid ratios and severely perturbs raft formation and raft associated functions in the plasma membrane.

Both human and murine AS Ms are products of conserved genes that share 82% identity. The human ASM gene gives rise to a 629 amino acid precursor protein that consists of a saposin-like domain and a phosphodiesterase domain, which is then modified by high mannose oligosaccharide residues. (13) The mannosylated precursor traffics either to the

11 lysosomal compartment (lysosomal sphingomyelinase/L-SMase) or to the extracellular space

(secretory sphingomyelinase/S-Mase). N-terminal amino acid sequence analysis revealed that

L-SMase starts with the amino acid sequence GHP ARLH whereas S-SMase begins with

HPLSPQGHPARLH. The difference in the N-terminal sequence results in different N

terminal proteolytic processing. While L-SMase undergoes a post-translational modification

that results in it' s trafficking to the lysosomes. S-SMase undergoes a post-translational

modification that re-directs the enzyme to the default secretory pathway. (44)

There are six potential N-linked glycosylation sites in human ASM, five of which are

glycosylated. Elimination of any of these glycosylation sites results in various degrees of

decreased enzyme activity due to lack of structural stability and/or misfolding of the

protein.(45)

L-SMase was shown to bind to a mannose 6-phosphate receptor affinity column and

to elute with free mannose 6-phosphate. This finding indicated that L-SMase is transported to

the lysosomes by the mannose 6-phosphate receptor. Although lysosomal ASM activity was

found at a reduced level in fibroblasts from patients with ICD, endocytosis of radiolabeled

extracellular ASM precursor by fibroblasts was not prevented by the addition of free mannose 6-phosphate,(14) indicating the possibility of an alternative pathway independent of the mannose 6-phosphate receptor. Interestingly, the N-terminus of the ASM contains a

sequence highly homologous to the saposin molecules. This finding raises the possibility that the ASM saposin-like domain is involved in lipid binding and/or trafficking of ASM to the

lysosomes. It is also plausible that the ASM saposin-like domain is involved in a protein protein interaction with an alternative sorting receptor to the mannose 6-phosphate receptor.

12 II. Targeting of Soluble Lysosomal Proteins

1. Mannose 6-phosphate dependent transport of lysosomal proteins.

1.1 Biosynthesis and phosphorylation

Soluble lysosomal proteins share a similar N-terminal signal sequence with most secretory proteins, which directs the lysosomal and secretory proteins into the rough endoplasmic reticulum.(l) After the insertion within the membrane of the endoplasmic reticulum, the signal sequence is cleaved and a high mannose oligosaccharide is added to selected asparagine residues. (1) The first reaction unique to the synthe sis of a lysosomal protein occurs shortly after the protein leaves the RER. Phosphomannosyl residues are generated through the action of two distinct enzymes. Firstly, an N-acetylglucosamine-l phosphotransferase transfers N-acetylglucosamine-l-phosphate from UDP-GlcNAc to one or more mannose residues on the lysosomal protein to give rise to a phosphodiester intermediate. Subsequently, an N -acetylglucosamine-l-phosphodiester a-N- acetylglucosaminidase removes the N-acetylglucosamine residue to generate an active phosphomonoester. (46-48)

Study with the lysosomal enzyme cathepsin D indirectly demonstrated that the initial phosphorylation event occurs in a post-RER, pre-Golgi compartment. The conversion of monophosphorylated species to diphosphorylated forms and the hydrolysis of the diester occur within the Golgi.(49) In fact, the attachment of an ER retention signal (KDEL) caused cathepsin D to accumulate within ER. Under this condition the oligosaccharides of the retained cathepsin D were phosphorylated with almost all of the residues being

13 phosphodiesters. Thus, the lack of conversion of phosphodiesters to phosphomonoesters substantiated that such a step occurs exclusively in the Golgi apparatus.

1.2 Phosphotransferase and I-cell disease

Intensive interests on phosphotransferase started from studies of patients with I-cell disease also known as mucolipidosis type II. I-cell disease is an inherited lysosomal storage disorder first described by Leroy and Demars in 1967. One unique feature ofthis disease is the presence of phase-dense intracytoplasmic inclusions in the fibroblasts from patients with this condition. These cells are termed "inclusion cells", and because of this feature the disorder was designated I-cell disease. Spranger and Wiedermann (1970) subsequently classified this disease as mucolipidosis type II (ML-II) since patients with this disorder exhibit clinical characteristics of mucopolysaccharidoses and sphingolipidoses.(50) Further studies revealed that I-cell disease is an autosomal recessive disorder caused by a deficiency of the enzyme UDP-N-acetylglucosamine: N-acetylglucosaminyl-1-phosphotransferase. This enzyme is the product of the GNPTA gene, which has been mapped to chromosome band

4q2l-q23. Cells from patients with I-cell disease express extremely low or undetectable levels of phosphotransferase leading to massive storage of carbohydrates and lipids in the lysosomes while a number of lysosomal enzymes that can degrade these substances are present in excess in tissue culture media and in extracellular fluids, such as serum and urine.(51)

When fibroblasts from I-cell disease patients were grown in the presence of sucrose, the phosphotransferase activity was restored to almost normal level and the activities of hydrolases in the lysosomes increased. One possible explanation for this effect is that the

14 sucrose somehow stabilizes the defective phosphotransferase in treated cells.(52) Moreover,

It was observed subsequently that I-cell disease fibroblasts were able to intemalize and use lysosomal enzymes produced by normal cells, whereas normal or other lysosomal disease fibroblasts were incapable of intemalizing lysosomal enzymes secreted by the I-cell disease fibroblasts.(53) Biochemical comparison oflysosomal enzymes from normal individuals with those from patients with I-cell disease led to the discovery of mannose 6-phosphate as the lysosomal sorting signa1.(54) In I-cell disease, the deficiency ofphosphotransferase prevents the formation of the mannose 6-phosphate recognition marker as the lysosomal enzymes are modified in the Golgi apparatus before being transported to the lysosome. Therefore, in the absence of mannose 6-phosphate many lysosomal enzymes cannot be transported to the lysosomes for normal processing and use.

Since the phosphotransferase plays a very important role in the transport of soluble lysosomal proteins, researchers were interested in how this enzyme worked. The enzyme contains a recognition subunit and a catalytic subunit, which selectively binds and phosphorylates lysosomal proteins. This reaction is based on the ability of the enzyme to recognize a protein domain that is common to all soluble lysosomal proteins which is absent in non-Iysosomal proteins.(55,56) However, no significant sequence identity has been found among c10ned lysosomal proteins. Since linearized or proteolytic fragments of lysosomal proteins do not serve as substrates of the phosphotransferase, it appears that the conformation of the lysosomal protein is important for the recognition by this enzyme instead of a simple linear sequence ofamino acids. (49, 50)

15 Sequence analysis by site-directed mutagenesis of a chimeric protein composed of a region of cathepsin D and the pepsinogen backbone indicated that a group of lysine residues were critical in permitting the phosphorylation of the fusion protein.(57,58) The analysis suggested that the three dimensional closeness of these lysine residues within the cathepsin D reglon was crucial for the recognition of the chimeric protein by the phosphotransferase.( 59,60)

1.3 The mannose 6-phosphate (M6P) receptors

Two distinct type 1 integral membrane glycoproteins that exhibited high affinity for mannose 6-phosphate residues were identified. They were localized in TGN, intracellular vesicles and plasma membrane.(61,62) One of the two M6P receptors is a 46 KDa protein that was originally identified and purified based on its enhanced ligand binding ability in the presence of divalent cations, particularly Mn +2, so it has been referred to as the cation dependent M6P receptor (CD-M6P-Rc). Sequence analysis of the bovine CD-M6P-Rc revealed that it contains a 28-residue N-terminal signal sequence, a 159-residue luminal domain, a single 25-residue transmembrane region and a 67-residue C-terminal cytoplasmic domain. The receptor was shown to have five potential Asn-linked glycosylation sites, four ofwhich were linked to oligosaccharides.(63)

The second receptor is a 215-300 KDa protein, which was also known as the insulin like growth factor II (IGF-II). However, the ligand binding ability of this receptor is independent on the presence of cations. Based on this property this receptor is referred as the cation-independentt M6P receptor (CI-M6P-Rc). Consequently, the CI-M6P-Rc is multifunctional; it binds lysosomal proteins bearing the M6P recognition marker as well as

16 the peptide hormone IGF-II.(64) The IGF-II/CI-M6P-Rc consists of a 44-residue N-terminal signal peptide, a 2269-residue luminal region composed of 15 homologous repeating domains, a single transmembrane region and a 163-residue C-terminal cytoplasmic tail.(64)

Biochemical studies showed that the binding sites for M6P and IGF-II are distinct and that the receptor can bind both ligands simultaneously.(65) However IGF-II can inhibit lysosomal enzyme binding and vice versa.(66) Although the biological significance ofthese findings is uncertain, three possibilities have been suggested: 1) CI-M6P-Rc participates in the clearance ofIGF-II from the circulation; 2) IGF-II binding to CI- M6P-Rc modulates the targeting of lysosomal enzymes; 3) IGF-II binding to CI- M6P-Rc is involved signal transduction. (67)

The luminal domain of CD-M6P-Rc, unlike the transmembrane and cytoplasmic regions, shares sequence similarity with each of the 15 repeating do mains of the CI-M6P-Rc.

Due to their common function in targeting lysosomal proteins and the similarity of their amino acid sequence, it has been postulated that the two receptors evolved from a common ancestral gene with the CI-M6P-Rc arising from the duplication of a single ancestral gene(68) .

Crystallographic analysis of the CD-M6P-Rc revealed the presence oftwo identical molecules complexed with a single M6P molecule. This result was consistent with previous fmdings of chemical cross-linking and equilibrium dialysis experiments that the CD-M6P-Rc is a dimmer containing two M6P binding sites. (69,70)

The quaternary structure of the CI- M6P-Rc has not been analyzed in detail, but available information suggests that this receptor could he a monomer or an oligomer. A CI

M6P-Rc monomer binds 2 moles of M6P or one mole of divalent phosphorylated

17 oligosaccharide. This indicates that only 2 of the 15 repeating motifs of this receptor are essential for M6P binding. (71) Subsequent experiments showed that only domains 3, 9 function in the binding ofM6P. A structure-based sequence alignment generated between the

CD-M6P-Rc and domains 3,9 of the IGF-II/CI-M6P-Rc by maintaining the positions of the cysteine residues involved in disulfide bond, showed that these domains posses conserved residues that are responsible for ligand binding present in the CD-M6P-Rc.(72)

Both receptors bind M6P with essentially the same affinity (7-8x10·6M), however, the

CI-M6P-Rc binds diphosphorylated oligosaccharide with a much higher affinity (2x10·9M) than the CD-M6P-Rc.(71 ,73)

1.4 Role of M6P receptors in lysosomal targeting

In vitro (Le. in cultured cells) the CI-M6P-Rc is the dominant receptor in lysosomal targeting of soluble hydrolases as it is able to compensate for the loss of the CD-M6P-Rc. In fact, cells deficient in the CI-M6P-Rc only sort 30-40% of newly synthesized enzymes.

When antibodies that block ligand binding of the CD-M6P-Rc are added to CI-M6P-Rc deficient cells, the secretion of hydrolases is increased, substantiating the sorting function of the CD-M6P-Rc. However, cells that express both M6P-receptors do not hyper secrete lysosomal enzymes when anti-CD-M6P-Rc antibodies are added. Finally, addition of anti-CI

M6P-Rc antibodies induces lysosomal enzyme secretion, even in the presence of functional

CD-M6P-Rc. Transfection of CI-M6P-Rc deficient cell Hnes with a CI-M6P-Rc cDNA results in complete or almost complete correction of the hyper secretion of lysosomal enzymes. Transfection of a CD-M6P-Rc cDNA to the same cellline partially corrected the

18 hyper secretion of lysosomal enzymes, reinforcing that the CD-M6P-Rc is less efficient than the CI-M6P-Rc in lysosomal protein targeting. (74,75)

The emerging idea is that the CI-M6P-Rc intemalizes extracellular lysosomal enzymes via receptor-mediated endocytosis, while the CD-M6P-Rc is entirely ineffective in the extracellular intemalization of hydrolases. (76) BOth M6P-Rcs bind phosphomannosyl residues optimally in a pH range of 6.3-6.5, but on1y the CI-M6P-Rc can also bind ligands at neutral pH. These explain that the inability of the CD-M6P-Rc to function in endocytosis is due to its poor binding capacity at the cell surface rather than its failure to recycle back to the plasma membrane. (77)

1.5 Final destination of the mannose 6- phosphate receptors

The mannose 6-phophate receptors mediate the delivery of newly synthesized soluble lysosomal proteins to the lysosomes by binding to M6P residues found on N-linked oligosaccharides. In vivo, the resulting M6P-Rc/Lysosomal protein complex is transported from the Golgi apparatus to an acidified, pre-Iysosomal compartment where the low pH of the compartment induces the complex to dissociate. The released lysosomal proteins are delivered into lysosomes while the M6P receptors either return to the trans-Golgi network to repeat the process or move to the plasma membrane where they function to intemalize exogenous ligands. (1)

2. Mannose 6-phosphate independent transport of lysosomal proteins

2.1 Mannose 6-phosphate independent vacuolar targeting in yeast

In the yeast Saccharomyces cerevisiae, the sorting and delivery of vacuolar hydrolases is similar to that of lysosomal proteins in mammalian cells. For instance, yeast

19 mutants defective in vacuole acidification missort newly synthesized enzymes just as mammalian cell mutants defective in endosome acidification. (78)

The precursor of carboxypeptidase Y (CPY), a soluble vacuole protein, has been

shown by site-directed mutagenesis to contain a signal peptide and an N-terminal region necessary for vacuole targeting. Overexpression of CPY leads to missorting and secretion of newly synthesized enzyme, suggesting that CPY is transported to the vacuole by a putative hydrolase sorting receptor. (79-81)

The receptor for Golgi-vacuole transport of CPY was found to be a type 1 integral membrane protein VpslOp. VpslOp was discovered by the isolation of Saccharomyces cerevisiae mutants that missorted CPY. (7) More recently, Vps10p was found to bind CPY via its N-terminal domain. (82) Although VpslOp bears no homology to the M6P receptors its function in the yeast corresponds to the function of the M6P receptors in mammalian cells.

(82) Unlike the M6P receptors VpslOp sorts CPY by the recognition of a peptide sequence instead ofa carbohydrate. (83,84) A recent study showed that the cytoplasmic tail ofVps10p mediated lysosomal sorting in mammalian cells by means of its interaction with GGA

(Golgi-iocalizing y-adaptin ear homology domain ADP-ribosylation factor binding protein).

GGA is a novel ubiquitous coat protein mediating the formation of intracellular transport intermediates and selection of cargo. (85)

Interestingly, the missorting and secretion of CPY due to its overexpression in yeast was not accompanied by missorting of other vacuolar enzymes, suggesting the existence of several sorting receptors.

2.2 Mannose 6-phosphate independent sorting in mammalian cells

20 Although M6P receptors play a major role in the intracellular transport of newly

synthesized lysosomal enzymes in mammalian cells, several lines of evidence suggest the

existence of an alternative mechanism of lysosomal targeting. B-Iymphocytes from patients

with I-cell disease maintain near normal cellular levels of lysosomal enzymes despite their

inability to add mannose 6-phosphate to newly synthesized hydrolases. (86) An I-cell disease

(ICD) B lymphoblastoid cellline targets about 45% of the lysosomal protease cathepsin D to

dense lysosome in the absence of detectable mannose 6-phosphate residues. (3) The secretory

protein pepsinogen is closely related to cathepsin D; however, it is mostly excluded from

dense lysosomes, indicating that the lymphoblast-targeting pathway of cathepsin D is

specific.(3) Analysis of a number of cathepsinD/pepsinogen chimeras indicates that extensive

polypeptide determinants in the cathepsin D C-terminal region confer efficient lysosomal

sorting when introduced into a pepsinogen sequence. These results indicate that a specific

protein recognition event underlies the mannose 6-phosphate independent lysosomal sorting

in ICD lymphoblast. (3)

In fibroblast from ICD patients most hydrolases are not targeted to the lysosomes, and

as a consequence, they are missorted to the extracellular space. (87) However, not all

lysosomal proteins were aberrantly secreted since prosaposin and the GM2AP were found in the lysosomes of the defective cells. (88)

Studies of intracellular trafficking of prosaposin and GM2AP (the cofactor of p hexosaminidase A) in cultured cells transfected with a dominant-negative truncated human

sortilin demonstrated that both prosaposin and GM2AP use the protein sortilin instead of the mannose 6-phosphate receptor to reach the lysosomes.(89,90) A sortilin siRNA also blocked

21 the trafficking of prosaposin and GM2AP to the lysosomal compartment and induced the retention ofthese proteins in the Golgi apparatus. These results further substantiated that the trafficking of prosaposin and GM2AP to lysosomes was dependent on sortilin.

Similar to prosaposin, immunocytochemical analysis of tunicamycin treated cells demonstrated that glycosylation was not essential for the targeting of the AOAH precursor to the lysosomes. (10) This finding suggests that the lysosomal targeting of the AOAH precursor was independent of the mannose 6-phosphate receptor and opens the question as to whether sortilin may be responsible for the targeting of some hydrolases in addition to the sphingolipid activator proteins, prosaposin and GM2AP.

III. The VpslOp superfamily

1. A novel receptor family

Sortilin is a member of the Vpsl0p superfamily. The Vpsl0 proteins are a novel family of heterogeneous type 1 transmembrane receptors. These receptors are expressed in many tissues and they are targets of a variety of ligands.(91,92) ln mammalian cells the

VpS10p family consists of5 members: SorLA (250KDa),(93) sortilin (lOOKDa),(6) SorCS1,

SorCS2 and SorCS3 (130KDa).(91) They are characterized by a domain homologous to the yeast vacuolar sorting protein VPSlOp in their luminal domain and a short cytoplasmic region harboring signaIs for rapid internalization and intracellular sorting. (85) Consequently, they are also known as V ps 1Op domain containing receptors. Their names start with the pre fix "sor" as an abbreviation for "sorting receptor related".

2. The genes ofhuman VpslOp family

22 Analysis of the exon-intron structures of the human VpslOp domain containing receptor genes has shown that they contain many short exons separated by large introns, several of which, extend over more than 50 KB. The three Sorcs genes encompass more than 500 KB of genomic DNA and represent some of the large st known human genes. (92)

The function of the large introns is unknown. However, they may be involved in the regulation of splicing or they may be the result of evolution. The first coding exon of aU human VpslOp receptor proteins has a high G/C content, indicative of CpG inlands. A CG island is a short stretch of DNA in which the frequency of the CG sequence is higher than other regions. It is also caUed the CpG Island, where "p" indicates that "c" and "G" are connected by a phosphodiester bond. Methylation of CpG islands might be involved in the developmental or tissue-specific regulation of the VpslOp receptors. (94)

3. Two common structural features of Vpsl Op receptors

AU members of the VpslOp receptor family are synthesized as pro-receptors containing one or two consensus sequence for c1eavage by furin [RX(RIK) R] in their N terminal region. In the case of sortilin, the furin c1eavage occurs during the passage of the receptor through the trans-Golgi network. (95) Affinity chromatography shows that pro peptide c1eavage is necessary and sufficient for activation of RAP (receptor associated protein) binding to sortilin. (92) Moreover, the sortilin propeptide and two sortilin ligands, neurotensin and RAP, bind to the same or overlapping sites on the luminal domain of the fully processed receptor. (92) The propeptide therefore appears to provide a static hindrance, preventing the ligands from gaining access to the binding site in uncleaved pro-receptors.

(92) Another feature of VpslOp family is a conserved region with 10 cysteines in the C-

23 terminus of the Vps 1Op domain. In fact, this is the most conserved region of aU members of this family. These 10 cysteines form 5 disulfide bridges in sorti lin and SorLA. In sortilin, the deletion of this segment abolishes the binding of the receptor to RAP, providing evidence that the 10 cysteines module constitute the major binding site in the Vps10p domain. (96)

4. Vps10p family receptors bind unrelated ligands

Most members of the Vps10p family bind more than one ligand and exhibit multifunction. SorLA is a highly conserved "putative" sorting receptor located mainly in the

TGN. Taken from the N-terminus, the luminal region comprises a VpslOp domain, a cluster of LDLR (low density lipoprotein receptor re1ated protein), class-A repeats and six fibronectin type III repeats. Thus, it is also called mosaic receptor. SorLA binds neurotensin

(NT) and head activator (HA) via its Vps10p domain, and apolipoprotein E and lipoprotein lipase (LpL) via its cluster of class-A repeats.(93,97,98) SorLA is structurally and functionally related to sortilin and to receptors of LDLR family, and might be involved in protein sorting and signal transduction. (98) SorLA is also implicated in the generation of

Alzheimer's disease as weIl as in atherosclerotic plaque formation.(99,100) Its mouse homologue, mSorLA, exhibits a unique pattern of expression in the developing brain, suggesting an important function in the development of this organ.

Sortilin, also named neurotensin receptor 3 (NTR3), was originally purified from brain extract. (101) Sortilin contains a single VpslOp domain within its luminal region. It binds receptor-associated protein (RAP), neurotensin and prosaposin. (89,101) Thus, sortilin functions as a lysosomal sorting receptor and participates in signal transduction in cooperation with the NTRI. A large pool of sortilin is localized within the Golgi apparatus

24 and intracellular vesic1es. However, in response to insulin, sortilin is translocated with the

glucose transporter Glut4 to the plasma membrane. (102,103) Recent studies showed that

sortilin mediates the uptake and degradation of lipoprotein lipase (LpL). (104) SorCS1 was the first identified member of a sub-group of the mammalian V ps 1Op receptor family that its

luminal region comprises an N-terminal Vps10p domain, followed by a leucine-rich domain that is implicated in protein-protein interactions. Two other known members of the subgroup are SorCS2 and SorCS3. Their ligands are unknown except that mature Sorcs 1 binds its own pro-peptide with low affinity. Given they are most abundant in the developing and mature brain, the SorCSs may have functions in the central nervous system. (105)

5. Structure and functions of the cytoplasmic tails of Vps 1Op receptors

Each VpslOp receptor carries a short cytoplasmic tail of 50 to 80 amino acids comprising typical motifs for interaction with cytoplasmic adaptor molecules. Functional sorting sites, like dileucines, acidic clusters and tyrosine-based motifs, involved in endocytosis and intracellular transports have so far been established in sortilin and SorLA.

(85,106)

Sortilin cytoplasmic domain contains several potential signal sequences that conform to established consensus motifs, known to be involved in adaptor protein binding, endocytosis, basolateral targeting and Golgi-endosome sorting. Furthermore, the sortilin cytoplasmic tail was shown to bind the VHS domain of GGA2, suggesting that sortilin cytoplasmic domain conveys signals for Golgi-lysosome transport. (85)

The 54-residue cytoplasmic tail of SorLA comprises a putative intemalization motif

(F ANSHY), an acidic cluster (DDLGEDDED), and a C-terminal patch of hydrophobic

25 residues (VPMVIA) proceeded by two acidic residues (DD). GGAI and 2 were shown to bind SorLA with differential requirements via three critical residues, two acidic residues and one hydrophobie amino acid in the C-terminal segment of its cytoplasmic tail (104). Unlike sortilin and the mannose 6-phosphate receptor, the GGA binding segment in SorLA contains neither an acidic cluster nor a dileucine motif, which suggests that key residues in SorLA and sortilin conform to a new motif, \l'XX0 for interaction with GGA and ~00 for interaction with GGA2, defining minimum requirements for GGA binding to cytoplasmic receptor domains. (106)

6. Sortilin: a multifunctional protein

Sortilin was originally cloned and purified from mouse and human brain extracts by affinity chromatography with a RAP column. In human, the gene maps to chromosome 1p and encodes an 833-amino-acid polypeptide. Human Sortilin is a 100KDa, non-G-protein coupled type 1 transmembrane protein consisting of a 44-amino acid N-terminal propeptide,

41 44 followed by a furin cleavage signal (R WRR ), a large luminal domain, a single transmembrane region and a short cytoplasmic tail. (6) The major pool of sortilin accumulates in the trans-Golgi network and in vesicles. Sortilin co-localizes with CI-M6P

Re, (107) while about 10% is expressed in plasma membrane. The plasma membrane expression of sortilin is increased after neurotensin-induced sequestration of neurotensin receptor 1. (NTR1)

Sortilin is synthesized as a precursor and converted to mature protein after the removal of the propeptide by furin before or during its passage through the Golgi apparatus.

The propeptide exhibits high affinity binding to the fully processed sortilin, and the binding

26 is competed by RAP and NT. Interestingly, both RAP and NT are unable to bind sortilin in the absence of maturation by furin. Studies by Petersen et al. (2004) have shown that the pro~ peptide of sortilin has at least two functions. First, sortilin depends on its propeptide for proper protein folding and normal passage through the biosynthetic pathway. (95)

Additionally, the sortilin propeptide acts as a safeguard protecting the cells against formation of death~signaling intracellular complexes. (93) In contrast, other VpslOp receptors, such as

SorLA and SorCS3, do not need their propeptide for their normal processing. (96) Finally, unlike other VpslOp members, the luminal domain of sortilin exclusively comprised a single

Vps10p domain. (6)

Sortilin appears to he a multifunctional protein since it is capable to bind different ligands such as neurotensin, RAP, prosaposin and LpL.(6,89,101,104) Sortilin is also engaged in intracellular sorting as well as in endocytosis and signal transduction.(108,109)

For instance, sortilin mediates rapid endocytosis of lipoprotein lipase, neurotensin (NT) and the precursor ofnerve growth factor (pro~NGF).(104,110) It also forms a complex with the G protein~coupled neurotensin receptor~ 1 on the plasma membrane modulating the NT signaling, and is able to target lysosomal proteins such as prosaposin and GM2AP to the lysosomes.(90,111) Sortilin is essential to pro-NGF induction ofneuronal death via complex formation with p75NTR on the cell membrane.(112) A recent study on the hydrophobic protein conotoxin~ TxVI demonstrates that sortilin interacts with TxVI in the ER and facilitates its export from ER to Golgi. (113)

The cytoplasmic do main of sortilin lacks characteristic sequence for signal transduction, but shares a homologous sequence with the CI-M6P~Rc, which binds GGAs.

27 Cellular trafficking of chimeric receptors containing the luminal and transmembrane domains

of the M6P-Rc followed by the sortilin cytoplasmic tail showed that the sorting signaIs

'PXX0 and dileucine motifs of sortilin mediates the rapid endocytosis of the chimeric protein, which subsequently was found in the TGN. The CI-M6P-Rc/sortilin chimera was

almost as efficient as the CI-M6P-Rc itself for the transport of newly synthesized lysosomal proteins to the lysosomes.(8S) Furthermore, using the cytoplasmic tail of sortilin as bait, yeast two-hybrid experiment showed that sortilin binds the VHS domain of the cytosolic

sorting protein GGA2.(8S)

In conclusion, the structural features of the luminal and cytoplasmic do mains of sortilin indicate that this protein functions as a sorting and trafficking receptor for lysosomal

soluble proteins. Nonetheless, sorne researchers believe that sortilin is also a neuropeptide receptor, which unlike other receptors, does not belong to the famiIy of the G-protein coupled receptors.

IV. GGAs are required for the sorting soluble lysosomal proteins

1. A novel family proteins interacting with ADP-ribosylation proteins (ARFs)

The GGA (Golgi-iocalizing, y-adaptin ear homology do main, ARF-binding protein) proteins constitute a multidomain protein family implicated in trafficking between TGN and endosomes. (114) Using a yeast two-hybrid screen ofhuman cDNA libraries with activated

ARF3 as the bait, two proteins, GGA1 and GGA2 were discovered nearly simultaneously by severallabs. (115-118) GGA3 was identified in a BLAST search of the GeneBank database.

(111) GGAI and GGA2 were also found in the yeast. (119) GGA proteins are conserved

28 throughout eukaryotes and aIl the GGA proteins identified so far are 65-80 KDa in size.

(119)

Immunoelectron lllicroscopy analyses showed that human GGAs localized to electron-dense coats associated with TGN membranes, suggesting a function in the TON region. The three human GGAs showed overlapping but subtle differences in subcellular staining. In addition to their shared TGN localization, GGA2 showed a diffuse cytosolic distribution, GGA3 showed a granular cytoplasmic pattern. (116,118)

2. Structure of GGA proteins

Alignment of amino acid sequences of human and yeast GGAs demonstrates that

GGAs consist of four domains. The N-terminal 150 residues constitute the "VHS" domain.

This domain was so named because it was initially found in three proteins: Vps27, Hrs and

STAM. The crystal structure of the VHS domain from another protein, Tom1, suggests that the VHS domain mediates protein-protein or prote in-membrane interaction. (120) The most highly conserved do main is the 170 residues long "GAT" domain. With approximately 65% identity among GGAs, the GAT domain contains two coiled-coil regions. The name ofthis do main derives from the sequence homology between GGA and Tom1. (117) The role ofthe two coiled-coil regions is unc1ear, since GGA appears to he monomeric in the cytosol. (116)

The "HINGE" domain, of variable lengths among GGAs, contains one or more clathrin binding motifs. Two conserved c1athrin-box motifs are present in the GGA2 hinge (LIDLE and LLDLL), whereas only one motifhas heen found in the GGA1 hinge domain. (117) The

C-terminal "GAE" domain is composed of 120 residues and has homology to the ear domain of y-adaptin.

29 3. Function ofGGAs in the yeast

In yeasts, GGA1 and GGA2 share 50% amino acid identity. While deletion of either

GGA alone causes minor or no defects, deletion ofboth GGAs lead to remarkable trafficking

defects of carboxypeptidase Y (CPY) and carboxypeptidase S (CPS).(121) Furthermore, a

chimeric Pep 12p protein normally delivered from the TGN to late endosomes, is missorted

into early endosomes in yeast strains lacking both GGA1 and GGA2. These data suggests

that GGAs are key components of a specifie pathway from the TGN to late endosomes. (122)

4. GGAs interacting proteins

In mammalian cells, the VHS domain of GGA proteins was shown to interact with

the dileucine-sorting motif present in the cytoplasmic taïl of CI-M6P-Rc and CD-M6P-Rc.

(123,124) Similarly, GGAs bind to the acidic-cluster-dileucine motif (ACLL) of the

cytoplasmic tail of sortilin (8,85) as well as the lipoprotein receptor-related protein 3 (LRP3).

(8) SorLA, a sorting receptor of the VpslOp family, binds GGAs 1 and 2 through a methionine-based signal and a pair ofpreceding acidic residues. (106)

The GAT domain of GGAs binds specifically GTP-bound ARP proteins found in the membrane of the TGN. This indicates that GGAs are effectors of ARP.(125) A number of experiments show that the localization of GGAs at the TGN is due to interaction with ARP since the VHS domain alone cannot recruit GGAs to TGN membrane. However, ARP is absent from clathrin-coated vesicles containing GGAs. Therefore, it is likely that ARP binds and recruits GGAs to the TGN membrane and then hands-offGGAs to cargos. (115-117,123)

GGAs and cargo are then incorporated into a transport intermediate that excludes ARP. (126)

30 Finally, while the HINGE domain of GGAs interacts with clathrin in vitro and

promotes the recruitment of clathrin to membranes, the GAE do main strengthens this

interaction. Nevertheless, the function of the GAE domain itself is still unclear. (127) In

conclusion, the GGAs are monomeric, ARF-dependent clathrin adaptors involved in ARF

and clathrin interactions required for the trafficking of cargo to the lysosomes.

5. Sortilin/GGA-mediated pathway and the sorting of ASM

Since the mechanism for lysosomal targeting of ASM has not been conclusively

discemed, the aim of this thesis is to investigate whether or not sortilin functions as an

alternative sorting receptor for targeting ASM to the lysosomes. To address this question, we

examined the effects of two dominant-negative competitors that are known to block the targeting of sortilin to the lysosomes. The negative competitors are a truncated sortilin

lacking its cytoplasmic region and a truncated GGA lacking the clathrin binding sequence.

These dominant-negative proteins were overexpressed in COS-7 cells and ICD cells to

analyze their effect on the transport of ASM to the lysosomes. In addition, the direct binding between sortilin and ASM was investigated in this study.

31 MATERIALS AND METHODS

Reagents and antibodies

DMEM, methionine, cysteine-free DMEM, fetal bovine serum, diaIyzed fetai bovine

serum; penicillin-streptomycin and L-glutamine were purchased from Invitrogen. DNA polymerase for PCR reaction was Turbo Plu from Stratagene. DNA was purified using High

Speed Midi-prep Kit from Qiagen. Polyfect Transfection Reagent was purchased from

Qiagen. Restricted enzymes were from New England BioLab. The DNA Ligation Kit was from Roche. esS] methionine was from Amersham Biosciences. Nitrocellulose blotting membrane (0.45J.1m thick) was from BioRad. Complete protease inhibitor cocktail was from

Roche. Paraformaldehyde was from Fisher Scientific. AU the other chemicals and reagents were purchased from Sigma.

Anti-myc antibody (9EI0), rabbit anti ASM polyc1onal antibody, monoclonal anti

Lamp 1 antibody, goat anti-mouse HRP, goat anti-mouse FITC and goat anti-mouse TR antibodies were from Santa Cruz Biotechnology. Anti-Golgin 97 antibody, goat anti- rabbit

Alexa 488, goat anti-rabbit Alexa 594 and lysotracker Red DN99 were from Molecular

Probes. PolYc1onal rabbit anti-sortilin antibody was a generous gift from Dr. C.M. Petersen

(University of Aarhus).

DNA constructs

ASM-myc encoding full-Iength mouse acid sphingomyelinase was produced by PCR amplification of complete cDNA of Mus musculus sphingomyelin phosphodiesterase 1

(cDNA clone MGC: 25355 IMAGE: 4482098), followed by ligation ofthe PCR product into pcDNA3.1 digested with BamHI and EcoRI.

32 WtSortilin-EGFP encoding full-Iength human sortilin was produced by PCR amplification of sortilin cDNA in pBK-CMV (provided by Dr.C.M. Petersen), followed by ligation of the PCR product into pEGFP vector digested with Agel and }{ho!.

DnSortilin-EGFP encoding luminaI and transmembrane domains ofwild type human sortilin was produced by PCR amplification followed by ligation of the PCR product into pEGFP vector digested with Agel and }{ho!. Sortilin-myc construct was previously made in our labo (89) Truncated GGA-EGFP construct is a kind gift from DrJuan S. Bonifacino

(NIH).

Celllines and cell culture

COS-7 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10 % fetal bovine serum, penicillin, streptomycin and L-glutamine, and maintained in an atmosphere of5% C02 at 37°C.

Fibroblasts from patients with Mucolipidosis Il (I-cell disease) were provided by Dr.

Mirella Filocamo, Diagnosi Pre-Postnatale MaIattie Metaboliche-lstitueo.G.Gaslini. (ItaIy).

The cells were cultured in Dulbecco's modified Eagle's medium supplemented with 15 % fetaI bovine serum, penicillin, streptomycin and L-glutamine, and maintained in an atmosphere of 5% CO2 at 37°C. The fibroblasts used in the experiments were not oider than

15 passages.

Transfection of COS-7 cells

The day before transfection, 1.5x105 COS-7 cells were seeded in each 60 mm dish in

3 ml of DMEM containing 10% fetai bovine serum. Cells were incubated at 37°C and 5%

C02 in an incubator. On the day of transfection, the cells were 70-80% confluent. The oid

33 medium was changed with 1.5 ml of fresh DMEM containing serum. 1.5J-lg of DNA was mixed with 100J-lI of serwn-free, antibiotics-free DMEM, and then 10J-l1 of Polyfect

Transfection Reagent was added to the DNA solution. The DNA-PolyFect-DMEM solution was mixed and incubated at room temperature for 10 minutes. After that, 600J-l1 of DMEM containing serwn was added to the DNA-PolyFect-DMEM solution, mixed with it and then the total volume of 700 J-li of mixture were transferred immediately to the 60 mm dish. The transfected cells were incubated at 37°C and 5% CO2 for 24 hours to allow gene expression.

Electroporation

Four hours before electroporation, cell culture medium was changed. Immediately before electroporation, the fibroblast cells were suspended with trypsin-EDTA, washed with ice-cold EPBS twice. Then the cell number was counted, and 5xl06 cells were resuspended in 800 J-li of ice-cold EPBS containing 20J-lg plasmid DNA. The cell-DNA mixture was transferred into pre-chilled sterilized 0.4mm cuvette and incubated on ice for 5 minutes. The electroporation was preformed at 250 volts, 960 J-lF or 300 volts, 500 J-lF on Genepuiser II

(BioRad). After pulse, the cells were incubated on ice for 5 to 10 minutes, and then cultured in DMEM with 15% fetai bovine serum at 37°C, 5% COz.

Immunofluorescence staining and confocal microscopy

24 hours before transfection, cells were subcultured and seeded into 24-well plate on sterilized coverslips. 24 hours after transfection, cells were washed with PBS twice and fixed with 4% paraformaldehyde in PBS for 10-15 minutes at room temperature. After fixation, cells were washed with PBS containing 20 mM glycine three times. Then cells were blocked with PBS containing 2% bovine serwn albumin, 2% normal goat serwn and 0.05% saponin

34 for 30 minutes at room temperature. Primary antibody incubation was performed in a 1:100 dilution at room temperature for 1 hour or in a 1:500 dilution at 4°C ovemight. The cells were washed three times with PBS, and then incubated with 1:500 diluted secondary antibody conjugated with fluorescence dye for 1 hour at room temperature. The cells were then washed with PBS three to five times, followed by washing with double-distilled water twice. The coyer slips were mounted with Geltol and the slides were kept at 4°C ovemight before viewing. The immunofluorescence staining results were analyzed with a confocal microscope LSM 410 (Carl Zeiss).

Co-immunoprecipitation assay

COS-7 cells transfected with pCDNA3.1 or ASM-myc were lysed in buffer containing 150mM NaCI, 2.5mM KCI, 10mM glycine, 0.1% Triton X-lOO, 50mM Tris.CI pH7.4 and a protease inhibitor cocktaiL Subsequently, Img of protein lysate was prec1eared with 50J.LI of protein G-Sepharose beads at 4°C for 1 hour. The supematants were incubated with protein G-Sepharose beads coupled with 1J.Lg anti-ASM antibody at 4°C overnight.

Beads were washed three times in ice-cold lysis buffer, one time in ice-cold PBS and boiled in 3x SDS-sample buffer. Proteins were separated on 7.5% SDS-PAGE and revealed by immunoblotting with a polyclonal anti-sortilin antibody or with a monoclonal anti-CI-MPR antibody.

Metabolic labeling and pulse-chase experiment

Fibroblasts from patients with ICD were transfected with a pEGFP vector with or without a subcloned dominant-negative sortilin cDNA. 24 hours after transfection, cells

(lOOmm plate) were washed twice with sterile PBS supplemented with 1mM MgClz and

35 O.lmM CaC}z. Starvation of the ceUs was performed for 2 hours in 3ml of methionine-free, cysteine-free DMEM supplemented with 5% dialyzed fetal bovine serum. Then cells were pulse-labeled by addition of 150llCi/ml esS] methionine into starve medium for 1 hour. CeUs were chased with DMEM containing 10% fetal bovine serum and 5nM non-Iabeled methionine and 2.5nM non-labeled cysteine for 0 hour, 0.5 hour, 1 hour, 2 hours and 3 hours, respectively. In this study COS-7 ceUs were also used for similar studies. 24 ho urs after transfection, the COS-7 cells were starved for 1 hour, pulse labeled for 0.5 hour and chased for 0 hour, 0.5 hour, 1 hour and 1.5 hours, respectively. After chasing, the culture medium was collected into a new test tube and the cellular debris was removed by centrifuging the medium at 1500 rpm at 4°C for 10 minutes. The ceUs were washed twice with ice-cold PBS and harvested in lml of lysis buffer (PBS containing 0.5% NP-40, lOmM EDTA and a protease inhibitor cocktail.) The ce Ils were then lysed on ice for 20 minutes. The supernatant of ceU lysate were coUected by centrifugation at 12,800 rpm at 4°C for 10 minutes. Pre treated media and celI lysate were pre-cleared with 50111 of protein A-Sepharose beads by incubation at 4°C for 1 hour. Following pre-clear, the media and the cell lysate were

Încubated with 50 III of protein A-Sepharose beads coupled with antÎ-ASM antibody at 4°C overnight. The next day, the immunocomplex-beads were washed with ice-cold lysis buffer for 3 times and with Îce-cold PBS once. The immunocomplexes were eluted from protein A

Sepharose beads with 50111 of3x SDS-PAGE sample buffer. The samples were boiled for 10 minutes at 90°C and subjected to reducing SDS-P AGE, analyzed by autoradiography.

36 RESULTS

Lysosomal targeting ofASM

In order to he exported from the Golgi apparatus, a soluble lysosomal hydrolase such as ASM must bind a membrane associated sorting receptor that interacts with adaptor proteins. This interaction initiates the recruitment of clathrin and formation of cargo vesicles.

Both the CI-M6P-Rc and the CD-M6P-Rc are well-characterized sorting receptors, which bind and route most soluble hydrolases to the lysosomes through M6P groups tagged to the enzymes. Since ASM is a highly glycosylated molecule, it was believed that it might be transported by one or both M6P receptors. However, recent studies suggested that ASM might bind an alternative membrane associated receptor, since competition with free mannose 6-phosphate did not completely prevent the internalization of ASM from the media of cultured cells.

To investigate if ASM is transported by another sorting receptor; we examined the intracellular localization of ASM in fibroblasts from patients with I-cell disease (ICD). These fibroblasts are deficient in phosphotransferase activity and they are unable to add mannose 6- phosphate residues to lysosomal enzymes. Due to this deficiency, ICD fibroblasts missort most soluble hydrolases, which are released to the extracellular space by a default mechanism. Immunofluorescence staining of ASM-myc transiently expressed in ICD cells showed that anti-myc antibody stained the perinuclear (Golgi) region and peripheral granular structures. (Figs.1A, ID). The granular staining of ASM-myc co-Iocalized with lysosomes that were stained with lysotracker DN99 and with anti-Lamp 1 antibody (Figs.1C, IF). These

37 results suggest that ASM is capable of trafficking to the lysosomes in the absence of mannose 6-phosphate tagging.

In a second set of experiments we compared the lysosomal staining of ASM between normal (Figs. 2A-C) and ICD fibroblasts (Figs. 2D-F) using confocal microscopy. Our results showed that the granular ASM staining was weaker in ICD cells. This fmding suggests that ASM reached the lysosomes via the M6P-Rc and that in the absence of a functional M6P-Rc pathway, ASM reach the lysosomes via an alternative receptor.

Sorti/in interacts with GGA through its cytoplasmic domain

Evidences from our lab as weIl as from other laboratories have implieated sortilin in the targeting of lysosomalproteins in mammalian cells.(85,89) Similarly, the yeast sortilin homologue, VPS 10, bas been implicated in the vacuolar targeting of carboxypeptidase Y.(7)

Structural analysis of sortilin showed that its luminal and cytoplasmic domains exhibited high homology to receptors involved in intraeellular sorting and trafficking. Interestingly, the cytoplasmic tail of sortilin is essential to target this prote in to the lysosomes.

Thus, to eorroborate our hypothesis that sortilin is the alternative reeeptor of ASM, we used four sortilin constructs. Two full-Iength human sortilin eonstruets were linked either to an EGFP tag or to a myc tag (Fig. 3A). Two other construets consisted oftruncated human sortilin lacking its cytoplasmic tail linked to either an EGFP tag or a myc tag (Fig. 3 B).

Western Blotting showed that the full-iength sortilin-EGFP construct expressed a 125 KDa fusion protein in COS-7 ce Ils, and the truncated sortilin-EGFP construct expressed a fusion protein of a slightly lower molecular weight (Fig. 3C). Confocal microscopy demonstrated that full-Iength sortilin-EGFP was predominantly located in the perinuclear (Golgi) region

38 and in peripheral granular structures (Fig. 4A). Compared to full-length sortilin-EGFP, truncated sortilin-EGFP was retained in the perinuclear region (Fig. 4D). The perinuclear staining was confirmed to correspond to the Golgi apparatus since the anti-sortilin staining overlapped with anti-Golgin 97 staining (Figs. 4E,F).

Based on previous results published by our laboratory, we predicted that the interaction between the cytoplasmic tail of sortilin and GGAs is crucial for the trafficking of sortilin from the Golgi to the endosomal compartment. To confirm the interaction of sortilin and GGA, we co-transfected full-Iength sortilin-myc and a dominant-negative GGA-EOFP into COS-7 cells (Figs. 5D-E). The dominant-negative GGA-EGFP construct encodes a truncated GOA containing the VHS and GAT domains only. This GOA construct is unable to bind clathrin due to the absence of the HINGE and EAR domains. The immunostaining of full-length sortilin-myc was exclusively located within the perinuclear (Golgi) region (Fig.

5D) of cells expressing dominant-negative GGA-EGFP (Fig. 5E). This result suggests that dominant-negative GGA blocked the exit of sortilin from Golgi apparatus (Fig. 5F).

Intracellular localization of fuU-Iength sortilin-myc (Fig. 5A-C) in COS-7 ceUs expressing

EGFP was similar to that of full-Iength sortilin-EGFP (Fig. 4A). Co-immunoprecipitation experiments in COS-7 ceUs showed that full-length sortilin bound GOA while sortilin lacking its cytoplasmic tail did not bind GGA (Fig. 6).

Dominant-negative sortilin altered the lysosomal targeting ofASM

To further address the possibility that sortilin is the alternative sorting receptor of

ASM, we co-transfected COS-7 ceUs with ASM-myc and with the dominant-negative sortilin construct lacking the cytoplasmic tail (Figs. 7G-I). Although the dominant-negative sortilin is

39 capable of binding a cargo molecule it is unable to exit the Golgi apparatus. The truncated sortilin was linked to an EGFP tag to be distinguished from endogenous sortilin. COS-7 cells co-transfected with ASM-myc and pEGFP vector (Figs. 7A-C) or full-Iength sortilin-EGFP

(Figs. 7D-F) were used as controls. The ASM-myc expressed in the transfected cells was immunostained with chicken anti-myc antibody followed by incubation with a goat anti chicken IgG conjugated with Alexa 594 fluorescence. The lysosomal targeting of ASM-myc in the presence or absence of truncated sortilin was analyzed by confocal microscopy. The anti-myc antibody yielded a perinuclear reaction characteristic of Golgi staining and a granular cytoplasmic staining characteristic of lysosomes in the cells co-transfected with

ASM-myc and pEGFP or full-Iength sortilin-EGFP (Figs. 7A, D). In the cells co-expressing

A.8M-myc and truncated sortilin-EGFP, the granular cytoplasmic staining was reduced as compared to the control cells (Fig.7G). Thus, truncated sortilin appears to partially inhibit the lysosomal targeting of ASM in COS-7 cells.

ASM targeting is abolished when the M6P and sortilin pathways are blocked

In this set of experiments we examined the consequence of ASM targeting in M6P and sortilin pathways deficient cells. First, we examined the ASM staining pattern in COS-7 cells co-transfected with ASM-myc and truncated GGA-EGFP (Figs. 8D-I). The truncated

GGA-EGFP construct lacks the HINGE and EAR domains, which in turn leads to a failure of clathrin recruitment by GGA in the trans-Golgi network. Under this condition we found that the lysosomal transport of ASM was completely prevented by truncated GGA and that immunostaining of ASM-myc was restricted to the Golgi region (Figs. 8D, G). Furthermore,

ASM-myc co-Iocalized with truncated GGA-EGFP (Figs. 8F, 1). Similar staining pattern was

40 observed in ICD fibroblasts co-expressing ASM-myc and truncated sortilin-EGFP (Figs.9D

F). However, lysosomal associated membrane protein 1 (LAMP 1) attained lysosomes in ICD fibroblasts that transiently expressing truncated sortilin-EGFP (Figs. 9A-C). Because LAMP

1 is targeted to lysosomes via an adaptor protein complex 3(AP-3) mediated pathway. These results indicate that the lysosomal targeting of ASM is mediated by both M6P-Rc and sortilin, which in turn are dependent on GGAs.

Dominant-negative sortilin enhanced the secretion ofASMe

In I-cell disease, the deficiency of phosphotransferase causes the misssorting of lysosomal enzymes to the extracellular space. Under this condition, several hydrolases have been detected to be present in excess in serum and urine.(87) Therefore, we postulated that the inhibition of the sortilin pathway enhances the release of ASM to the extracellular compartment in ICD cells. To this end we investigated the fate ofnewly synthesized ASM in the presence of over-expressed truncated sortilin in COS-7 cells and ICD fibroblasts (Fig.

10).

COS-7 cells transiently co-expressing ASM-myc and truncated sortilin-EGFP or pEGFP empty vector were starved for 60 minutes in the methionine, cysteine-free DMEM, pulse labeled with 150,...,ci/ml CsS] methionine for 30 minutes and chased in DMEM supplemented with 5mM unlabeled methionine and cysteine. Cell lysates and media were immunoprecipitated 0, 30, 60 and 120 minutes after the initiation of chasing with anti-myc antibody. The ASM signal in the precipitates were resolved by SDS-PAGE and visualized by autoradiography. The results showed that 90 minutes after chasing, more ASM was found in the culture media of cells over-expressing truncated sortilin as compared to the culture media

41 of control cells. The amc;>unt of ASM in the lysates had no obvious difference between cells over-expressing truncated sortilin and control cells (Fig. 10A).

To further study the involvement of sortilin in the sorting of ASM, we performed pulse-chase experiment in ICD fibroblasts. ICD cells were transfected with pEGFP empty vector or with truncated sortilin-EGFP via electroporation. Twenty-four hours later, the expression of transfected EGFP protein was examined under fluorescence microscope. The cells were subsequently starved for 2 hours, pulsed with 200J.1cilml e5S] methionine for 1 hour and chased for 0, 0.5, 1, 2 and 3 hours, respectively. Cell lysates and media were immunoprecipitated by anti-ASM antibody. Analysis by autoradiography revealed that in the presence oftruncated sortilin, secretion of ASM into the culture media started as early as half hour after chasing, whereas in the absence of truncated sortilin, secretion of ASM was delayed. On the other hand, nascent ASM in the lysates of cells expressing truncated sortilin disappeared faster as compared to the control cells (Fig. 1OB). In addition, a semi-quantitative analysis showed that during the whole chasing process, more ASM was secreted into culture media of cells expressing truncated sortilin (Fig. 10C). These results suggest that truncated sortilin enhanced the secretion of ASM in COS-? cell and ICD cells.

Sorlilin binds and interacts with ASM.

To verify a potential interaction between ASM and sortilin, co-immunoprecipitation assays were performed using COS-? cells. Considering that the binding of ASM to sortilin should occur in the lumen of the trans-Golgi network, the optimal pH of our incubation reaction was the equivalent of the intraluminal pH of this compartment, i.e., pH 6.5. First, non-transfected COS ? cells (Fig. lIA, lanel) and cells transiently transfected with

42 pCDNA3.l (Fig. lIA, lane2), sortilin-myc (Fig. lIA, lane3), sortilin-myc and ASM-myc

(Fig.IIA, lane4) were lysed and immunoprecipitated with anti-sortilin antibody coupled protein A-Sepharose beads, the precipitates were separated on 7.5% SDS-PAGE, after transferring, the pulled-down protein was detected with anti-myc antibody (9ElO). The two bands in lane 4 indicated that ASM-myc was co-precipitated with sortilin-myc. Inversely, figure lIB showed that in COS-7 cells transiently transfected with ASM-myc, sortilin was co-immunoprecipitated with ASM-myc (Fig. lIB, lane2). On the other hand, in COS-7 cells transiently transfected with pCDNA3.1 empty vector, sortilin was not pulled-down by anti mye antibody (Fig.IIB, lanel).

43 Fig.1 Intracellular localization of ASM in ICD cells.

(A, D): ICD cells were transfected with ASM-myc and stained with a chicken antibody to myc epitope followed by Alexa 594 conjugated goat anti-chicken IgG (A) or Alexa 488 conjugated goat anti-chicken IgG (D). (B, E): The lysosomes of ICD cells were stained with mouse monoclonal anti-Iamp 1 IgG followed by FITC conjugated goat anti-mouse IgG (B) or incubation with Lysotracker DN99 Red (E). (C): The merged image of A and B. (F): The merged image of D and E. Note that ASM is present in the lysosomes ofICD cells.

Fig.2 Lysosomal targeting of ASM in normal fibroblasts and in ICD cells.

(A-C): Normal fibroblasts were transfected with ASM-myc and stained with a chicken antibody to mye epitope followed by Alexa 594 conjugated goat anti-chicken IgO. (D-F):

ICD cells were transfected with ASM-myc and stained with chicken antibody to mye epitope followed by Alexa 594 conjugated goat anti-chicken IgO. The immunostaining in ICD cells is decreased.

Fig. 3 Expression of sortilin-EGFP.

EGFP tagged full-Iength sortilin construct (A) and EGFP tagged truncated (dominant negative) sorti lin lacking its cytoplasmic tail (B) were in vivo translated in COS-7 cells and run on a 10% SDS-PAGE. The full-Iength sortilin-EGFP is seen as a 125 KDa (lane 2), while the truncated sortlin-EGFP was slightly smaller (lane 3). COS-7 cell transfected with pEGFP empty vector was used as a negative control. A

Il Propeptide • Cytoplasmic tail • Luminal domain • EGFPtag = Transmembrane region

C 1 2 3 CTL FLSOR ASOR

-..~ 1 ..... 130KDa ..... 100KDa Fig.4 Intracellular localization of full-Iength sortilin and truncated sortilin.

(A, D): COS-7 cells were transfected with full-Iength sortilin-EGFP (A) or truncated sortilin

EGFP (D). (B, E): Golgi apparatuses were stained with mouse monoclonal anti-Golgin 97

IgG followed by Texas Red conjugated goat anti-mouse IgG. (C): A merged image of A and

B. (F): A merged image of D and E. Note that full-Iength sortilin is seen in the Golgi apparatus and peripheral granules whereas truncated sortilin is restricted to the Golgi apparatus.

Fig.5 Effects of truncated GGA on intracellular trafficking of sortilin.

(A-C): COS-7 cells co-transfected with a plasmid encoding EGFP (B) and a plasmid encoding sortilin-myc were stained with chicken antibody to myc epitope (A) followed by

Alexa 594 conjugated goat anti-chicken IgG. (D-F): COS-7 cells were co-transfected with a plasmid encoding a truncated GGA-EGFP (E) and a plasmid encoding sortilin-myc, and stained with chicken antibody to myc epitope (D) followed by Alexa 594 conjugated goat anti-chicken IgG. (C): A merge of image A and B. (F): A merge of image D and E. Note that truncated GOA causes Golgi retention of sortilin.

Fig.6 Co-immunoprecipitation assay showing association of sortilin and GGA.

(A): Western blotting of cell lysates from COS-7 cells transfected with pCDNA3.1 empty vector (lane1), full-length sortilin-myc and GGA-EGFP (lane2), truncated sortilin-myc and

GGA-EGFP (lane3). The upper bands are full-length and truncated sortilins, the lower bands are GGAs. (B): In vivo co-immunoprecipitation assays using the same lysates as indicated in

(A). Lane 2 demonstrated that full-Iength sortilin was pulled down by GGA. Lane 3 demonstrated that truncated sortilin was not pulled down by GGA. ...Jo, ...Jo, ...Jo, (JI ...... 00 N CO NOCO 0 CCC C C "1» 1»" " 1» " 1» "1» >

...... (\~< N ~ d'O w ~ ~)( \PO QG' ~)( ~~ QG' Q~ ...Jo, ...Jo, ...Jo, (JI ...... 0 0 N ~~~ CO N 0 CO 0 Q~ C C C C C 1»" 1»" 1»" "1» 1»" = ~

(\~ < Fig.7 Effects offull-Iength and truncated sortilins on lysosomal targeting of ASM.

(A-C): COS-7 cells co-transfected with a plasmid encoding EGFP (B) and aplasmid encoding ASM-myc, stained with chicken antibody to myc epitope (A) followed by Alexa

594 conjugated goat anti-chicken IgG. (D-F): COS-7 cells were co-transfected with a plasmid encoding full-Iength sortilin-EGFP (E) and a plasmid encoding ASM-myc, stained with chicken antibody to myc epitope (D) followed by Alexa 594 conjugated goat anti-chicken

IgG. (G-I): COS-7 cells were co-transfected with a plasmid encoding truncated sortilin-EGFP

(H) and a plasmid encoding ASM-myc, stained with chicken antibody to myc epitope (G) followed by Alexa 594 conjugated goat anti-chicken IgG. The third image in each row (C, F,

1) is a merge of the corresponding tirst and second images. Note that truncated sortilin decreased the granular staining of ASM.

Fig.8 Effect oftruncated GGA on lysosomal trafficking of ASM in COS-7 Cells.

(A-C): COS-7 cells co-transfected with a plasmid encoding EGFP (B) and a plasmid encoding ASM-myc, stained with chicken antibody to myc epitope (A) followed by Alexa

594 conjugated goat anti-chicken IgG. (D-I): COS-7 cells co-transfected with a plasmid encoding truncated GGA-EGFP (E, H) and a plasmid encoding ASM-myc, stained with chicken antibody to myc epitope (D, G) followed by Alexa 594 conjugated goat anti-chicken

IgG. The third image in each row (C, F, 1) is a merge of the corresponding first and second images. Note that truncated GGA abolished, the transport of ASM to the lysosomes.

Fig.9 Effect of truncated sortilin on lysosomal targeting of ASM in ICD cells.

(A-C): ICD cells transfected with truncated sortilin-EGFP (B) were stained with anti-LAMP

1 antibody (A) followed by Texas red conjugated goat anti-mouse IgG. (D-F): ICD cells were co-transfected with ASM-myc (D) and truncated sortilin-EGFP (E), stained with chicken antibody to myc epitope followed by Alexa 594 conjugated goat anti-chicken IgG. (C): A merged image of A and B. (F): A merged image of D and E. Note that truncated sortilin abolished the ASM granular staining in ICD cells.

Fig.lO Influence oftruncated sortilin on the sorting of ASM in COS-7 cells and ICD cells.

COS-7 cells (A) or ICD cells (B) transfected with truncated sortilin or pCDNA3.l empty vector were pulsed for 30 or 60 minutes with esS] methionine and chased for variant times as indicated. Cell lysates and media were immunoprecipitated with anti-ASM IgG and the precipitates were separated by SDS-PAGE (7.5%) and visualized by autoradiography. (C):

Since the I-cell experiment was the main experimental condition; a quantitative analysis was performed by measuring the optical density (OD) of the bands. The quantitation revealed that

ASM levels were lower in the lysates and higher in the media of I-cells that were transfected with truncated sortilin (purple line). A ceillysate cultUre media

ASM-MYC+pCDNA3.1 ASM-MYC+ASOR ASM-MYC+pCDNA3.1 ASM-MYC+4S0R

0' JO' 60' 90' 0' JO' 60' 90' 0' JO' 60' 90' 0' JO' 60' 90'

B 4S0R-EGFP ASOR-EGFP ••••••••• ••• • ••• 0' JO' 60' 120' 180' 0' JO' 60' 120' 180' 0' 30' 60' 120'180' 0' 30' 60' 120'180' c

asm in the Iysates of 1cells asm ln the culture media of 1cells

16 14 14 12 12 10 J 10 --+-Seriesl 1 8 ---+--Series1 ' !! 8 __ Series2 . i 6 ----Series2, 1 6 4 4 2 2 o 4 2 4 5 chasingtime challingtime Fig.lI Co-immunoprecipitation showing interaction between sortilin and ASM.

(A): The lysates of non transfected COS-7 cells or COS-7 cells transfected with empty vector, sortilin-myc or sortilin-myc and ASM-myc were immunoprecipitated with anti sortilin antiserum, blotted with anti-myc IgG, ASM was pulled down by sortilin (lane 4).

Note that no ASM was pulled down in the control (lane 1,2,3). (B): The lysate ofCOS-7 cells transfected with empty vector or ASM-myc was immunoprecipitated with anti-myc IgG and blotted with anti-sortilin antiserum. Sortilin was pulled down by ASM (lane 2) while no sortilin was pulled down in the control cells (lane 1). '"""" "0~~ N .()C'~ ~"\\'J& T-c'ery ~ , ~~~.-

.()C'~ '"""" ~~ N • -f.rl>) ~T • ~ ~C' DISCUSSION

I-cell disease (ICD) is an inherited lysosomal storage disorder that arises from a defect in the catalytic activity of N-acetylglucosaminyl-l-phosphotransferase. This enzyme is required for addition of mannose 6-phosphate (M6P) residues to most lysosomally targeted proteins in the cis-compartment ofthe Golgi apparatus. ICD was first described by Leroy and

DeMars (1967),(128) and subsequently classified as mucolipidosis type II (ML-II) by

Spranger and Wiedermann (1970),(50) because the clinical characteristics ofICD resembled those of mucopolysaccharidoses and sphingolipidoses.

ICD is characterized by a number of lysosomal enzymes present in excess ID extracellular fluid, due to the deficiency of the M6P sorting pathway.(51,54) Initially, the

M6P-Rc was thought to be the only lysosomal sorting receptor in mammalian cells.

However, severallaboratories reported that various lysosomal hydrolases, such as ASM, and sphingolipid activator proteins could still reach the lysosomesat lower rates in ICD.(3,129)

Based on these observations we hypothesized that ASM is transported to the lysosomes via an M6P-independent pathway in ICD and other cells. Our results on the intracellular localization of ASM in ICD cells validated our hypothesis. An ASM-myc fusion protein transiently expressed in ICD cells co-Iocalized with two lysosomal markers in granular structures, indicating that ASM reached the lysosomes without M6P residues. Since the M6P-Rc pathway is non-functional in ICD cells, the lysosomal targeting of ASM probably occurred via an alternative sorting pathway. Nonetheless, the labelling of ASM in the lysosomes of ICD cells was comparatively lower than control cells and occurred in a small population of lysosomes. This was not surprising given the fact that lysosomes are

44 heterogeneous structures that receive material from different compartments, including the plasma membrane (via endocytosis), Golgi apparatus (via membrane flow and protein targeting) and intracellular organelles (via autophagy).(130,131)

The decrease of ASM labeling in the lysosomal compartment of ICD cells suggests that the M6P-Rc is also implicated in the lysosomal sorting of ASM. In conclusion, our results demonstrated that the lysosomal targeting of ASM is partially dependent on the M6P

Rc pathway as well as an alternative targeting pathway.

Schissel et al (44) reported that in cultured mammalian cells ASM is found in two forms: a secretory form (S-ASM) and a lysosomal form (L-ASM). However, both sphingomyelinases are encoded by the same gene and arise from a single mRN A species.(132) It seems that the common precursor of ASM undergoes different posttranslational modifications in the Golgi apparatus giving rise to two products: the S-ASM released to the extracellular space via constitutive secretion, and the L-ASM transported from the Golgi apparatus to late endosomes and lysosomes via sorting receptors.( 44) In addition, the existence of a secreted form of sphingomyelinase raises the question as to whether or not

S-ASM is endocytosed via plasma membrane receptors and subsequently transported to endosomes and lysosomes. However, in this investigation we have examined the possibility that ASM is targeted to the lysosomes by an alternative sorting receptor directly from the trans-Golgi network (TGN). The rationale for this hypothesis was based on the consideration that accumulation of lipids plus the unbalanced distribution of cholesterol in fibroblasts in

ICD allow for only limited endocytosis of ASM. (133,134) A more important consideration was the fact that ASM contains an N-terminal saposin-like motifthat has been implicated in

45 a protein-protein interaction with the sorting receptor "sortilin".(13) Sortilin is a member of the novel VpslOp receptor family. Sortilin was initially found in the plasma membrane of neurons (135) and more recently in the Golgi apparatus of different cell types.(8,90,109) In the present investigation we have observed that a sortilin-EGFP fusion protein was found in the Golgi compartment and in peripheral vesicles reminiscent of endosomes. This intracellular distribution of sortilin substantiated its potential role as a sorting receptor that operates between the Golgi apparatus and endosomes.

Sequence analysis of sortilin revealed that its luminal domain and its cytoplasmic region contain structural features found in receptors invo lved in lysosomal targeting. In fact, the luminal domain of sortilin shares a substantial sequence homology to each oftwo luminal domains ofthe yeast vacuolar sorting protein VpslOp.(6) In yeast, VpslOp is involved in the sorting of the vacuolar hydrolase, carboxypeptidase Y.(7) The cytoplasmic tail of sortilin contains an acidic cluster-dileucine motif that binds the monomeric adaptor protein

GGA.(85) The cytoplasmic tail of sortilin is similar to the equivalent domain of the mannose

6-phosphate receptor. Interestingly, this region of sortilin was shown to substitute the function of the cytoplasmic tail of the cation-independent M6P-Rc (CI-M6P-Rc).(85) Based on these structural and functional features we hypothesized that sortilin was the receptor involved in the M6P-independent sorting of ASM.

To test this hypothesis we used a truncated sortilin lacking its cytoplasmic tail which was introduced into COS-7 cells as a dominant-negative competitor. Although dominant negative sortilin can bind lysosomal proteins it cannot exit the Golgi apparatus due to the lack of GGA binding motif. The truncated sortilin was linked to an EGFP tag to distinguish

46 from endogenous sortilin. Our experiments demonstrated that truncated sortilin was indeed retained in the Golgi apparatus.

To examine if sortilin was involved in the lysosomal targeting of AS M, we used truncated sortilin as a dominant-negative competitor in COS-7 cells. Our results showed that co-expression of ASM-myc and truncated sortilin-EGFP decreased the immunostaining of

ASM-myc in the peripheral granular structures as compared to control COS-7 cells co expressing ASM-myc and wild-type sortilin-EGFP. Since the M6P-Rc pathway functions normally in COS-7 cells, a certain amount of ASM-myc was observed in the lysosomes of the cells transfected with truncated sortilin. This result confirmed that sortilin was partially responsible for targeting ASM to the lysosomes. On the other hand, the immunostaining of

ASM in the lysosomes of COS-7 cells co-expressing ASM-myc and full-Iength sortilin

EGFP was comparable to that of control cells, suggesting that the over-expression of full length sortilin did not enhance the lysosomal targeting of ASM. A possible explanation for this result was that the monomeric adaptor GGA acted as a limiting factor in cells over expressing functional sortilin.

Although the above results demonstrated that both the M6P receptor and sortilin are implicated in the lysosomal transport of ASM, they could not exclude the possibility that the lysosomal ASM observed in ICD cells resulted from the intemalization of secreted ASM. To discard this possibility and to demonstrate whether or not sortilin acts in parallel to the mannose 6-phosphate receptor we investigated the intracellular trafficking of ASM in COS-7 cells transiently expressing truncated GGA. GGA is a clathrin-mediated ARF binding monomeric adaptor protein involved in the targeting of receptors from the TGN to late

47 endosomes.(117,136) GGA binds both sortilin and M6P-Rc via an N-terminal VHS domain and recruits clathrin to the membrane of the TGN via clathrin-binding sites within its HINGE domain. Therefore, GGA is a link between cargo receptors and clathrin. The truncated GGA used in our experiments lacks the HINGE and EAR do mains involved in the recruitment of clathrin.(123,137) The use of this dominant-negative competitor has heen shown to cause

Golgi retention of sortilin and M6P-Rc in COS-7 cells. Thus, we co-expressed truncated

GGA and ASM-myc in COS-7 cens. Under this condition ASM-myc was only detected in the Golgi apparatus. This indicates that ASM could not he targeted to lysosomes in the absence of functional sortilin and M6P sorting pathways. Since truncated GGA does not interfere with receptor-mediated endocytosis our result ruled out the possibility that internalized secreted ASM was the origin oflysosomal ASM in ICD cens.

To further confum that sortilin was involved in lysosomal sorting of ASM in ICD cells, these cens were co-transfected with ASM-myc and truncated sortilin-EGFP.

Immunostaining of the se fibroblasts showed a similar pattern to COS-7 cells co-expressing

ASM-myc and truncated GGA. Thus, the inhibition of lysosomal targeting of ASM in ICD cells corroborated that ASM uses sortilin as an alternative targeting receptor.

Soluble lysosomal enzymes sorted via the M6P receptor are generally released to the extracellular space in ICD cells. Therefore, the inhibition of sortilin should also result in the partial secretion of ASM. To address this question, we examined the fate of newly synthesized ASM using metabolic labelling and pulse-chase experiments in both COS-7 cells and ICD-cens. We disturhed the sortilin pathway by over-expressing truncated sortilin. Our results showed that truncated sortilin accelerated the secretion of ASM into the culture media

48 in ICD cells, corroborating that sortilin is required for the sorting of ASM in addition to the

M6P-Rc. Whether 100 % of the newly synthesized ASM was secreted or not rernains unknown. However, it is possible that a srnaU proportion of ASM was retained and/or degraded before leaking to the extracellular space.

As discussed elsewhere, the ASM gene encodes two forms of the same protein: lysosornal ASM (L-ASM) and secreted ASM (S-ASM). Generally, both isomers are synthesized simultaneously.(132) Surprisingly, the ASM level in the media of COS-7 ceUs was lowas compared to the ASM levels in the lysates the same cells. Moreover, the secretion of ASM in COS-7 ceUs was decreased as compared to cultured ICD cells. Although these results are difficult to interpret, it is plausible that the differences in the secretion rate of

ASM might be attributed to the use of different of celllines. In fact, it bas been reported that macrophages(132) and endothelial cells(138) found in atherosclerotic lesions secrete large amounts of ASM. These observations suggest that fibroblasts rnay exhibit a propensity to secrete more ASM than other cell types.

Using co-immunoprecipitation assays we presented strong evidence for a direct association between sortilin and ASM in COS-7 ceUs. The binding sites for the interaction of

ASM and sortilin are still unknown in both molecules. However, the luminal region of sortilin contains a Vps10p domain found in other members ofthis family ofreceptors. Within the C-terminal region of the Vps10p dornain exists a highly hydrophobic cysteine-rich stretch ofamino acids that corresponds to the most conserved sub-region of the VpslOp dornain.(96)

Based on tbis conservation it is likely that sortilin binds ASM and other ligands via this region. Similarly, the ASM region responsible for the binding sortilin is also unknown.

49 However, previous studies in our laboratory demonstrated that the soluble lysosomal protein

prosaposin requires the presence of saposin-like motifs to he targeted to the lysosomal

compartment.(12) Proteins containing saposin-like motifs constitute a large and diverse

group of proteins found in a variety of eukaryotic cells from plants and animaIs. Together

they form the "saposin-like protein (SAPLIP) family. AlI SAPLIP members share one or

more cysteine-rich saposin-like sequences. These conserved cysteines, form two or three

intra-domain disulfide bonds that create a common structural framework upon which other

conserved amino acids form several amphipathic a-helices(9) These motifs have been

implicated in the targeting of these proteins to their final destination, usually lysosomes.

(9,16)

In fact, the C-terminus of prosaposin contains a saposin-like domain that is

significantly similar to the N-terminus of surfactant protein B (SP-B). When this domain was

deleted, prosaposin could not he targeted to the lysosomal compartment (12) Interestingly,

SP-B also requires the presence of the N-terminus for its transient routing to multivesicular

and lamellar bodies.(II) Thus, we propose that the saposin-like motif found in ASM is most

likely the binding site for sortilin.

Although we have shown that both the M6P-Rc and sortilin mediate the lysosomal targeting of ASM, the M6P-Rc could not he co-immunoprecipitated by the anti-ASM antibody and vice versa. One possible explanation for our results could he attributed to conformational changes that occur after the binding hetween ASM and M6P-Rc, which result in the masking of the active epitopes.

50 Taken together, we have presented strong evidence that both, sortilin and the M6P

Rc, mediate the lysosomal targeting of ASM in COS-7. In ICD cells we demonstrated that

ASM reaches the lysosomal compartment via sortilin. In conclusion, both receptors are involved in the sorting ofASM to the lysosomes. Our results also demonstrated for the first time that sortilin is not only involved in the targeting of sphingolipid activator proteins, but also in the targeting of a soluble hydrolase. This finding suggests that sortilin may play a greater role as an alternative receptor to the M6P-Rc.

51 CONCLUSIONS

1) We have discovered a new mechanism of lysosomal targeting for the acid sphingomyelinase (ASM). This mechanism involves the receptor "sortilin" which acts as an alternative receptor to the M6P-Rê.

2) In the absence of the M6P pathway in ICD fibroblasts, ASM is transported to the lysosomes via sortilin.

3) Co-immunoprecipitation assays demonstrated that sortilin specifically binds acid sphingomyelinasè.

4) The involvement of sortilin in the trafficking of ASM to the lysosomes was corroborated via the perturbation of sortilin and monomeric adaptor GGAs with dominant negative constructs.

5) The present study indicates that sortilin may play a general role in the transport of lysosomal proteins containing saposin-like motifs.

52 ABBREVIATION

A alanine

AOP adenosine diphosphate

AOAH acyloxyacyl hydrolase

AP-3 adaptor protein complex 3

ARF ADP ribosylation factor

ASM acid sphingomyelinase

C cysteine

CCV clathrin coated vesicle

CO spectroscopy circular dichroism spectroscopy

CO-M6P-RC Cation-dependent mannose 6-phosphate receptor cONA complementary DNA

CI-M6P-Rc Cation-independent mannose 6-phosphate receptor

CPS carboxypeptidase S

Cpy carboxypeptidase Y o asparatic acid

0609 tricyclodecan-9-yl xanthate potassium

E glutamic acid

EGFP enhanced green fluorescence protein

ER endoplasmic reticulum

F phenyalanine

FBi fumonisin B 1

FITC fluorescein isothiocyanate

G glycine

GGA Golgi localizing gamma-adaptin ear containing ARF binding protein

53 GleNAc N-acetyl-D-glucosamine

Gm2AP ganglioside M2 activator protein

GTP guanosine triphosphate

GNPTA glucosamine (UDP-N-acetyl}-Iysosomal-enzyme N-acetylglucosamine phosphotransferase

H histidine

HA head activator

isoleucine

ICD I-cell disease

IGF Il insulin-like growth factor Il

K lysine

Kda kilodalton

L leucine

Lamp Iysosomal associated membrane protein

LDLR low-density lipoprotein receptor

LPL lipoprotein lipase

LPS lipopolysaccharide

LRP3 low-density lipoprotein receptor related protein 3

L-SMase Iysosomal sphingomyelinase

M methionine

M6P mannose 6-phosphate

M6P-RC mannose 6-phosphate receptor

ML-II mucolipidosis Il

N asparagine

NK-Iysin Natural-killer cell Iysin

NMR nuclear magnetic resonance

NPD Niemman-Pick disease

54 NT neurotensin

NTR1 neurotensin receptor 1

NTR3 neurotensin receptor 3

P proline

PCR polymerase chain reaction proNGF proprotein of nerve growth factor

PSI plant specifie insert

Q glutamine

R arginine

RAP receptor associated protein

RER rough endoplasmic reticulum

S serine

SAPLIP saposin-like protein

Sorcs sortilin-related Vps10 domain containing receptor

SorLA sorting protein-related receptor containing LDLR class A repeats

SPB surfactant prote in B

SPM sphingomyeline

S-SMase secretory sphingomyelinase

T threonine

TR Texas red

TGN trans-Golgi network

TRITC tetramethylrhodamine isothiocyanate

UDP uridine diphophate

V valine

Vps10p vacuolar protein sorting 10 protein

W tryptophan

55 WT wild type

Y tyrosine

56 REFERENCE

1. Lodish, H. R, Arnold; Zipursky, S. Lawrence; Matsudaira, Paul; Baltimore, David; Darnell, James E. (1999) Molecular cel! biology, 4th Ed., W.H.Freeman&Co, NewYork 2. Gabel, C. A, Goldberg, D. E., and Komfeld, S. (1983) Proc Natl Acad Sei USA 80(3), 775-779 3. Glickman, J. N., and Komfeld, S. (1993) J Cel! Biol 123(1), 99-108 4. Rijnboutt, S., Aerts, H. M., Geuze, H. J., Tager, J. M., and Strous, G. J. (1991) J Biol Chem 266(8), 4862-4868 5. Glombitza, G. J., Becker, E., Kaiser, H. W., and Sandhoff, K. (1997) J Biol Chem 272(8), 5199-5207 6. Petersen, C. M., Nielsen, M. S., Nykjaer, A., Jacobsen, L., Tommerup, N., Rasmussen, H. H., Roigaard, H., Gliemann, J., Madsen, P., and Moestrup, S. K. (1997) J Biol Chem 272(6),3599-3605 7. Marcusson, E. G., Horazdovsky, RF., Cereghino, J. L., Gharakhanian, E., and Emr, S. D. (1994) Cel! 77(4), 579-586 8. Takatsu, H., Katoh, Y., Shiba, Y., and Nakayama, K. (2001) J Biol Chem 276(30), 28541-28545 9. Kervinen, J., Tobin, G. J., Costa, J., Waugh, D. S., Wlodawer, A., and Zdanov, A (1999) Emba J18(14), 3947-3955 10. Staab, J. F., Ginkel, D. L., Rosenberg, G. R, and Munford, R S. (1994) J Biol Chem 269(38),23736-23742 11. Lin, S., PhiIlips, K. S., Wilder, M. R, and Weaver, T. E. (1996) Biochim Biophys Acta 1312(3), 177-185 12. Zhao, Q., and Morales, C. R (2000) J Biol Chem 275(32), 24829-24839 13. He, X., Okino, N., Dhami, R, Dagan, A, Gatt, S., Schulze, H., Sandhoff, K., and Schuchman, E. H. (2003) J Biol Chem 278(35),32978-32986

57 14. Hurwitz, R., Ferlinz, K., Vielhaber, G., Moczall, H., and Sandhoff, K. (1994) J Biol Chem 269(7), 5440-5445 15. Dell'Angelica, E. C., and Payne, G. S. (2001) Celll06(4), 395-398 16. Munford, R. S., Sheppard, P. O., and O'Hara, P. J. (1995) J Lipid Res 36(8), 1653- 1663 17. Hiraiwa, M., Soeda, S., Kishimoto, Y., and O'Brien, J. S. (1992) Proc Natl Acad Sei USA 89(23),11254-11258 18. Batenburg, J. J. (1992) Am J Physiol262(4 Pt 1), L367-385 19. Azuma, N., O'Brien, J. S., Moser, H. W., and Kishimoto, Y. (1994) Arch Biochem Biophys 311(2), 354-357 20. Grabowski, G. A., and Horowitz, M. (1997) Baillieres Clin Haematoll0(4), 635-656 21. Jatzkewitz, H., and Stinshoff, K. (1973) FEBS Let! 32(1), 129-131 22. Munford, R. S., and Hunter, J. P. (1992) J Biol Chem 267(14), 10116-10121 23. Nogee, L. M., de Mello, D. E., Dehner, L. P., and Colten, H. R. (1993) N Engl J Med 328(6), 406-410 24. Ricci, V., Stroppiano, M., Corsolini, F., Di Rocco, M., Parenti, G., Regis, S., Grossi, S., Biancheri, R., Mazzotti, R., and Filocamo, M. (2004) Hum Mutat 24(1), 105 25. Pittis, M. G., Ricci, V., Guerci, V.!., Marcais, C., Ciana, G., Dardis, A., Gerin, F., Stroppiano, M., Vanier, M. T., Filocamo, M., and Bembi, B. (2004) Hum Mutat 24(2), 186-187 26. Andersson, M., Gunne, H., Agerberth, B., Boman, A., Bergman, T., Sillard, R., Jornvall, H., Mutt, V., Olsson, B., Wigzell, H., and et al. (1995) Embo J 14(8), 1615- 1625 27. Tormakangas, K., Runeberg-Roos, P., Ostman, A., Tilgmann, C., Sarkkinen, P., Kervinen, J., Mikola, L., and Kalkkinen, N. (1991)Adv Exp Med Biol 306, 355-359 28. Runeberg-Roos, P., Kervinen, J., Kovaleva, V., Raikhel, N. V., and Gal, S. (1994) Plant Physioll05(1), 321-329 29. Schutte, C. G., Lemm, T., Glombitza, G. J., and Sandhoff, K. (1998) Prote in Sci 7(4), 1039-1045

58 30. Liepinsh, E., Andersson, M., Ruysschaert, J. M., and Otting, G. (1997) Nat Struct Biol 4(10), 793-795 31. Ahn, V. E., Faull, K. F., Whitelegge, J. P., Fluharty, A. L., and Prive, G. G. (2003) Proc Natl Acad Sei USA 100(1),38-43 32. Brockman, J. M., Wang, Z., Notter, RH., and Dluhy, R A. (2003) Biophys J84(1), 326-340 33. Vaccaro, A. M., Tatti, M., Ciaffoni, F., Salvioli, R, Maras, B., and Barca, A. (1993) FEBS Lett 336(1), 159-162 34. Lefrancois, S., May, T., Knight, C., Bourbeau, D., and Morales, C. R (2002) J Biol Chem 277(19),17188-17199 35. Tatti, M., Salvioli, R, Ciaffoni, F., Pucci, P., Andolfo, A., Amoresano, A., and Vaccaro, A. M. (1999) Eur J Biochem 263(2), 486-494 36. Schnabel, D., Schroder, M., and Sandhoff, K. (1991) FEBS Lett 284(1),57-59 37. Rafi, M. A., de Gala, G., Zhang, X. L., and Wenger, D. A. (1993) Somat Cell Mol Genet 19(1), 1-7 38. Vaccaro, A. M., Salvioli, R, Tatti, M., and Ciaffoni, F. (1999) Neurochem Res 24(2), 307-314 39. Misasi, R, Sorice, M., Garofalo, T., Griggi, T., Campana, W. M., Giammatteo, M., Pavan, A., Hiraiwa, M., Pontieri, G. M., and O'Brien, 1. S. (1998) J Neurochem 71(6), 2313-2321 40. Hagen, F. S., Grant, F. J., Kuijper, J. L., Slaughter, C. A., Moomaw, C. R, Orth, K., O'Hara, P. J., and Munford, R S. (1991) Biochemistry 30(34),8415-8423 41. Tormakangas, K., Hadlington, J. L., Pimpl, P., Hillmer, S., Brandizzi, F., Teeri, T. H., and Denecke, J. (2001) Plant Cell13(9), 2021-2032 42. Lampert, F., and Teller, W. (1967) Monatsschr Kinderheilkd 115(8),439-443 43. Miranda, S. R, Erlich, S., Friedrich, V. L., Jr., Haskins, M. E., Gatt, S., and Schuchman, E. H. (1998) Transplantation 65(7), 884-892 44. Schissel, S. L., Keesler, G. A., Schuchman, E. H., Williams, K. J., and Tabas, 1. (1998) J Biol Chem 273(29), 18250-18259

59 45. Newrzella, D., and Stoffel, W. (1996) J Biol Chem 271(50),32089-32095 46. Reitman, M. L., and Komfeld, S. (1981) J Biol Chem 256(9), 4275-4281 47. Reitman, M. L., and Komfeld, S. (1981) J Biol Chem 256(23), 11977-11980 48. Waheed, A., Hasilik, A., and von Figura, K. (1982) J Biol Chem 257(20), 12322- 12331 49. Pelham, H. R (1988) Embo J7(4), 913-918 50. Spranger, J. W., and Wiedemann, H. R (1970) Humangenetik 9(2), 113-139 51. Leroy, J. G., and Spranger, J. W. (1970) N Engl J Med283(11), 598-599 52. Okada, S., Inui, K., Furukawa, M., Midorikawa, M., Nishimoto, J., Yabuuchi, H., Kato, T., Watanabe, M., Gasa, S., and Makita, A. (1987) Enzyme 38(1-4),267-272 53. Ullrich, K., and von Figura, K. (1979) Biochem JI82(1), 245-247 54. Vladutiu, G. D., and Rattazzi, M. C. (1979) J Clin Invest 63(4),595-601 55. Little, L., Alcouloumre, M., Drotar, A. M., Herman, S., Robertson, R, Yeh, R Y., and Miller, A. L. (1987) Biochem J248(1), 151-159 56. Little, L. E., Mueller, O. T., Honey, N. K., Shows, T. B., and Miller, A. L. (1986)J Biol Chem 261(2), 733-738 57. Cantor, A. B., Baranski, T. J., and Komfeld, S. (1992) J Biol Chem 267(32), 23349- 23356 58. Baranski, T. J., Cantor, A. B., and Komfeld, S. (1992) J Biol Chem 267(32), 23342- 23348 59. Baranski, T. J., Koelsch, G., Hartsuck, J. A., and Kornfe1d, S. (1991) J Biol Chem 266(34),23365-23372 60. Baranski, T. J., Faust, P. L., and Kornfeld, S. (1990) Cel! 63(2), 281-291 61. Brown, W. J., Goodhouse, J., and Farquhar, M. G. (1986) J Cel! Biol 103(4), 1235- 1247 62. Willingham, M. C., Pastan, 1. H., Sahagian, G. G., Jourdian, G. W., and Neufeld, E. F. (1981) Proc Natl Acad Sei USA 78(11), 6967-6971 63. Dahms, N. M., Lobel, P., Breitmeyer, J., Chirgwin, J. M., and Komfe1d, S. (1987) Cel! 50(2), 181-192

60 64. Morgan, D. O., Edman, J. C., Standring, D. N., Fried, V. A., Smith, M. C., Roth, R A, and Rutter, W. J. (1987) Nature 329(6137),301-307 65. Waheed, A., Braulke, T., Junghans, U., and von Figura, K. (1988) Biochem Biophys Res Commun 152(3), 1248-1254 66. MacDonald, R G., Pfeffer, S. R, Coussens, L., Tepper, M. A., Brocklebank, C. M., Mole, J. E., Anderson, J. K., Chen, E., Czech, M. P., and Ullrich, A (1988) Science 239(4844), 1134-113 7 67. Oka, Y., Rozek, L. M., and Czech, M. P. (1985) J Biol Chem 260(16),9435-9442 68. Klier, H. J., von Figura, K., and Pohlmann, R (1991) Eur J Biochem 197(1), 23-28 69. Stein, M., Meyer, H. E., Hasilik, A., and von Figura, K. (1987) Biol Chem Hoppe Seyler 368(8),927-936 70. Roberts, D. L., Weix, D. J., Dahms, N. M., and Kim, J. J. (1998) Cel! 93(4), 639-648 71. Tong, P. Y., Gregory, W., and Kornfeld, S. (1989) J Biol Chem 264(14), 7962-7969 72. Hancock, M. K., Haskins, D. J., Sun, G., and Dahms, N. M. (2002) J Biol Chem

277(13), 11255~11264 73. Tong, P. Y., and Komfeld, S. (1989) J Biol Chem 264(14), 7970-7975 74. Lobel, P., Fujimoto, K., Ye, RD., Griffiths, G., and Kornfeld, S. (1989) Cel! 57(5), 787-796 75. Kyle, J. W., NoIan, C. M., Oshima, A., and SIy, W. S. (1988) J Biol Chem 263(31),

16230~16235 76. Stein, M., Zijderhand-Bleekemolen, J. E., Geuze, H., Hasilik, A., and von Figura, K. (1987) Embo J 6(9), 2677-2681 77. Stein, M., Braulke, T., Krentler, C., Hasilik, A., and von Figura, K. (1987) Biol Chem Hoppe Seyler 368(8), 937-947 78. Banta, L. M., Robinson, J. S., Klionsky, D. J., and Emr, S. D. (1988) J Cel! Biol 107(4), 1369-1383 79. Johnson, L. M., Bankaitis, V. A, and Emr, S. D. (1987) Cel! 48(5), 875-885 80. Stevens, T. H., Rothman, J. H., Payne, G. S., and Schekman, R (1986) J Cel! Biol 102(5), 1551-1557

61 81. Rothman, J. H., Hunter, C. P., Valls, L. A, and Stevens, T. H. (1986) Proc Natl Acad Sei USA 83(10), 3248-3252 82. Jorgensen, M. U., Emr, S. D., and Winther, J. R. (1999) Eur J Biochem 260(2), 461- 469 83. Winther,1. R., Stevens, T. H., and Kielland-Brandt, M. C. (1991) Eur J Biochem 197(3),681-689 84. Valls, L. A, Winther, J. R., and Stevens, T. H. (1990) JCell Biol 111(2), 361-368 85. Nielsen, M. S., Madsen, P., Christensen, E. 1., Nykjaer, A., Gliemann, J., Kasper, D., Pohlmann, R., and Petersen, C. M. (2001) Embo J20(9), 2180-2190 86. Tsuji, A., Omura, K., and Suzuki, Y. (1988) Clin Chim Acta 176(1), 115-121 87. Lightbody, J., Wiesmann, D., Hadom, B., and Herschkowitz, N. (1971) Lancet 1(7696), 451 88. Rijnboutt, S., Aerts, H. M., Geuze, H. J., Tager, J. M., and Strous, G. J. (1991) J Biol Chem 266(8),4862-4868. 89. Lefrancois, S., Zeng, J., Hassan, A. J., Canuel, M., and Morales, C. R. (2003) Embo J 22(24),6430-6437 90. Hassan, A. J., Zeng, J., Ni, X., and Morales, C. R. (2004) Mol Reprod Dev 68(4), 476-483 91. Hermey, G., Riedel, 1. B., Hampe, W., Schaller, H. C., and Hermans-Borgmeyer, I. (1999) Biochem Biophys Res Commun 266(2),347-351 92. Jacobsen, L., Madsen, P., Jacobsen, C., Nielsen, M. S., Gliemann, J., and Petersen, C. M. (2001) J Biol Chem 276(25),22788-22796 93. Jacobsen, L., Madsen, P., Moestrup, S. K., Lund, A H., Tommerup, N., Nykjaer, A., Sottrup-Jensen, L., Gliemann, J., and Petersen, C. M. (1996) J Biol Chem 271(49), 31379-31383 94. Hampe, W., Rezgaoui, M., Hermans-Borgmeyer, 1., and Schaller, H. C. (2001) Hum Genet 108(6), 529-536 95. Munck Petersen, C., Nielsen, M. S., Jacobsen, C., Tauris, J., Jacobsen, L., Gliemann, J., Moestrup, S. K., and Madsen, P. (1999) Emba JI8(3), 595-604

62 96. Westergaard, U. B., Sorensen, E. S., Hermey, G., Nielsen, M. S., Nykjaer, A, Kirkegaard, K., Jacobsen, C., Gliemann, J., Madsen, P., and Petersen, C. M. (2004) J Biol Chem 279(48),50221-50229 97. Gliemann, J., Hermey, G., Nykjaer, A, Petersen, C. M., Jacobsen, C., and Andreasen, P. A (2004) Biochem J381(Pt 1),203-212 98. Lintzel, J., Franke, 1., Riedel, 1. B., Schaller, H. C., and Hampe, W. (2002) Biol Chem 383(11), 1727-1733 99. Wolozin, B. (2004) Arch NeuroI61(8), 1178-1180

100. Zhu, Y., Bujo, H., Yamazaki, H., Ohwak~ K., Jiang, M., Hirayama, S., Kanak~ T.,

Shibasaki, M., Takahash~ K., Schneider, W. J., and Saito, Y. (2004) Circ Res 94(6), 752-758 101. Zsurger, N., Mazella, J., and Vincent, J. P. (1994) Brain Res 639(2), 245-252 102. Morris, N. J., Ross, S. A, Lane, W. S., Moestrup, S. K., Petersen, C. M., Keller, S. R, and Lienhard, G. E. (1998) J Biol Chem 273(6),3582-3587 103. Lin, B. Z., Pilch, P. F., and Kandror, K. V. (1997) J Biol Chem 272(39),24145-24147 104. Nielsen, M. S., Jacobsen, C., Olivecrona, G., Gliemann, J., and Petersen, C. M. (1999) J Biol Chem 274(13),8832-8836 105. Hermey, G., Plath, N., Hubner, C. A, Kuhl, D., Schaller, H. C., and Hermans Borgmeyer,1. (2004) J Neurochem 88(6), 1470-1476 106. Jacobsen, L., Madsen, P., Nielsen, M. S., Geraerts, W. P., Gliemann, J., Smit, A B., and Petersen, C. M. (2002) FEBS Lett 511(1-3), 155-158 107. Griffiths, G., Matteoni, R, Back, R, and Hoflack, B. (1990) J CeU Sei 95 (Pt 3),441- 461 108. Dicou, E., Vincent, J. P., and Mazella, J. (2004) J Neurosci Res 78(1),92-99 109. Morinville, A, Martin, S., Lavallee, M., Vincent, J. P., Beaudet, A, and Mazella, J. (2004) Int J Biochem CeU Biol 36(11), 2153-2168 110. Navarro, V., Martin, S., Sarret, P., Nielsen, M. S., Petersen, C. M., Vincent, J., and Mazella, J. (2001) FEBS Lett 495(1-2), 100-105

63 111. Mart~ S., Navarro, V., Vincent, 1 P., and Mazella, 1 (2002) Gastroenterology 123(4), 1135-1143 112. Nykjaer, A, Lee, R, Teng, K. K., Jansen, P., Madsen, P., Nielsen, M. S., Jacobsen, C., Kliemannel, M., Schwarz, E., Willnow, T. E., Hempstead, B. L., and Petersen, C. M. (2004) Nature 427(6977), 843-848 113. Conticello, S. G., Kowalsman, N. D., Jacobsen, C., Yudkovsky, G., Sato, K., Elazar, Z., Petersen, C. M., Aronheim, A, and Fainzilber, M. (2003) J Biol Chem 278(29), 26311-26314 114. Ghosh, P., and Kornfeld, S. (2004) Eur J Cel! Biol 83(6), 257-262 115. Poussu, A, Lohi, O., and Lehto, V. P. (2000)J Biol Chem 275(10),7176-7183 116. Hirst, J., Lui, W. W., Bright, N. A, Totty, N., Seaman, M. N., and Robinson, M. S. (2000) J Cel! Biol 149(1), 67-80 117. Dell'Angelica, E. C., Puertollano, R, Mullins, C., Aguilar, R C., Vargas, J. D.,

Hartnel~ L. M., and Bonifacino, J. S. (2000) J Cel! Biol 149(1), 81-94 118. Boman, A L., Zhang, C., Zhu, X., and Kahn, R A. (2000) Mol Biol Cel! 11(4), 1241- 1255 119. Boman, A. L. (2001) J Cel! Sci 114(Pt 19),3413-3418 120. Misra, S., Beach, B. M., and Hurley, J. H. (2000) Biochemistry 39(37), 11282-11290 121. Mullins, C., and Bonifacino, J. S. (2001) Mol Cel! Biol 21(23), 7981-7994 122. Black, M. W., and Pelham, H. R (2000) J Cel! Biol 151(3), 587-600 123. Puertollano, R., Randazzo, P. A., Presley, J. F., Hartnell, L. M., and Bonifacino, 1 S. (2001) Cel! 105(1), 93-102 124. Puertollano, R, Aguilar, R C., Gorshkova, 1., Crouch, R J., and Bonifacino, 1 S. (2001) Science 292(5522), 1712-1716 125. Collins, B. M., Watson, P. l, and Owen, D. 1 (2003) Dev Cel! 4(3), 321-332 126. Hirsch, D. S., Stanley, K. T., Chen, L. x., Jacques, K. M., Puertollano, R, and Randazzo, P. A (2003) Traffic 4(1), 26-35 127. Page, L. J., Sowerby, P. J., Lui, W. W., and Robinson, M. S. (1999) JCel! Biol 146(5),993-1004

64 128. Leroy, J. G. a. D., RI. (1967) Science 157,804-806 129. Jobb, E. A, and Callahan, J. W. (1989) Biochem Cell Biol 67(11-12), 801-807 130. Tang, X. M., Clermont, Y., and Hermo, 1. (1988) J Androl9(4),284-293 131. C.RMorales, Q. Z. a. S. 1. (1999) BioceIl23(3), 149-160 132. Schissel, S. 1., Schuchman, E. H., Williams, K. J., and Tabas, 1. (1996) J Biol Chem 271(31), 18431-18436 133. Dhami, R, and Schuchman, E. H. (2004)J Biol Chem 279(2),1526-1532 134. Grassme, H., Jendrossek, V., Bock, J., Riehle, A, and Gulbins, E. (2002) J /mmunol 168(1),298-307 135. Chabry, J., Gaudriault, G., Vincent, J. P., and Mazella, J. (1993) J Biol Chem 268(23), 17138-17144 136. Takatsu, H., Yoshino, K., and Nakayama, K. (2000) Biochem Biophys Res Commun 271(3), 719-725 137. Bilodeau, P. S., Winistorfer, S. C., Allaman, M. M., Surendhran, K., Kearney, W. R, Robertson, A D., and Piper, R C. (2004) J Biol Chem 279(52),54808-54816 138. Marathe, S., Schissel, S. 1., Yellin, M. J., Beatini, N., Mintzer, R, Williams, K. J., and Tabas, 1. (1998) J Biol Chem 273(7), 4081-4088