EUKARYOTIC CELL, Nov. 2005, p. 1851–1862 Vol. 4, No. 111535-9778/05/$08.00�0 doi:10.1128/EC.4.11.1851–1862.2005Copyright © 2005, American Society for Microbiology. All Rights Reserved.

Saccharomyces cerevisiae Npc2p Is a Functionally ConservedHomologue of the Human Niemann-Pick Disease

Type C 2 Protein, hNPC2†Adam C. Berger,1,2,3 Thomas H. Vanderford,4,5 Kim M. Gernert,6 J. Wylie Nichols,7

Victor Faundez,2 and Anita H. Corbett1*Department of Biochemistry, Emory University School of Medicine, Atlanta, Georgia 303221; Department of Medicine,

Emory University School of Medicine, Atlanta, Georgia 303224; BIMCORE (Biomolecular Computing Resource),Emory University, Atlanta, Georgia 303226; Department of Physiology, Emory University School of Medicine,

Atlanta, Georgia 303227; Department of Cell Biology, Emory University School of Medicine,Atlanta, GA 303222; and Graduate Program in Biochemistry, Cell, and Developmental

Biology3 and Graduate Program in Population Biology, Ecology, and Evolution,5

Emory University, Atlanta, Georgia 30322

Received 14 March 2005/Accepted 24 August 2005

Niemann-Pick Disease Type C (NP-C) is a fatal neurodegenerative disease, which is biochemically distin-guished by the lysosomal accumulation of exogenously derived cholesterol. Mutation of either the hNPC1 orhNPC2 gene is causative for NP-C. We report the identification of the yeast homologue of human NPC2,Saccharomyces cerevisiae Npc2p. We demonstrate that scNpc2p is evolutionarily related to the mammalianNPC2 family of proteins. We also show, through colocalization, subcellular fractionation, and secretionanalyses, that yeast Npc2p is treated similarly to human NPC2 when expressed in mammalian cells. Impor-tantly, we show that yeast Npc2p can efficiently revert the unesterified cholesterol and GM1 accumulation seenin hNPC2�/� patient fibroblasts demonstrating that it is a functional homologue of human NPC2. The presentstudy reveals that the fundamental process of NPC2-mediated lipid transport has been maintained throughoutevolution.

Niemann-Pick Disease Type C (NP-C) is a fatal neurovis-ceral disorder that is characterized by the accumulation oflow-density-lipoprotein (LDL)-derived cholesterol in endoly-sosomal compartments (37). NP-C is caused by mutation ofeither of two genes: hNPC1 (5, 30) or hNPC2 (34). Mutationsin hNPC1 account for 95% of all patients (6, 37). Recently, itwas demonstrated that the yeast homologue of hNPC1,scNcr1p, can functionally complement the loss of hNPC1 frommutant CHO cells (31). This finding suggests that there is anevolutionary conservation of function between species.

Human NPC2 (hNPC2), which is mutated in approximately5% of NP-C disease (37), encodes a conserved 151-amino-acidsecreted glycoprotein with an endoplasmic reticulum signalsequence (24). The NPC2 protein contains an MD-2-relatedlipid-recognition (ML) domain (21), which is predicted to me-diate direct binding to lipids (21). This prediction is based inpart on the finding that the MD-2 protein, the founding mem-ber of the ML domain family of proteins, binds directly tolipopolysaccharide (49). Structural analysis of an ML domain-containing protein, the dust mite allergen Der P 2, reveals thatthis protein folds into two �-strands with an internal cavity thatis occupied by a hydrophobic ligand predicted to be a lipid(10). As seen for the other ML domain-containing proteins

tested thus far (10, 49), the human, mouse, and porcine NPC2orthologues all bind directly to lipid ligands, specifically cho-lesterol or cholesterol analogs (13, 26, 35). The recently solvedcrystal structure of bovine NPC2 (bNPC2) is consistent with amodel wherein cholesterol binds within a loosely packed hy-drophobic protein core (13). Taken together, these bindingand structural studies suggest that NPC2 plays a direct role insterol transport.

Here we report the identification of the yeast homologue ofhNPC2 that we have termed Npc2p (scNpc2p). Our analysis ofSaccharomyces cerevisiae Npc2p through phylogenetic and ho-mology modeling studies suggests that this protein has beenconserved from yeast to mammals. Most importantly, we showthat yeast Npc2p can functionally replace human NPC2 inhNPC2�/� patient fibroblasts by reestablishing the transport ofcholesterol and the ganglioside GM1. These findings demon-strate that yeast Npc2p is a functional homologue of hNPC2and reveal that the yeast protein, like hNPC2 (34), can facili-tate intracellular lipid trafficking.

MATERIALS AND METHODS

Strains, plasmids, and cell culture. All DNA manipulations were performedaccording to standard protocols (42), and all yeast media were prepared bystandard methods (1). All yeast strains, cell lines, plasmids, and viruses aredescribed in Table 1. Chemicals were obtained from Fisher Scientific (Pittsburgh,PA), Sigma Chemical Co. (St. Louis, MO), Stratagene (La Jolla, CA), or USBiological (Swampscott, MA) unless otherwise noted. A complete deletion ofscNPC2 was created as previously described by using a PCR-based strategy (27).Briefly, the diploid wild-type strain ACY247 was used as a starting strain tocreate a heterozygous diploid which was sporulated and subjected to tetraddissection. Spores were backcrossed twice to the wild-type haploid strain

* Corresponding author. Mailing address: Department of Biochemis-try, Emory University School of Medicine, 1510 Clifton Rd., NE, Atlanta,GA 30322. Phone: (404) 727-4546. Fax: (404) 727-2738. E-mail: acorbe2@emory.edu.

† Supplemental material for this article may be found at http://ec.asm.org/.

1851

on February 9, 2021 by guest

http://ec.asm.org/

Dow

nloaded from

ACY193 generating two separate isogenic isolates, ACY783 and ACY1260.Gene-specific deletion was confirmed by PCR analysis with one oligonucleotidewithin the HIS3 locus and one oligonucleotide in the flanking genomic region.

HEK293 cells were grown in Dulbecco modified Eagle medium (DMEM; 4.5 gof glucose/liter; Cellgro, Herndon, VA) supplemented with 10% fetal bovineserum (FBS) (HyClone, Logan, UT), 100 U of penicillin/ml, and 100 �g ofstreptomycin/ml (Cellgro). Mouse fibroblasts were grown in DMEM with 10%FBS, 100 U of penicillin/ml, 100 �g of streptomycin/ml, and 200 �g of hygro-mycin B/ml. Human NPC2�/� and normal skin fibroblasts (14) were kindlyprovided by Daniel S. Ory (Washington University, St. Louis, MO) and weregrown in DMEM supplemented with 10% FBS, 2 mM glutamine (Cellgro),100 U of penicillin/ml, and 100 �g of streptomycin/ml.

Antibodies. Primary antibodies used in the present study included anti-GM130used at 1:250 and anti-AP-1 � used at 1:500 (BD Transduction Laboratories),anti-mouse LAMPI or LAMPII used at 1:500 and anti-human LAMPI used at1:100 (Developmental Studies Hybridoma Bank, University of Iowa), anti-cathe-psin D used at 1:500 (Upstate, Waltham, MA), anti-green fluorescent protein(GFP) used at 1:5,000 (Synaptic Systems, Gottingen, Germany), and anti-tubulinused at 1:1,000 (kindly provided by Harish C. Joshi, Emory University, Atlanta,GA). Secondary antibodies used were: Alexa-conjugated goat anti-mouse andgoat anti-rabbit 568 used at a 1:1,000 dilution (Molecular Probes/InvitrogenDetection Technologies, Carlsbad, CA) and goat anti-rabbit horseradish perox-idase (HRP) and goat anti-mouse HRP used at 1:7,000 (Zymed Laboratories/Invitrogen Immunodetection, Carlsbad, CA).

Adenovirus production. Adenoviruses encoding yeast Npc2p-GFP and humanNPC2-GFP were created through standard molecular biological methods.Briefly, yeast NPC2 was amplified by PCR from a yeast genomic NPC2 plasmid(pAC1181) and human NPC2 was amplified by PCR from human peripheralblood mononuclear cell cDNA (kind gift from Silvija I. Staprans, Emory Uni-versity, Atlanta, GA). Both genes were first cloned into the pEGFP vector(Clontech, Palo Alto, CA) to create in-frame C-terminal GFP fusions. PCR wasused to amplify both genes with GFP followed by cloning into the AdEasyAdenoviral Vector system (Stratagene). In short, both genes were cloned NotI-HindIII into the pShuttle-CMV vector followed by homologous recombinationinto the pAdEasy-1 adenoviral backbone. The resulting recombinants weretransfected into HEK293 cells by using Lipofectamine 2000 (Invitrogen, Carls-bad, CA) for plaque formation. Viruses were amplified in HEK293 cells andpurified by using the BD Adeno-X virus purification kit (Clontech) as directed.

Phylogenetic analysis of ML-containing proteins. The putative yeast NPC2homologue protein sequence was aligned with other ML containing proteins (seeTable S1 in the supplemental material) by using CLUSTAL X (46) and by hand.Amino acid sequences were used to minimize the effect of saturation on the data.

The gap-stripped alignment was 98 amino acids long. Tree-puzzle (version 5.0;National Supercomputer Centre in Sweden, Linkoping University [http://www.nsc.liu.se/software/biology/puzzle5/]) (43) was used to calculate a maximum-likelihood distance matrix, estimating from the data both the amino acidfrequencies and the alpha parameter describing the gamma distribution of site-by-site rate variation. A phylogenetic tree was obtained from the maximum-likelihood distance matrix by using the neighbor-joining method in Neighborfrom the Phylip package (version 3.5c; J. Felsenstein, Department of GenomeSciences, University of Washington [http://evolution.genetics.washington.edu/phylip.html]). A total of 999 bootstrap replicates of the alignment were per-formed by using Seqboot, Neighbor, and Consense from Phylip, and Puzzleboot(version 1.03; Department of Biochemistry and Molecular Biology, DalhousieUniversity [www.tree-puzzle.de/#puzzleboot]). Those nodes with greater than50% bootstrap support are labeled in Fig. 1B. No significant difference in diver-gence was found between the dust-mites (clade 1) and the animalia NPC2proteins (clade 2). The relative nearness of these two clades to yeast Npc2p andtheir clustering within a larger bootstrap supported clade that approximatelyrecapitulates the animalia species tree suggests that this whole clade representsthe evolutionary progression of NPC2 proteins from fungi to humans.

Homology model construction. Molecular modeling of yeast Npc2p was per-formed by using the program Modeler (version 7v7; Departments of Biophar-maceutical Sciences and Pharmaceutical Chemistry and California Institute forQuantitative Biomedical Research, University of California San Francisco[http://salilab.org/modeler/]) (12, 32, 41) for model construction and SYBYL(version 7.0; Tripos, Inc. [http://www.tripos.com]) for analysis and refinement.The model was based upon both the bNPC2 structure (PDB entry 1NEP; 13) andthe dust mite allergen Der P 2 structure (PDB entry 1KTJ; 10). Sequencealignments were obtained from the PFAM database (www.sanger.ac.uk/Software/Pfam/) and manually adjusted. For analysis purposes, protein alignments weregenerated only for the mature proteins. Structural constraints in the model werelimited to the conserved backbone regions and cysteine bridges. Refinement ofthe model was achieved by using the SYBYL Analyze Protein subroutine toevaluate model stereochemistry, followed by manual structure minimization. Allimages were created by using the Swiss-PdbViewer DeepView (version 3.7; SwissInstitute of Bioinformatics [http://www.expasy.org/spdbv/]) (16). Cavity analysiswas performed by using PASS (version 2.0.36; Computational Chemistry List[http://www.ccl.net/cca/software/UNIX/pass/overview.shtml]) (4) to generate buriedprobes whose volume was assessed by using SYBYL.

Direct fluorescence microscopy. scNpc2p was localized in living yeast cells asa C-terminal GFP fusion protein. To visualize the GFP fusion protein, wild-typeyeast cells (ACY402) were transformed with the scNPC2-GFP plasmid(pAC1185) and viewed by direct fluorescence microscopy using an Olympus

TABLE 1. Yeast strains, cell lines, plasmids, and viruses

Strain, cell line, plasmid, or virus Genotype and/or description Source or reference

ACY193 MAT� ura3-52 leu2�1 his3�200 50ACY247 MATa/� ura3-52/ura3-52 leu2�1/leu2�1 his3

�200/his3�200 ade2/ADE2 ade3/ADE3lys2/LYS2 trp1/TRP1

22

ACY402 (BY4741) MATa his3 leu2 met15 ura3 Research Genetics, Invitrogen Corp., Carlsbad, CAACY783 MATa NPC2::HIS3 ura3 lys2 leu2 This studyACY1248 MATa ARV1::KanMX4 his3 leu2 met15 ura3 Research GeneticsACY1037 MATa LEM3::KanMX4 his3 leu2 met15 ura3 18ACY1260 MATa NPC2::HIS3 ura3 lys2 leu2 This studyHEK293 E1�/� V. FaundezWild-type mouse fibroblasts Wild type 38Wild-type human skin fibroblasts Wild type 14Human NPC2�/� skin fibroblasts hNPC2�/� 14pAC3 (pRS315) LEU2 CEN 45pAC45 pNUF2-NUF2-GFP URA3 2� 23pAC1181 pNPC2-scNPC2 LEU2 CEN scNPC2 was amplified by PCR from genomic

DNA and cloned into pAC3 cut with PstI and XhoIpAC1185 pNPC2-scNPC2-GFP URA3 2� scNPC2 was amplified by PCR from genomic DNA

and cloned into pAC45 cut with PstI and XhoIpAC488 pEGFP-N3 ClontechpAC1611 pShuttle-CMV StratagenepAC1612 pAdEasy-1 StratagenescNpc2p-GFP adenovirus pCMV-scNPC2-GFP This studyhNPC2-GFP adenovirus pCMV-hNPC2-GFP This studyGFP adenovirus pCMV-GFP V. Faundez

1852 BERGER ET AL. EUKARYOT. CELL

on February 9, 2021 by guest

http://ec.asm.org/

Dow

nloaded from

BX60 epifluorescence microscope equipped with a GFP optimized barrier filterand a Photometrics Quantix digital camera.

Phosphatidylcholine internalization in yeast. Cells were grown overnight insynthetic complete media containing 2% glucose and diluted the following morn-ing to early log phase at 25°C. NBD-PC (1-myristoyl-2-[6-[(7-nitro-2-1,3-benz-oxadiazol-4-yl)amino]hexanoyl]-sn-glycero-3-phosphocholine) (Avanti-PolarLipids, Alabaster, AL), dissolved in chloroform, was dried under a stream ofnitrogen, resuspended in dimethyl sulfoxide, and then added to the culture at afinal concentration of 5 �M. Cultures were then incubated with shaking for30 min at 30°C. The cell suspension was washed two times with synthetic com-plete media containing 2% glucose, followed by three washes with SC-NaN3

(synthetic complete media lacking glucose but containing 2% sorbitol and 20 mMsodium azide), and cells were analyzed by direct fluorescence microscopy asdescribed above.

Filipin assays. Yeast filipin assays were performed as follows. Cells weregrown overnight at 30°C in synthetic complete media followed by dilution ofcultures to an optical density of 0.1 and growth for 3 h. Filipin (Sigma) wasdissolved in dimethyl sulfoxide and added to cultures at a final concentration of5 �g/ml for 5 min at 30°C. Live cells were visualized by direct fluorescencemicroscopy as described above.

Filipin assays performed on human fibroblasts were carried out as follows.hNPC2�/� patient fibroblasts (14) grown on coverslips were infected withscNPC2-GFP, hNPC2-GFP, or a GFP control adenovirus and grown for 37 h.Cells were washed in phosphate-buffered saline (PBS) with Ca2�/Mg2� and fixedin 4% paraformaldehyde for 20 min at 4°C. Paraformaldehyde was quenched bywashing twice with PBS containing 25 mM glycine, followed by a single wash withPBS. Cells were incubated with 50 �g of filipin/ml in PBS for 2 h at 37°C,followed by two washes with PBS. Slides were mounted in Gelvatol, and cellswere visualized by confocal/two-photon microscopy using a Zeiss Axiovert 100 Mmicroscope coupled to HeNe1, argon ion, and Verdi-pumped titanium:sapphirelasers. Filipin was excited with the Titanium:sapphire laser, and GFP was excitedby using the argon ion laser. For fluorescence microscopy analysis, multipleGFP-positive cells were imaged for filipin fluorescence. Quantification of themicroscopic analysis was performed by systematically analyzing �150 differentGFP-positive cells from fibroblasts individually infected with viruses encodinghNPC2-GFP, scNpc2p-GFP, or GFP alone. These individual GFP-positive cellswere then qualitatively scored for loss of filipin fluorescence intensity. Quantifiedcells were analyzed from two independent experiments. All images were viewedand acquired by using a Plan Apochromat 63/1.4 oil differential interferencecontrast (DIC) objective and Zeiss LSM 510 sp1 software.

Yeast growth analysis. Yeast cells were grown overnight at 30°C in syntheticcomplete media supplemented with 2% glucose, diluted to early log phase thenext morning, and grown for 3 h at the same temperature. The cell densities ofthe cultures were then measured by using a Bio-Rad SmartSpec3000, and cellswere diluted to an optical density at 600 nm (OD600) of 0.1, followed by a further40-fold dilution into synthetic media containing 2% glucose. A total of 100 �l ofthe diluted cell culture was added into an individual well from a 96-well plate.The plate was shaken continuously at 30°C for 30 h with cell density readingstaken every half hour at OD600 with a ELx808 automated microplate reader(Bio-Tek Instruments, Inc., Winooski, VT).

Yeast Nystatin assays. Yeast cells were grown and diluted as described abovefor yeast growth analysis. Cells were counted, serially diluted, spotted ontocontrol rich media containing no drug or 5 �M Nystatin (Sigma), and grown atroom temperature.

Indirect immunofluorescence microscopy. Immunofluorescence was per-formed as previously described (11). Briefly, cultured cells were placed on iceand fixed in 4% paraformaldehyde in PBS for 20 min. After fixation, parafor-maldehyde was quenched by two washes with PBS containing 25 mM glycine andthen once in PBS. Cells were permeabilized by incubating them with block (15%horse serum, 0.02% saponin in PBS, 2% bovine serum albumin, and 1% fish skingelatin) for 1 h at room temperature. Primary antibody incubation in block wasperformed for 1 h at 37°C. Cells were washed three times with block in betweenantibody incubations and once with PBS prior to mounting in Gelvatol. Allimages were viewed and acquired as described above for filipin assays performedon human fibroblasts.

Cultured cell fractionation. Fractionation of cultured cells was performed asdescribed previously (40). Briefly, fibroblasts were infected with either scNPC2-GFP or hNPC2-GFP adenovirus and grown for 37 h prior to collection. Cellswere homogenized in budding buffer (38 mM potassium aspartate, 38 mMpotassium glutamate, 38 mM potassium gluconate, 20 mM morpholinepropane-sulfonic acid-KOH [pH 7.2], 5 mM reduced glutathione, 5 mM Na2CO3, 2.5 mMMgSO4, 2 mM EGTA) by using a cell cracker with a 12-�m clearance asdescribed previously (8). The homogenate was sedimented at 3,000 rpm for 5 min

to create an S1 supernatant. The resultant S1 supernatants were sedimented in10 to 45% velocity sucrose gradients at 34,000 rpm for 1 h using a BeckmanSW55 rotor. All procedures were performed at 4°C in the presence of theprotease inhibitor Complete (Roche Applied Sciences, Indianapolis, IN). Immu-noblotting was performed by standard methods (48). Briefly, proteins wereresolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and trans-ferred to a polyvinylidene difluoride membrane. Blots were probed with ananti-GFP rabbit polyclonal antibody (1:5,000 dilution) (44). Immunoreactivebands were visualized by chemiluminescence.

Secretion assay. Secretion of scNpc2p was assessed by infecting 2 105

wild-type mouse fibroblasts with adenovirus encoding scNpc2p-GFP, hNPC2-GFP, or GFP alone. After 7 h, the cells were washed with DMEM and thengrown in 800 �l of fresh medium. After 72 h, the medium was collected, and celllysates were prepared by washing cells twice with PBS, followed by lysis. Either65 �l of medium or 30 �g of lysate was used for sodium dodecyl sulfate-polyacrylamide gel electrophoresis and immunoblot analysis as described abovefor cultured cell fractionation.

CTxB assays. Cholera toxin B (CTxB) assays performed on human fibroblastswere carried out as follows. hNPC2�/� fibroblasts were grown and infected asdescribed for filipin assays. Prior to fixation, cells were washed twice with HamF-12 media supplemented with 25 mM HEPES (pH 7.4) and 0.01% bovineserum albumin. Cells were incubated with 20 nM Alexa 555-conjugated CTxB(Molecular Probes/Invitrogen Detection Technologies, Carlsbad, CA) for 1 h at37°C and then processed for fixation and microscopy as described for filipinassays performed on human fibroblasts. Confocal microscopy was carried out byusing a Zeiss Axiovert 100M microscope coupled to HeNe1, argon ion, andVerdi-pumped titanium:sapphire lasers. CTxB was excited with the HeNe1 laser,and GFP was excited by using the argon ion laser. Quantification of the micro-scopic analysis was performed by analyzing GFP-positive cells from fibroblastsindividually infected with viruses encoding hNPC2-GFP, scNpc2p-GFP, or GFPalone. These individual GFP-positive cells were then qualitatively scored for lossof CTxB fluorescence intensity. Quantified cells were analyzed from two inde-pendent experiments. All images were viewed and acquired by using a PlanApochromat 63/1.4 oil DIC objective and Zeiss LSM 510 sp1 software.

RESULTS

S. cerevisiae Homologue of hNPC2. The S. cerevisiae genomeencodes an apparent homologue of the hNPC2 protein that wehave termed scNpc2p (23% identical and 46% similar tohNPC2). scNPC2 corresponds to an uncharacterized openreading frame, YDL046w, which encodes a 173-amino-acidprotein containing a single predicted N-terminal signal peptide(Fig. 1). Mammalian NPC2 proteins contain a conserved lipidbinding domain, termed the ML domain, which is found ineukaryotes, plants, and fungi (21). Interestingly, scNpc2p is theonly S. cerevisiae protein which contains this functional domain(21). Recent work shows that mammalian NPC2 binds to cho-lesterol (13, 26). Site-directed mutagenesis of the hNPC2 pro-tein identified six residues that are critical for hNPC2 functionas assessed by rescue of the cholesterol accumulation defectobserved in hNPC2 mutant fibroblasts (see Table S2 in thesupplemental material) (26). Protein sequence alignmentshows that three of these six functionally important residuesare conserved in scNpc2p (Fig. 1A).

To analyze the evolutionary relationship between yeastNpc2p and mammalian NPC2, 39 ML domain-containing pro-teins from animals and fungi were assessed for divergencefrom yeast Npc2p by using a maximum-likelihood phylogenetictree built from aligned amino acid sequences and rooted onfour ML-containing fungal proteins (Fig. 1B and Table S1 inthe supplemental material). An analysis of phylogenetic dis-tance from scNpc2p, averaged within the three major boot-strap supported clades (Fig. 1B, clades 1 to 3) and compared byanalysis of variance, demonstrates that yeast Npc2p is moreclosely related to the animalia NPC2 proteins than it is to anyof the MD proteins (clade 2 versus clade 3, Newman-Keuls

VOL. 4, 2005 IDENTIFICATION OF A FUNCTIONAL YEAST NPC2 HOMOLOGUE 1853

on February 9, 2021 by guest

http://ec.asm.org/

Dow

nloaded from

test, P 0.0001). Thus, scNpc2p is more closely related tomammalian NPC2 than it is to other ML domain-containingproteins (e.g., the MD proteins), supporting the hypothesisthat scNpc2p is the yeast orthologue of hNPC2.

To assess whether the primary sequence conservation ofyeast and mammalian NPC2 proteins could translate to a con-servation of secondary and tertiary protein structure as well,we created a structural model of yeast Npc2p using the solved

crystal structures of bovine NPC2 (13) and the dust mite al-lergen Der P 2 (10). The yeast model, which was created byusing the sequence alignment shown in Fig. 1A, encompassedonly the amino acids aligned with the mature bNPC2 protein(amino acids 40 to 170). When folded onto the bNPC2 struc-ture, the scNpc2p sequence fits well into seven beta strands (Ato G) arranged into two opposing sheets. We find that themajority of amino acids that are mutated in the NP-C patient

FIG. 1. Yeast Npc2p homology. (A) Protein sequence alignment ofhuman NPC2, bovine NPC2, and S. cerevisiae Npc2p. Identical residuesare shaded and conserved residues are boxed. Amino acids that areessential for cholesterol binding or rescue of hNPC2 mutant phenotypes(26) are marked with an asterisk. Amino acids that are sites of mutationin patients (25, 33, 34, 36) are denoted with a star. The ML domains ofhNPC2, bNPC2, and scNpc2p, which span amino acids 24 to 145, 24 to145, and 44 to 168, respectively, are highlighted in gray (amino acidnumbering includes signal peptide; cleavage sites denoted with arrows).Amino acids used for structural analysis are positioned between the filledarrow heads. (B) Maximum-likelihood phylogenetic tree built by quartet-puzzling of ML domain-containing proteins from animalia and fungi.Nodes with �50% bootstrap support are indicated. Clades 1, 2, and 3 arered, blue, and green, respectively. S. cerevisiae Npc2p is denoted in bold-face. Accession numbers for analyzed proteins are listed in SupplementalTable 1. (C) Structural model of scNpc2p. The structure of bovine NPC2(PDB entry 1NEP) was solved by Friedland et al. (13). It is compared toa homology model of yeast Npc2p generated as described in Materialsand Methods. The backbones of scNpc2p and bNPC2 are depicted in blueand gray and in red and gray, respectively. Amino acid side chains aredepicted in red for scNpc2p and blue for bNPC2. The N and C termini areindicated by an N or C in black, and � strands are labeled A to G in white.Amino acids are indicated with numbering beginning from the first struc-tural residue indicated in panel A. Disulfide bonds are indicated betweencysteine residues, and the � strands are indicated in ribbon diagram.

1854 BERGER ET AL. EUKARYOT. CELL

on February 9, 2021 by guest

http://ec.asm.org/

Dow

nloaded from

FIG. 2. scNPC2 function in yeast. (A) Localization of scNpc2p in yeast. A plasmid encoding scNPC2-GFP (pAC1185) was transformed intowild-type yeast cells, and the GFP tagged protein was visualized by direct fluorescence microscopy. A corresponding DIC image is shown.(B) Growth curve. Cells were grown in synthetic medium for 30 h with cell density readings (OD600) taken every 30 min. Cell density versus timeis plotted for wild-type cells (black) and npc2� cells (gray). The average of three independent experiments is plotted. Standard deviations areindicated. (C) NBD-PC uptake and trafficking is unaffected by deletion of scNPC2. Cells were incubated with 5 �M NBD-PC, washed withSC-NaN3, and assessed by direct fluorescence microscopy. Internalization and localization of NBD-PC is shown for npc2� (panel a), wild-type(panel c), and lem3� (panel e) cells. For the lem3� panel, the image was exposed 20-fold longer (10-s exposure time) than for npc2� and wild-typeimages (0.5-s exposure time) in order to visualize cells. Exposure of lem3� cells incubated with NBD-PC for 0.5 s results in panels devoid of anyfluorescence (data not shown). The fluorescence seen in the presented lem3� panel is similar to that observed for cells not incubated with NBD-PC(data not shown). Corresponding DIC images are shown (panels b, d, and f). (D) Ergosterol accumulates at the plasma membrane of npc2� cells.Cells were incubated with filipin, washed, and visualized by direct fluorescence microscopy. Filipin localization is shown for npc2� and wild-typecells. Corresponding DIC images are shown. (E) npc2� cells are sensitive to nystatin. Cells were grown overnight, diluted to early log phase, grownfor 3 h, and then serially diluted onto plates containing no drug (Control) or 5 �M nystatin. Growth is shown for two separate npc2� isolates,wild-type (nystatin-resistant control), and arv1� (nystatin-sensitive control) cells.

1855

on February 9, 2021 by guest

http://ec.asm.org/

Dow

nloaded from

1856

on February 9, 2021 by guest

http://ec.asm.org/

Dow

nloaded from

population are structurally conserved in the yeast model andare potentially important for either structural integrity or pro-tein-protein interactions (Fig. S1 in the supplemental mate-rial). Thus, the modeling reveals that yeast Npc2p can assumethe same three-dimensional structure as bNPC2.

Intracellular localization of scNpc2p. Our analyses suggestthat a homologue of the hNPC2 protein exists in yeast. Toexamine whether the yeast and mammalian NPC2 proteinsreside in functionally related compartments, we examined thelocalization of scNpc2p in yeast cells. The hNPC2 protein isfound in lysosomal compartments (3, 34, 51). We thereforeexamined whether scNpc2p localizes to the equivalent struc-ture in yeast, the vacuole. The subcellular localization ofscNpc2p was assessed by creating the C-terminal GFP fusionprotein, scNpc2p-GFP which is expressed from its endogenousyeast promoter. Immunoblot analysis shows that the scNpc2p-GFP fusion protein migrates at approximately 51 kDa and,similar to human NPC2 (7), can be deglycosylated by Endo Htreatment to run at its predicted size of �47 kDa (data notshown). We find that scNpc2p-GFP is localized to the lumen ofthe vacuole in wild-type cells (Fig. 2A). These results are con-sistent with those obtained in the recent global localizationstudy of yeast proteins (20). Thus, scNpc2p has a similar lo-calization pattern to that found for hNPC2.

Analysis of npc2� yeast cells. A deficiency in hNPC2 leads toan accumulation of lipids in endocytic compartments, suggestingthat human NPC2 plays a role in lipid trafficking. This hypothesisis supported by recent analyses demonstrating that hNPC2 bindsdirectly to cholesterol (26). Furthermore, analysis of the bNPC2structure reveals that the protein has the potential to accommo-date cholesterol within a cavity in its hydrophobic core (13). Ourhomology modeling suggests that the hydrophobic interior core ofscNpc2p may be able to accommodate an ergosterol molecule(Fig. S2 in the supplemental material), suggesting that it may playa role in yeast lipid metabolism. In order to examine the functionthat scNpc2p plays in yeast cells, we generated two isogenic hap-loid yeast strains deficient for scNPC2 (see Materials and Meth-ods). scNPC2 is not essential for yeast viability and, like cellslacking the yeast hNPC1 homologue, scNcr1p (2, 31), npc2� cellsshow no readily discernible phenotype when grown at varioustemperatures (18, 25, 30, or 37°C) or on minimal or rich media(Fig. 2B and E and data not shown). This lack of differentialgrowth is evidenced by growth curve analyses demonstratingthat npc2� cells grow at a rate similar to that of wild-type cells(Fig. 2B).

We assessed lipid internalization and transport to the vacu-ole in npc2� cells by using a previously characterized fluores-cently labeled phosphatidylcholine (NBD-PC) analog that isinternalized across the plasma membrane by transmembranetransport and conveyed to the vacuole through the prevacuolarcompartment (2, 15, 17, 19). As expected, wild-type cells ac-cumulate NBD-PC in the vacuole (Fig. 2Cc) and lem3� mutantcells, which fail to internalize NBD-PC (18), do not uptake the

lipid (Fig. 2Ce). npc2� cells, similar to wild-type cells, accu-mulate NBD-PC within the vacuole (Fig. 2Ca), suggesting thatneither internalization nor intracellular trafficking of NBD-PCis greatly affected in these cells.

Since the proposed function of hNPC2 is in cholesteroltransport (29), we assessed whether a deficiency in scNPC2would have an effect on the equivalent yeast molecule, ergos-terol. Similar to what has been previously demonstrated (31),ergosterol accumulates in the plasma membrane of wild-typeyeast as detected by filipin fluorescence. Likewise, we saw nosignificant change in ergosterol localization in npc2� cells(Fig. 2D). However, we did find that npc2� cells are sensitiveto the polyene antibiotic nystatin (Fig. 2E), which forms com-plexes with ergosterol (9). This sensitivity was observed for twoisogenic npc2� deletion mutants. As controls, wild-type cellsshow no nystatin sensitivity and a known nystatin sensitivemutant, arv1� (47), shows significantly decreased growth.Since there is no apparent difference in growth rate betweennpc2� and wild-type cells (Fig. 2E, control plate, and seeFig. 2B), nystatin sensitivity suggests that npc2� cells have anergosterol-dependent plasma membrane perturbation.

scNpc2p localization in cultured mammalian cells is similarto the localization of hNPC2. hNPC2 has previously been re-ported to colocalize with a subset of LAMPI (3), cathepsin D(51) and AP-1 � (3) positive structures, indicating that somefraction of the protein resides in lysosomal and Golgi compart-ments at steady state. To test the extent of conservation be-tween species, we assessed the localization pattern of the yeastNpc2 protein expressed in mammalian cells. The subcellularlocalizations of yeast Npc2p and, as a control, hNPC2, wereanalyzed by creating C-terminal GFP fusion proteins. Mouseor human fibroblasts were infected with either scNpc2p-GFPor hNPC2-GFP adenovirus and the GFP fusion proteins werecolocalized with LAMPI, LAMPII, and cathepsin D (markersfor the late endosome and lysosome) as well as with GM130and AP-1 � (Golgi markers). Both hNPC2 and scNpc2p dis-tribution and colocalization with Golgi and lysosomal or-ganelles were indistinguishable in either mouse (Fig. 3) orhuman (Fig. 4) fibroblasts. The finding that hNPC2 andscNpc2p distribute similarly was confirmed by biochemicalfractionation. The subcellular fractionation pattern of bothproteins was examined by velocity sucrose gradient, and theresults from this analysis demonstrate that hNPC2 andscNpc2p elute in identical fractions (Fig. 5A), indicating thatthey reside in similarly sized vesicles and organelles.

hNPC2 is a soluble, secreted protein (34). To test whetherscNpc2p can also be secreted from mammalian cells, we in-fected wild-type mouse fibroblasts with adenoviruses encodingscNpc2p-GFP, hNPC2-GFP, or GFP alone. Cells were grownfor 72 h, followed by collection of the media to evaluate thelevels of secreted protein. Lysates were also prepared from theinfected cells as a control for cellular NPC2 protein level ineach sample. Our immunoblot analysis reveals that both

FIG. 3. Subcellular distribution of hNPC2 and scNpc2p in wild-type mouse fibroblasts. Cells were infected with an adenovirus encoding eitherhNPC2-GFP or scNpc2p-GFP. The GFP-tagged proteins (green) were assessed for their codistribution with LAMPI and LAMPII, markers of lateendosomes/lysosomes, and AP-1, a Golgi marker, (red) by confocal microscopy. Colocalization is indicated in yellow in the merged images.Enlarged images are magnified 200%. Bars, 10 �m.

VOL. 4, 2005 IDENTIFICATION OF A FUNCTIONAL YEAST NPC2 HOMOLOGUE 1857

on February 9, 2021 by guest

http://ec.asm.org/

Dow

nloaded from

1858 BERGER ET AL. EUKARYOT. CELL

on February 9, 2021 by guest

http://ec.asm.org/

Dow

nloaded from

scNpc2p-GFP and hNPC2-GFP are detectable in the culturemedia (Fig. 5B). These results confirm that like hNPC2,scNpc2p can be secreted into the media.

Yeast Npc2p expression functionally complements loss ofhNPC2. Human NPC2�/� fibroblasts accumulate unesterifiedcholesterol and the ganglioside GM1 in late endosomes andlysosomes (34). We tested whether yeast Npc2p could reversethe hNPC2�/� lipid accumulation defects. Human NPC2�/�

fibroblasts (Fig. 6), as well as wild-type control cells (data notshown), were infected with adenovirus encoding scNpc2p-GFP, hNPC2-GFP, or GFP alone. Cholesterol accumulationwas assessed by filipin staining and direct fluorescence two-photon confocal microscopy. GM1 accumulation was assessedby CTxB staining and direct fluorescence confocal microscopy.To quantitate the degree of complementation by scNpc2p,�150 GFP-positive cells from two independent experimentswere counted and scored for reversion of cholesterol and GM1

accumulation defects. The results indicate that, as expected,hNPC2-GFP efficiently restores the transport of cholesteroland GM1 in hNPC2�/� cells (Fig. 6A and C), whereas GFPalone does not (Fig. 6A and C). Importantly, expression ofyeast Npc2p-GFP also restores cholesterol and GM1 transportin hNPC2�/� cells (Fig. 6A and C). When the ability of hNPC2and scNpc2p to rescue the lipid accumulation defects is quan-tified, we find that scNpc2p is able to rescue to the same extentas hNPC2 (Fig. 6B and D). These findings in human cellsindicate that yeast Npc2p is able to functionally complementthe loss of hNPC2 and demonstrate that the yeast Npc2 proteinis a functional homologue of hNPC2.

DISCUSSION

In the present study we have identified a functional homo-logue of the Niemann-Pick disease type C protein, hNPC2, in

FIG. 4. Subcellular distribution of hNPC2 and scNpc2p in wild-type human skin fibroblasts. Skin fibroblasts were infected with an adenovirusencoding either hNPC2-GFP or scNpc2p-GFP. The GFP-tagged proteins (green) were assessed for their codistribution with LAMPI, a marker ofthe late endosome/lysosome, cathepsin D, a lysosomal marker, and GM130, a Golgi marker, (red) by confocal microscopy. Colocalization isindicated in yellow in the merged images. Enlarged images are magnified 200%. Bars, 10 �m.

FIG. 5. (A) Subcellular fractionation of hNPC2 and scNpc2p. Ly-sates from wild-type mouse fibroblasts infected with adenoviruses en-coding hNPC2-GFP or scNpc2p-GFP were analyzed by velocity su-crose gradient sedimentation as described in Materials and Methods.Fractions were collected for immunoblot analysis and blots wereprobed for GFP. Densitometry is plotted versus sucrose fraction per-centage for hNPC2 (black) and scNpc2p (gray). (B) Secretion ofNpc2p. Wild-type mouse fibroblasts were infected with adenovirusencoding hNPC2-GFP, scNpc2p-GFP, or GFP alone. Media and lysatewere collected for immunoblot analysis as described in Materials andMethods. All blots were also probed for tubulin as a lysate loadingcontrol. The results are representative of three independent experi-ments.

VOL. 4, 2005 IDENTIFICATION OF A FUNCTIONAL YEAST NPC2 HOMOLOGUE 1859

on February 9, 2021 by guest

http://ec.asm.org/

Dow

nloaded from

FIG. 6. Yeast Npc2p is a functional homologue of hNPC2. hNPC2�/� patient fibroblasts were infected with an adenovirus encoding hNPC2-GFP as a positive control, scNpc2p-GFP, or GFP alone as a negative control. (A) After infection, cells were fixed and incubated with filipin todetect unesterified cholesterol. Fluorescence was viewed by two-photon/confocal microscopy. Bar, 10 �m. (B) Quantification of rescue ofcholesterol accumulation. The percentage of GFP-positive cells scored filipin negative (% GFP� filipin-negative cells) is shown for hNPC2,scNpc2p, and GFP alone (empty vector). For each sample, n � 152. (C) Cells were incubated with 20 nM CTxB for 1 h at 37°C prior to fixationto detect GM1. Fluorescence was viewed by confocal microscopy. Bar, 10 �m. (D) Quantification of rescue of GM1 accumulation. The percentageof GFP-positive cells scored CTxB negative (% GFP� CTxB-negative cells) is shown for hNPC2, scNpc2p, and GFP alone (empty vector). For eachsample, n � 150.

1860

on February 9, 2021 by guest

http://ec.asm.org/

Dow

nloaded from

yeast. The hNPC2 protein is hypothesized to play a critical rolein cellular cholesterol homeostasis, specifically the retrogrademovement of cholesterol from late endosomes (29), with mu-tations in this protein leading to the fatal neurological disor-der, NP-C (37). This protein is a member of a specialized classof lipid-binding proteins (the ML domain-containing family[21]), and as such, binds directly to (26) and facilitates themovement of cholesterol (34).

On the basis of phylogenetic and molecular modeling anal-yses, we suggest that the yeast genome encodes a homologue ofthe mammalian NPC2 protein. Specifically, we find the evolu-tionary distance between scNpc2p and the mammalian NPC2proteins is smaller than the distance between scNpc2p and themajority of ML domain-containing proteins examined (Fig. 1).Indeed, this phylogenetic analysis suggests that the yeast pro-tein may be a progenitor of the mammalian NPC2 proteins.Interestingly, scNpc2p is very closely related to both Aspergillusoryzae and Neurospora crassa phospholipid transfer proteins(39) (Fig. 1). Based upon this relationship as well as therelationship to the cholesterol binding NPC2 proteins (13,26, 35), we suggest that yeast Npc2p may also bind directly toa lipid cargo.

Our in vivo analyses provide further evidence to support theconclusion that scNpc2p is a functional homologue of hNPC2.Specifically, we find that yeast cells lacking scNPC2 are sensi-tive to an ergosterol interacting drug, nystatin (Fig. 2E). Thisfinding is similar to what has been observed for the yeasthomologue of NPC1, scNcr1p, where a perturbation of thisprotein also causes cells to become sensitive to this drug (31).This finding uncovers a functional link between the two yeastNP-C proteins.

When scNpc2p was expressed in mammalian cells, we foundthat the yeast protein has a similar localization and identicalsubcellular fractionation pattern to that observed for hNPC2(Fig. 3, 4, and 5A). We found that scNpc2p-GFP and hNPC2-GFP partially colocalize with the late endosomal and lysosomalmarkers LAMPI, LAMPII, and Cathepsin D (Fig. 3 and 4), aswell as with the Golgi markers GM130 and AP-1 � (Fig. 3 and4), suggesting that adenoviral expressed hNPC2-GFP andscNpc2p-GFP have a steady-state localization in both late en-dosomal/lysosomal structures and Golgi. Although these pro-teins are expressed from adenoviruses, these results are con-sistent with previous localization studies of endogenoushNPC2 (3, 51). Overall, these results suggest that both hNPC2and scNpc2p reach the same intracellular compartments.

As a definitive test of the evolutionary conservation ofhNPC2, we analyzed the function of scNpc2p in mammaliancells. We found that expression of yeast Npc2p in hNPC2�/�

fibroblasts is able to correct both the cholesterol and GM1accumulation defects observed in these cells (Fig. 6A and C).These results demonstrate that scNpc2p is able to facilitate themovement of both cholesterol and GM1. Since hNPC2 is hy-pothesized to mediate cholesterol transport through directbinding to this lipid (26), the ability of the yeast protein toinduce cholesterol and GM1 transport provides supportive ev-idence for our suggestion that yeast Npc2p can bind directly toa lipid cargo. Thus, yeast Npc2p is homologous to mammalianNPC2, not only in sequence and structure but also in function.These results highlight the mechanistic conservation of lipid

recognition and transport, particularly that of sterols, betweenspecies.

The present study, in combination with previous work, dem-onstrates that the yeast genome encodes functional homo-logues of both hNPC1 (2, 31) and hNPC2. The finding thatboth NP-C proteins are expressed in a single-celled eukaryotecoupled with their ubiquitous expression in human cells (28,37), suggests that they play a role in basic cellular metabolism,likely in lipid or cholesterol/ergosterol homeostasis. The dem-onstration that both scNcr1p and scNpc2p are able to effi-ciently clear cholesterol from NP-C patient cells suggests thatthe mechanism of recognizing and transporting cholesterol byhNPC1 and hNPC2 has also been conserved in these yeastproteins. Therefore, studies in simple genetic model systems,such as yeast, may be helpful in defining the normal cellularfunction of these proteins, thereby providing insight into themolecular mechanisms that underlie NP-C disease.

ACKNOWLEDGMENTS

We thank Harish C. Joshi, Daniel S. Ory, and Silvija I. Staprans forsupplying reagents. We thank Murray Stewart for critical reading ofthe manuscript and John Logsdon for helpful suggestions on our phy-logenetic analysis.

This study was supported by a predoctoral NRSA fellowship fromthe National Institute of Neurological Disorders and Stroke, Na-tional Institutes of Health, to A.C.B. (NS044743), a grant from theNational Institutes of Health to V.F. (NS42599), and a grant fromthe National Niemann-Pick Disease Foundation to A.H.C.

REFERENCES

1. Adams, A., D. E. Gottschling, C. A. Kaiser, and T. Stearns. 1997. Methodsin yeast genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor,N.Y.

2. Berger, A. C., P. K. Hanson, J. W. Nichols, and A. H. Corbett. 2005. A yeastmodel system for functional analysis of the Niemann-Pick type C protein 1(NPC1) homologue, Ncr1p. Traffic 6:907–917.

3. Blom, T. S., M. D. Linder, K. Snow, H. Pihko, M. W. Hess, E. Jokitalo, V.Veckman, A. C. Syvanen, and E. Ikonen. 2003. Defective endocytic traffick-ing of NPC1 and NPC2 underlying infantile Niemann-Pick type C disease.Hum. Mol. Genet. 12:257–272.

4. Brady, G. P., Jr., and P. F. Stouten. 2000. Fast prediction and visualizationof protein binding pockets with PASS. J. Comput. Aided Mol. Des. 14:383–401.

5. Carstea, E. D., J. A. Morris, K. G. Coleman, S. K. Loftus, D. Zhang, C.Cummings, J. Gu, M. A. Rosenfeld, W. J. Pavan, D. B. Krizman, J. Nagle,M. H. Polymeropoulos, S. L. Sturley, Y. A. Ioannou, M. E. Higgins, M.Comly, A. Cooney, A. Brown, C. R. Kaneski, E. J. Blanchette-Mackie, N. K.Dwyer, E. B. Neufeld, T. Y. Chang, L. Liscum, J. F. Strauss III, K. Ohno,M. Zeigler, R. Carmi, J. Sokol, D. Markie, R. R. O’Neill, O. P. van Diggelen,M. Elleder, M. C. Patterson, R. O. Brady, M. T. Vanier, P. G. Pentchev, andD. A. Tagle. 1997. Niemann-Pick C1 disease gene: homology to mediators ofcholesterol homeostasis. Science 277:228–231.

6. Carstea, E. D., M. H. Polymeropoulos, C. C. Parker, S. D. Detera-Wadleigh,R. R. O’Neill, M. C. Patterson, E. Goldin, H. Xiao, R. E. Straub, M. T.Vanier, R. O. Brady, and P. G. Pentchev. 1993. Linkage of Niemann-Pickdisease type C to human chromosome 18. Proc. Natl. Acad. Sci. USA 90:2002–2004.

7. Chikh, K., S. Vey, C. Simonot, M. T. Vanier, and G. Millat. 2004. Niemann-Pick type C disease: importance of N glycosylation sites for function andcellular location of the NPC2 protein. Mol. Genet. Metab. 83:220–230.

8. Clift-O’Grady, L., C. Desnos, Y. Lichtenstein, V. Faundez, J. T. Horng, andR. B. Kelly. 1998. Reconstitution of synaptic vesicle biogenesis from PC12cell membranes. Methods 16:150–159.

9. de Kruijff, B., and R. A. Demel. 1974. Polyene antibiotic-sterol interactionsin membranes of Acholeplasma laidlawii cells and lecithin liposomes. 3.Molecular structure of the polyene antibiotic-cholesterol complexes. Bio-chim. Biophys. Acta 339:57–70.

10. Derewenda, U., J. Li, Z. Derewenda, Z. Dauter, G. A. Mueller, G. S. Rule,and D. C. Benjamin. 2002. The crystal structure of a major dust mite allergenDer p 2, and its biological implications. J. Mol. Biol. 318:189–197.

11. Faundez, V., J. T. Horng, and R. B. Kelly. 1997. ADP-ribosylation factor 1 isrequired for synaptic vesicle budding in PC12 cells. J. Cell Biol. 138:505–515.

12. Fiser, A., R. K. Do, and A. Sali. 2000. Modeling of loops in protein structures.Protein Sci. 9:1753–1773.

VOL. 4, 2005 IDENTIFICATION OF A FUNCTIONAL YEAST NPC2 HOMOLOGUE 1861

on February 9, 2021 by guest

http://ec.asm.org/

Dow

nloaded from

13. Friedland, N., H. L. Liou, P. Lobel, and A. M. Stock. 2003. Structure of acholesterol-binding protein deficient in Niemann-Pick type C2 disease. Proc.Natl. Acad. Sci. USA 100:2512–2517.

14. Frolov, A., S. E. Zielinski, J. R. Crowley, N. Dudley-Rucker, J. E. Schaffer,and D. S. Ory. 2003. NPC1 and NPC2 regulate cellular cholesterol ho-meostasis through generation of low density lipoprotein cholesterol-derivedoxysterols. J. Biol. Chem. 278:25517–25525.

15. Grant, A. M., P. K. Hanson, L. Malone, and J. W. Nichols. 2001. NBD-labeled phosphatidylcholine and phosphatidylethanolamine are internalizedby transbilayer transport across the yeast plasma membrane. Traffic 2:37–50.

16. Guex, N., and M. C. Peitsch. 1997. SWISS-MODEL and the Swiss-Pdb-Viewer: an environment for comparative protein modeling. Electrophoresis18:2714–2723.

17. Hanson, P. K., A. M. Grant, and J. W. Nichols. 2002. NBD-labeled phos-phatidylcholine enters the yeast vacuole via the pre-vacuolar compartment.J. Cell Sci. 115:2725–2733.

18. Hanson, P. K., L. Malone, J. L. Birchmore, and J. W. Nichols. 2003. Lem3pis essential for the uptake and potency of alkylphosphocholine drugs, edel-fosine and miltefosine. J. Biol. Chem. 278:36041–36050.

19. Hanson, P. K., and J. W. Nichols. 2001. Energy-dependent flip of fluores-cence-labeled phospholipids is regulated by nutrient starvation and tran-scription factors, PDR1 and PDR3. J. Biol. Chem. 276:9861–9867.

20. Huh, W. K., J. V. Falvo, L. C. Gerke, A. S. Carroll, R. W. Howson, J. S.Weissman, and E. K. O’Shea. 2003. Global analysis of protein localization inbudding yeast. Nature 425:686–691.

21. Inohara, N., and G. Nunez. 2002. ML: a conserved domain involved in innateimmunity and lipid metabolism. Trends Biochem. Sci. 27:219–221.

22. Jones, A. L., B. B. Quimby, J. K. Hood, P. Ferrigno, P. H. Keshava, P. A.Silver, and A. H. Corbett. 2000. SAC3 may link nuclear protein export to cellcycle progression. Proc. Natl. Acad. Sci. USA 97:3224–3229.

23. Kahana, J. A., B. J. Schnapp, and P. A. Silver. 1995. Kinetics of spindle polebody separation in budding yeast. Proc. Natl. Acad. Sci. USA 92:9707–9711.

24. Kirchhoff, C., C. Osterhoff, and L. Young. 1996. Molecular cloning andcharacterization of HE1, a major secretory protein of the human epididymis.Biol. Reprod. 54:847–856.

25. Klunemann, H. H., M. Elleder, W. E. Kaminski, K. Snow, J. M. Peyser, J. F.O’Brien, D. Munoz, G. Schmitz, H. E. Klein, and W. W. Pendlebury. 2002.Frontal lobe atrophy due to a mutation in the cholesterol binding proteinHE1/NPC2. Ann. Neurol. 52:743–749.

26. Ko, D. C., J. Binkley, A. Sidow, and M. P. Scott. 2003. The integrity of acholesterol-binding pocket in Niemann-Pick C2 protein is necessary to con-trol lysosome cholesterol levels. Proc. Natl. Acad. Sci. USA 100:2518–2525.

27. Koepp, D. M., D. H. Wong, A. H. Corbett, and P. A. Silver. 1996. Dynamiclocalization of the nuclear import receptor and its interactions with transportfactors. J. Cell Biol. 133:1163–1176.

28. Larsen, L. B., P. Ravn, A. Boisen, L. Berglund, and T. E. Petersen. 1997.Primary structure of EPV20, a secretory glycoprotein containing a previouslyuncharacterized type of domain. Eur. J. Biochem. 243:437–441.

29. Liscum, L. 2000. Niemann-Pick type C mutations cause lipid traffic jam.Traffic. 1:218–225.

30. Loftus, S. K., J. A. Morris, E. D. Carstea, J. Z. Gu, C. Cummings, A. Brown,J. Ellison, K. Ohno, M. A. Rosenfeld, D. A. Tagle, P. G. Pentchev, and W. J.Pavan. 1997. Murine model of Niemann-Pick C disease: mutation in acholesterol homeostasis gene. Science 277:232–235.

31. Malathi, K., K. Higaki, A. H. Tinkelenberg, D. A. Balderes, D. Almanzar-Paramio, L. J. Wilcox, N. Erdeniz, F. Redican, M. Padamsee, Y. Liu, S.Khan, F. Alcantara, E. D. Carstea, J. A. Morris, and S. L. Sturley. 2004.Mutagenesis of the putative sterol-sensing domain of yeast Niemann PickC-related protein reveals a primordial role in subcellular sphingolipid dis-tribution. J. Cell Biol. 164:547–556.

32. Marti-Renom, M. A., A. C. Stuart, A. Fiser, R. Sanchez, F. Melo, and A. Sali.2000. Comparative protein structure modeling of genes and genomes. Annu.Rev. Biophys. Biomol. Struct. 29:291–325.

33. Millat, G., K. Chikh, S. Naureckiene, D. E. Sleat, A. H. Fensom, K. Higaki,M. Elleder, P. Lobel, and M. T. Vanier. 2001. Niemann-pick disease type c:

spectrum of he1 mutations and genotype/phenotype correlations in the npc2group. Am. J. Hum. Genet. 69:1013–1021.

34. Naureckiene, S., D. E. Sleat, H. Lackland, A. Fensom, M. T. Vanier, R.Wattiaux, M. Jadot, and P. Lobel. 2000. Identification of HE1 as the secondgene of Niemann-Pick C disease. Science 290:2298–2301.

35. Okamura, N., S. Kiuchi, M. Tamba, T. Kashima, S. Hiramoto, T. Baba, F.Dacheux, J. L. Dacheux, Y. Sugita, and Y. Z. Jin. 1999. A porcine homologof the major secretory protein of human epididymis, HE1, specifically bindscholesterol. Biochim. Biophys. Acta 1438:377–387.

36. Park, W. D., J. F. O’Brien, P. A. Lundquist, D. L. Kraft, C. W. Vockley, P. S.Karnes, M. C. Patterson, and K. Snow. 2003. Identification of 58 novelmutations in Niemann-Pick disease type C: correlation with biochemicalphenotype and importance of PTC1-like domains in NPC1. Hum. Mutat.22:313–325.

37. Patterson, M. C., M. T. Vanier, K. Suzuki, J. A. Morris, E. Carstea, E. B.Neufeld, J. E. Blanchette-Mackie, and P. G. Pentchev. 2001. Niemann-Pickdisease type C: a lipid trafficking disorder, p. 3611–3633. In C. Scriver, A.Beaudet, W. Sly, and D. Vale (ed.), The metabolic and molecular bases ofinherited disease, 8th ed. McGraw-Hill Book Co., New York, N.Y.

38. Peden, A. A., V. Oorschot, B. A. Hesser, C. D. Austin, R. H. Scheller, and J.Klumperman. 2004. Localization of the AP-3 adaptor complex defines anovel endosomal exit site for lysosomal membrane proteins. J. Cell Biol.164:1065–1076.

39. Record, E., S. Moukha, and M. Asther. 1999. Characterization and expres-sion of the cDNA encoding a new kind of phospholipid transfer protein, thephosphatidylglycerol/phosphatidylinositol transfer protein from Aspergillusoryzae: evidence of a putative membrane targeted phospholipid transferprotein in fungi. Biochim. Biophys. Acta 1444:276–282.

40. Salazar, G., R. Love, M. L. Styers, E. Werner, A. Peden, S. Rodriguez, M.Gearing, B. H. Wainer, and V. Faundez. 2004. AP-3-dependent mechanismscontrol the targeting of a chloride channel (ClC-3) in neuronal and non-neuronal cells. J. Biol. Chem. 279:25430–25439.

41. Sali, A., and T. L. Blundell. 1993. Comparative protein modeling by satis-faction of spatial restraints. J. Mol. Biol. 234:779–815.

42. Sambrook, J., and D. Russell. 2001. Molecular cloning: a laboratory manual.Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.

43. Schmidt, H. A., K. Strimmer, M. Vingron, and A. von Haeseler. 2002.TREE-PUZZLE: maximum likelihood phylogenetic analysis using quartetsand parallel computing. Bioinformatics 18:502–504.

44. Seedorf, M., M. Damelin, J. Kahana, T. Taura, and P. A. Silver. 1999.Interactions between a nuclear transporter and a subset of nuclear porecomplex proteins depend on Ran GTPase. Mol. Cell. Biol. 19:1547–1557.

45. Sikorski, R. S., and P. Hieter. 1989. A system of shuttle vectors and yeasthost strains designed for efficient manipulation of DNA in Saccharomycescerevisiae. Genetics 122:19–27.

46. Thompson, J. D., T. J. Gibson, F. Plewniak, F. Jeanmougin, and D. G.Higgins. 1997. The CLUSTAL_X windows interface: flexible strategies formultiple sequence alignment aided by quality analysis tools. Nucleic AcidsRes. 25:4876–4882.

47. Tinkelenberg, A. H., Y. Liu, F. Alcantara, S. Khan, Z. Guo, M. Bard, and S. L.Sturley. 2000. Mutations in yeast ARV1 alter intracellular sterol distributionand are complemented by human ARV1. J. Biol. Chem. 275:40667–40670.

48. Towbin, H., T. Staehelin, and J. Gordon. 1979. Electrophoretic transfer ofproteins from polyacrylamide gels to nitrocellulose sheets: procedures andsome application. Proc. Natl. Acad. Sci. USA 76:4350–4354.

49. Viriyakosol, S., P. S. Tobias, R. L. Kitchens, and T. N. Kirkland. 2001.MD-2 binds to bacterial lipopolysaccharide. J. Biol. Chem. 276:38044–38051.

50. Winston, F., C. Dollard, and S. L. Ricupero-Hovasse. 1995. Construction ofa set of convenient Saccharomyces cerevisiae strains that are isogenic toS288C. Yeast 11:53–55.

51. Zhang, M., M. Sun, N. K. Dwyer, M. E. Comly, S. C. Patel, R. Sundaram,J. A. Hanover, and E. J. Blanchette-Mackie. 2003. Differential trafficking ofthe Niemann-Pick C1 and 2 proteins highlights distinct roles in late endocyticlipid trafficking. Acta Paediatr. Suppl. 92:63–73.

1862 BERGER ET AL. EUKARYOT. CELL

on February 9, 2021 by guest

http://ec.asm.org/

Dow

nloaded from