Niemann-Pick C1 Disease Gene: Homology to Mediators of Cholesterol Homeostasis

See allHide authors and affiliations

Science  11 Jul 1997:
Vol. 277, Issue 5323, pp. 228-231
DOI: 10.1126/science.277.5323.228


Niemann-Pick type C (NP-C) disease, a fatal neurovisceral disorder, is characterized by lysosomal accumulation of low density lipoprotein (LDL)–derived cholesterol. By positional cloning methods, a gene (NPC1) with insertion, deletion, and missense mutations has been identified in NP-C patients. Transfection of NP-C fibroblasts with wild-type NPC1 cDNA resulted in correction of their excessive lysosomal storage of LDL cholesterol, thereby defining the critical role of NPC1 in regulation of intracellular cholesterol trafficking. The 1278–amino acid NPC1 protein has sequence similarity to the morphogen receptor PATCHED and the putative sterol-sensing regions of SREBP cleavage-activating protein (SCAP) and 3-hydroxy-3-methyl-glutaryl coenzyme A (HMG-CoA) reductase.

Genetic disorders have helped to define critical steps of cellular metabolism. For example, elucidation of the biochemical and genetic defects underlying familial hypercholesterolemia (FH) laid the cornerstone for the discovery of the LDL-receptor pathway of cellular cholesterol metabolism (1). Similarly, the role of lysosomes and cholesteryl ester (CE) hydrolase in the processing of the CE core of LDL was revealed by Wolman's syndrome, a lysosomal CE storage disease (2).

Niemann-Pick type C (NP-C) disease is an inherited lipid storage disorder that affects the viscera and central nervous system (3). It occurs at low frequency (affecting one in 106 individuals) and is inherited in an autosomal recessive manner. Both linkage and complementation analyses have shown that at least two separate genes, NPC1 (major locus) andNPC2, induce identical clinical and biochemical phenotypes (4). Cells from NP-C patients are defective in the release of cholesterol from lysosomes (5). This lysosomal sequestration of LDL-derived cholesterol results in cholesterol processing errors, including delayed down-regulation of both LDL uptake and de novo sterol synthesis, as well as repressed cholesterol esterification. The NP-C phenotype suggests that trafficking of lysosomal cholesterol to other cellular membranes is a protein-mediated process.

The NPC interval, previously assigned to pericentromeric chromosome 18 (6), was narrowed to a 1-centimorgan region of 18q11 defined by markers D18S44 and D18S1388 (7, 8). A minimal set of overlapping yeast artificial chromosomes (YACs) and bacterial artificial chromosomes (BACs) was assembled (Fig.1) (7). YACs that define the region were introduced into cultured NP-C cells through spheroplast fusion. YAC 911_D_5 exclusively conferred a normal phenotype to mutant cells, establishing the presence of the NPC1 gene on this YAC (9 ) and reducing the NPC interval to markers D18S1382 and D18S1388, an estimated distance of 300 kb. BACs assembled across the narrowed NP-C interval were subcloned into exon-trapping vectors (10). Among the trapped exons that mapped to YAC 911_D_5 and its associated BACs (Fig. 1), the 3′ exon C59 showed identity to an expressed sequence tag (EST) H11600 (GenBank) and the corresponding cluster of 14 ESTs (WI-14881) identified by UNIGENE (11). Northern blot analysis of a multi-tissue RNA panel (Clontech) with EST clone H11600 as probe identified a transcript of ∼4.9 kb (7). To extend this clone 5′, we designed antisense primers from clone H11600 and used them to amplify sequences from adapter-ligated cDNA libraries (12). Through successive extensions, the sequence of the entire open reading frame (ORF) was identified. Primers corresponding to the 5′ most and 3′ most sequences were then used to amplify a single 4673–base pair (bp) clone, 704-1, containing the entire ORF. We now refer to clone 704-1 asNPC1. The authenticity of this clone was verified by three additional trapped internal exons that mapped to the NPC1ORF (Fig. 1).

Figure 1

Positional cloning of the NPC1 gene. The 1-cM genetic interval, covering a physical distance of ∼1500 kb, is defined by microsatellite markers D18S44 and D18S1388. It is represented schematically by equally spaced loci and is not drawn to scale. Complementation with YAC 911_D_5 (hatched) refined the interval to a region between D18S1382 and D18S1388. BACs were assembled across the NP-C interval and used to generate genomic subclones for exon trapping. Of the resultant trapped inserts, four of the verified exons—A88, A92, E49, and C59—mapped to NPC1. The 4673-bp cDNA is represented by an ORF of 3834 bp and a 713-bp 3′ UTR.

The NPC1 cDNA sequence predicts a protein of 1278 amino acids with an estimated molecular mass of 142 kD (Fig.2A) (13). The NH2-terminus contains 13 hydrophobic amino acids typical of signal peptides that target proteins to the endoplasmic reticulum (ER). Analysis of regions of hydrophobicity and structural motif comparisons predict an integral membrane protein with as many as 13 to 16 possible transmembrane (TM) regions. The COOH-terminus of NPC1 contains a di-leucine motif (LLNF) that serves as a lysosomal targeting sequence for Limp II, a lysosomal resident protein with multiple TM domains (14). This motif also mediates endocytosis (15). Database sequence comparisons revealed extensive resemblance (% identity/% similarity) to uncharacterized NP-C orthologs in mouse (16) (85/93), the yeast Saccharomyces cerevisiae(34/57), and the nematode Caenorhabditis elegans (30/55). A region between residues 55 to 165, which is free of TM domains, is highly conserved, suggesting that it has functional importance. Within this sequence lies a leucine zipper motif (residues 73 to 94) that may mediate protein multimerization as it does for certain transcription factors (17). Residue 506 is a putative tyrosine phosphorylation site (18). Fourteen putative glycosylation sites are conserved between the human and mouse proteins (Fig. 2A).

Figure 2

(A) Predicted amino acid sequence of NPC1. The sequence (cDNA GenBank accession number AF002020) begins with the first methionine. The NH2-terminal sequence, in bold italics, designates the predicted signal peptide (13). Overlined sequences represent a domain that is conserved in mouse (16), C. elegans (GenBank accession numberU53340), and S. cerevisiae (GenBank accession numberU33335) orthologs. Conserved cysteines are underlined. Boxed sequences contain one or more of the 16 predicted TM domains, and the bold sequences therein are those that overlap with the 12 putative TM segments of human PATCHED. Several of the boxed sequences contain two (residues 652 to 709; 744 to 788) or three (residues 1099 to 1164) predicted TM domains. The same TM prediction program and parameters were applied in human and mouse (16) sequences. Three marginally predicted TM domains in human (at residues 532 to 549, 1014 to 1040, and 1060 to 1085) were identified in addition to the remaining 13 conserved between human and mouse. Potential N-glycosylation sites conserved between human and mouse (16) are found at positions 70, 122, 185, 222, 314, 459, 478, 524, 557, 598, 916, 961, and 968. (B) Amino acid sequence homology of NPC1-related proteins. Partial sequences of the sterol-sensing domain are from human NPC1, human PATCHED (PTC; GenBank accession number U59464), human HMG-CoA reductase (HMG-CoA; GenBank accession number M11058), and SCAP from Chinese hamster (SCAP; GenBank accession number U67060). Sequences were aligned with CLUSTAL and shaded with GCG PRETTY software. Residues with greater than 35% identity and similarity are shaded in black or gray, respectively. Regions of overlap in the predicted TM domains of NPC1 and PATCHED are overlined. Arrow indicates the D443N mutation of SCAP that results in sterol insensitivity (20). Dashes represent break in actual amino acid sequence of respective proteins to allow sequence alignment with NPC1. Abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; and W, Trp.

The human NPC1 protein exhibits extensive homology with the TM regions of PATCHED, a morphogen receptor in Drosophila and the defective protein in basal cell nevus syndrome (19). Furthermore, across amino acids 615 to 797, a region containing five predicted TM domains, NPC1 shows homology with sterol regulatory element binding protein (SREBP) cleavage-activating protein, SCAP (20), a modulator of cholesterol-regulated transcription factor activation, and with 3-hydroxy-3-methyl-glutaryl coenzyme A (HMG-CoA) reductase (21), the regulatory enzyme of de novo cholesterol biosynthesis (Fig. 2B). This region of SCAP and HMG-CoA reductase is thought to contain sterol-sensing domains (20). The degree of resemblance (% identity/% similarity) of this region of human NPC1 with the comparable region in human PATCHED, hamster SCAP, and human HMG-CoA reductase is 35/63, 29/59, and 24/62, respectively.

Analysis of single-strand conformational polymorphisms (SSCPs) documented the presence of mutations in the NPC1 gene in NP-C patients (22). We identified eight distinct mutations in nine unrelated NP-C families (Table1). One mutation was a 4-bp insertion that resulted in a frameshift at codon 1205, resulting in a premature termination. There were two multiple nucleotide deletions, including a 73-bp deletion resulting in a frameshift that produced a premature termination at codon 632. To date, five missense mutations have been identified; in two instances identical mutations, Thr to Met at codon 1036 (C→T transition) or Asn to Ser at codon 1156 (A→G transition), were found in unrelated families. Two of the missense mutations (Asn to Ser at codon 1156 and Arg to His at codon 1186 ) altered amino acids that are phylogenetically conserved. None of the mutations was observed in control DNA samples from 68 unaffected and unrelated individuals.

Table 1

Mutations of NPC1 gene in Niemann-Pick type C patients from different families. nt, nucleotide; aa, amino acid; Cmpd heteroz., compound heterozygote; Homoz., homozygote.

View this table:

Introduction of NPC1 expression vectors into culturedNPC1-genotyped human fibroblasts by transient transfection restored a normal cellular phenotype as evaluated by filipin fluorescence (Fig. 3A) (23). In mock-transfected NP-C cultures, only 1.6 ± 1.0% (Fig. 3B) cells did not show aberrant lysosomal cholesterol storage. By contrast, in NPC1-transfected cultures about 21 ± 2% of the mutant cells showed no lysosomal cholesterol accumulation, indicating a significant (P = 0.002) recovery of the normal phenotype. Thus, positional cloning, mutation detection, and cDNA-based functional correction establishNPC1 as the gene responsible for the major form of NP-C disease.

Figure 3

Cytochemical detection of intracellular LDL-derived cholesterol accumulation in NP-C fibroblast cultures transiently transfected with NPC1 cDNA. Transfected NP-C cells were cultured with LDL (23). Cells were stained with filipin, a specific cytochemical marker of unesterified cholesterol (3), and viewed by fluorescence microscopy. Cells transfected with sense 5-4 cDNA (A) show corrected cells (arrows) which contain fewer filipin-fluorescent lysosomes. After transfection with antisense 7-5 cDNA (B), essentially all cells contain intensely filipin-fluorescent lysosomes characteristic of the NP-C phenotype. Bar, 34 μm.

We have previously documented that NP-C cells show (i) excessive lysosomal accumulation of LDL-derived cholesterol, (ii) premature enrichment of cholesterol in trans-cisternal Golgi compartments, and (iii) delayed relocation of cholesterol to and from the plasma membrane (3). The presence of a putative lysosomal targeting motif suggests that NPC1 may move along an endocytic pathway that parallels the distribution of endocytosed cholesterol. The second NP-C genotype (NPC2 ) has an identical pattern of disrupted intracellular cholesterol transport (4), indicating that the sterol transport pathway or pathways may involve multiple proteins acting in tandem or in sequence. Cholesterol-mediated transcriptional regulation involves several proteins in the ER, including a sterol-sensitive protease activator (SCAP) (20), unidentified proteases (24), and protease-mobilized transcription factors (25). When released from the ER, these proteins move to the nucleus to regulate transcription of genes that control the levels of cellular cholesterol and fatty acids. Conceivably, NPC1 has a SCAP-like role, perhaps involving interactions with other proteins such as NPC2. Alternatively, as suggested by its homology with PATCHED, NPC1 may be a receptor for proteins involved in cholesterol transport.

  • * These authors contributed equally to this study.

  • To whom correspondence should be addressed. E-mail: pentchev{at}


View Abstract

Stay Connected to Science

Navigate This Article