Mutations of a Drosophila NPC1 Gene Confer Sterol and Ecdysone Metabolic Defects
Megan L. Fluegel, Tracey J. Parker, Leo J. Pallanck

Abstract

The molecular mechanisms by which dietary cholesterol is trafficked within cells are poorly understood. Previous work indicates that the NPC1 family of proteins plays an important role in this process, although the precise functions performed by this protein family remain elusive. We have taken a genetic approach to further explore the NPC1 family in the fruit fly Drosophila melanogaster. The Drosophila genome encodes two NPC1 homologs, designated NPC1a and NPC1b, that exhibit 42% and 35% identity to the human NPC1 protein, respectively. Here we describe the results of mutational analysis of the NPC1a gene. The NPC1a gene is ubiquitously expressed, and a null allele of NPC1a confers early larval lethality. The recessive lethal phenotype of NPC1a mutants can be partially rescued on a diet of high cholesterol or one that includes the insect steroid hormone 20-hydroxyecdysone. We also find that expression of NPC1a in the ring gland is sufficient to rescue the lethality associated with the loss of NPC1a and that cholesterol levels in NPC1a mutant larvae are unchanged relative to controls. Our results suggest that NPC1a promotes efficient intracellular trafficking of sterols in many Drosophila tissues including the ring gland where sterols must be delivered to sites of ecdysone synthesis.

STEROLS are critical in eukaryotic biology as structural components of all membranes and as precursors for the synthesis of the large eukaryotic family of steroid hormones (Mouritsen and Zuckermann 2004; Payne and Hales 2004; Becher and McIlhinney 2005). While sterols are essential due to the important biological functions they serve, an overabundance of the primary mammalian sterol, cholesterol, is toxic and contributes to a variety of prevalent diseases including heart disease and stroke (Choy et al. 2004; Saini et al. 2004). Accumulating evidence also suggests that alterations in cholesterol homeostasis may contribute to neurodegenerative disorders such as Alzheimer's disease, although the mechanisms by which alterations in cholesterol metabolism affect neuronal integrity remain unclear (Burns and Duff 2002; Reiss et al. 2004; Wellington 2004).

The importance of sterols in eukaryotic biology and disease pathogenesis has stimulated intensive investigation of cholesterol homeostasis in vertebrates. This body of work has clarified many features of cholesterol metabolism, including its de novo synthesis from acetate in the endoplasmic reticulum (ER) (Bloch 1965; Lynen 1966), its packaging into and transport within lipoprotein particles (Oncley 1954; Fredrickson 1957), and the receptor-mediated uptake of these particles by cells (Brown and Goldstein 1986). While these advances have fostered the development of treatment strategies that affect the etiology of hypercholesterolemia, this disorder remains a major source of morbidity, and our knowledge of cholesterol metabolism and the health consequences associated with altered cholesterol homeostasis is far from complete. In particular, the molecular mechanisms by which cholesterol is absorbed by the intestine and trafficked within cells following its acquisition remain poorly understood (Turley and Dietschy 2003; Soccio and Breslow 2004).

Although our current understanding of the molecular events that regulate and direct the movement of cholesterol within cells is limited, previous work has delineated the route by which circulating cholesterol is acquired and trafficked intracellularly (Brown and Goldstein 1986; Liscum and Munn 1999). Cholesterol-containing low density lipoprotein (LDL) particles in the circulatory system are recognized and bound by the LDL receptor on the surface of cells and internalized through a clathrin-mediated endocytic process. The internalized LDL particles are trafficked within an endosome to the lysosome where the LDL particles are degraded. This hydrolysis releases free cholesterol from the LDL core and allows for the subsequent trafficking of cholesterol to the trans-Golgi network (TGN). Cholesterol can then be transported from the TGN to the plasma membrane or to the ER where intracellular cholesterol levels are monitored by a suite of proteins to control the rate of de novo cholesterol synthesis.

Molecular analyses of the genes responsible for rare heritable forms of hypercholesterolemia have made important contributions to our current understanding of cholesterol transport through the circulatory system and its uptake by cells (Goldstein and Brown 1973). Similar insight into the molecular mechanisms of intracellular cholesterol trafficking is beginning to emerge from studies of the genes associated with the autosomal recessive childhood neurodegenerative disorder Niemann-Pick type C disease (NPCD). Cells from NPCD patients display defects in intracellular lipid trafficking and accumulate cholesterol and other lipids within late endosomal and lysosomal compartments (Loftus et al. 1997; Morris and Carstea 1998; Zervas et al. 2001). Endocytosis and intracellular transport of LDL particles to lysosomes appear to proceed normally in cells from NPCD patients while the rates of cholesterol trafficking from lysosomes to the plasma membrane and the TGN are significantly reduced (Liscum et al. 1989; Wojtanik and Liscum 2003). Consequently, the induction of cellular cholesterol homeostatic regulatory responses associated with cholesterol enrichment of ER membranes is delayed in cells from NPCD patients.

NPCD results primarily from loss-of-function mutations of the NPC1 gene (Carstea et al. 1997). The ubiquitously expressed NPC1 polypeptide contains 13 transmembrane domains, a sterol-sensing domain, and a carboxy-terminal dileucine lysosomal targeting sequence (Carstea et al. 1997; Davies and Ioannou 2000). The sequence and domain structure of NPC1 is similar to that of the NPC1L1 protein, which appears to be required primarily or exclusively for the intestinal absorption of dietary sterols (Davies et al. 2000b; Altmann et al. 2004; Garcia-Calvo et al. 2005; Lammert and Wang 2005), and to the morphogen receptor Patched. The NPC1 protein also bears resemblance to the bacterial class of RND permeases and is able to transport fatty acids (Davies et al. 2000a), although the importance of this result is unclear as the flux of these species from lysosomes in NPCD cells is unperturbed (Passeggio and Liscum 2005). Consistent with the predicted topology and lysosomal targeting motifs present in NPC1, subcellular distribution analyses indicate that the NPC1 polypeptide is localized to membranes of late endosomes, lysosomes, and the TGN (Garver et al. 2000; Scott et al. 2004). In conjunction with the cholesterol accumulation phenotype seen upon loss of NPC1 function, these findings strongly suggest a direct role for this factor in the trafficking of sterols through the endocytic/secretory pathway, although precisely how NPC1 mediates these trafficking events is far from clear. Moreover, because of the findings that other lipids and proteins are aberrantly transported in NPC1 mutant cell lines, it is unclear whether the alterations in cholesterol transport are primary or secondary defects, or even whether the specific pathologies associated with NPCD are related to cholesterol trafficking defects or are instead due to defective trafficking of another metabolite altogether (Gondre-Lewis et al. 2003; Vanier and Millat 2003).

We are using Drosophila as a model system to explore the molecular mechanisms of sterol absorption and intracellular trafficking as well as the effects of altered sterol trafficking on neuronal integrity. Drosophila is a cholesterol auxotroph, and a dietary supply of sterols is required for proper development and viability in this organism (Clayton 1964). This dietary requirement is due, at least in part, to the role sterols play as precursors in the synthesis of the essential insect hormone ecdysone (Garen et al. 1977; Feldlaufer et al. 1995). Previous work indicates that dietary sterols are transported in the insect circulatory system in lipoprotein complexes called lipophorins that closely resemble their mammalian equivalents and that these lipoprotein complexes are acquired by cells through receptor-mediated endocytosis (Chino and Downer 1982; Soulages and Brenner 1991; Soulages and Wells 1994). Thus, the mechanisms by which cholesterol is transported through the circulatory system and acquired by cells appear to be highly conserved in insects and vertebrates.

We have focused on the Drosophila NPC1 gene family as a starting point for our studies of sterol trafficking. The Drosophila genome encodes two closely related members of the NPC1 family, here called NPC1a and NPC1b. To explore the functional roles of these proteins in sterol trafficking and neuronal integrity we have characterized their expression patterns and generated null alleles of these genes. In this report we describe the results of these analyses with respect to the NPC1a gene. Our studies indicate that the NPC1a gene is expressed broadly in many tissue types throughout development and is required for viability under normal growth conditions. NPC1a null mutants die at an early larval stage, but feeding NPC1a mutants the steroid hormone ecdysone extends the lethal phase of these mutants suggesting that reduced ecdysone synthesis is one consequence of the NPC1a lesion. Additionally, the inclusion of excess cholesterol in the diet markedly extends the lethal phase of NPC1a mutants and allows some mutants to survive to the adult stage of development. NPC1a mutants can also be rescued to the adult stage by ectopic expression of NPC1a in the ring gland, the site of ecdysone synthesis. These data suggest that NPC1a is required for efficient intracellular cholesterol trafficking to sites of ecdysone synthesis in the ring gland. Our model of NPC1 function offers the potential to further elucidate the mechanistic details of NPC1 function as well as to use classical genetic approaches to identify important factors that function in concert with NPC1.

MATERIALS AND METHODS

Fly strains:

Fly stocks bearing the Df(2L)J2 deletion; P-element insertions P{lacW}me31Bk06607 and P{EPgy2}EY00751; the Δ2-3 source of transposase; and the Gal4 drivers elav–Gal4 (Lin et al. 1994), twist–Gal4 (Thisse et al. 1987), 24B–Gal4 (Fyrberg et al. 1997), Lsp2–Gal4 (Cherbas et al. 2003), and tubulin–Gal4 were obtained from the Bloomington Stock Center. The ring gland drivers P0206–Gal4 (Suster and Bate 2002), phantom–Gal4 (M. O'Connor, personal communication), and 2-286–Gal4 (Timmons et al. 1997) were kindly provided by Lynn Riddiford, James Truman, and Allen Shearn, respectively. NPC1a allele stocks were maintained with a Cy–GFP balancer, kindly provided by Kendal Broadie. All experiments were performed at 25° unless otherwise noted. Flies were raised on standard cornmeal/molasses food.

Identification and analysis of Drosophila NPC1 homologs:

Genomic and cDNA sequences encoding the Drosophila NPC1a and NPC1b proteins were identified by searching the BDGP database using a human NPC1 polypeptide query sequence (AAD48006). One of the NPC1a cDNA clones (LD15757, Research Genetics) was fully sequenced, and this sequence was compared to genomic DNA sequence to verify the locations of splice junctions in the NPC1a gene. Because no NPC1b cDNA clones are available from public databases, NPC1b coding sequences were inferred from theoretical translation of a predicted Drosophila gene (CG12092) and from comparative genomic analysis of corresponding sequences from D. pseudoobscura and Anopheles gambiae. A multiple protein sequence alignment of the human NPC1 and NPC1L1 and Drosophila NPC1a and NPC1b sequences (obtained from the GenBank database) was created using Vector NTI Advance 9.0 (InforMax). This alignment is available as supplemental data at http://www.genetics.org/supplemental/. Percentage identity between human NPC1 and individual homologs was gathered by calculating pairwise alignments within AlignX. Domain identity values for the NPC1 homologs were calculated from the multiple protein alignment using human NPC1 domain designations given in Vanier and Millat (2003).

Generation of NPC1a alleles:

A fly stock bearing the P{lacW}me31Bk06607 insertion was crossed to a stock bearing a stable source of transposase to induce transposition. Flies exhibiting evidence of transposition were screened for insertions near NPC1a by PCR using primers located in the 5′ untranslated region of NPC1a. The Pros35LA1 insertion obtained from the mobilization of P{lacW}me31Bk06607 and the P{EPgy2}EY00751 insertion situated upstream of NPC1a were used in imprecise excision screens to identify NPC1a alleles using established procedures (Robertson et al. 1988). Approximately 400 and 250 putative excision events were recovered from the Pros35LA1 and P{EPgy2}EY00751 insertions, respectively. Molecular analyses of the lines generated from imprecise excision of the Pros35LA1 insertion indicated that none had deletions extending toward the NPC1a gene. The NPC1a14A, NPC1a25A, NPC1a27A, NPC1a57A, and NPC1a94A alleles of NPC1a, as well as the NPC1a65A precise excision control chromosome, were all derived from excision of the P{EPgy2}EY00751 insertion.

Northern blot analysis:

poly-(A)+ RNA from major Drosophila life stages was purchased (Clontech) and total RNA was isolated from tissue with TRIzol reagent (Sigma, St. Louis). RNA extraction was performed according to manufacturer instructions. Briefly, animals were collected at the desired stage and homogenized in TRIzol reagent. Debris was removed by centrifugation and the sample was extracted with chloroform. RNA was precipitated from the aqueous phase with isopropanol, resuspended in RNase-free water, and stored at −80° until use. NPC1a riboprobes were generated using PCR products corresponding to nucleotides 1819–2834 of the NPC1a coding sequence. The PCR primers (Invitrogen) used to generate the NPC1a probe were as follows: 5′-GTCCCGTAGATCCAGCCATA-3′ and 5′-CGGCATAGCCAACTCTTGAT-3′. The Pros35 riboprobe sequence is homologous to nucleotides 225–691 of Pros35. PCR amplification for this riboprobe used primers 5′-ACTTGTCGCCCTGTGTAA-3′ and 5′-TAGGCAGGGTACCGAGTATA-3′. The rp49 riboprobe sequence is homologous to nucleotides 18–466 of rp49. PCR amplification of this region used primers 5′-AACCCTCGAGATGACCATCCGCCCAGCATA-3′ and 5′-AACCTCTAGAGTGTATTCCGACCAGGTTAC-3′. PCR products for each target gene were cloned into the TOPO-II cloning kit (Invitrogen). Probe constructs were linearized upstream of the insert and purified for use in an antisense transcription reaction. Digoxigenin (DIG)-labeled riboprobes were transcribed using the Roche DIG-labeling kit or separately purchased kit components. Northern blots were conducted by standard methods (Sambrook et al. 1989). An alkaline phosphatase-conjugated antidigoxigenin antibody (Roche) was used at a 1:10,000 dilution to detect target sequences. Signal was generated through the use of CDP-Star (Roche) diluted 1:100.

Embryo in situ hybridization:

w1118 embryos were collected overnight and fixed according to standard protocols, and in situ staining was done according to established procedures (Tautz and Pfeifle 1989). NPC1a DIG-labeled riboprobes were synthesized as described above and hydrolyzed with 100 mm NaCO3 before use. Anti-DIG–alkaline phosphatase (1:2000) and nitro blue tetrazolium chloride/5-bromo-4-chloro-3-indolyl phosphate (NBT/BCIP) (Roche, 1:50) were used for detection of DIG signal.

Generation of transgenic constructs:

The NPC1a genomic rescue construct was generated by subcloning a 9.4-kb KpnI–SalI fragment containing the complete NPC1a coding sequence from the BAC clone RPC1-98 (BACPAC Resource Center, Oakland, CA) into a modified form of the Drosophila transformation vector pUAST, which lacks UAS motifs. The Pros35 genomic rescue construct was similarly generated by subcloning an 8-kb BamHI fragment containing the complete Pros35 sequence from the same BAC clone into the UAS-deficient pUAST vector. The UAS–NPC1a transgene was generated from cDNA clone LD15757, which contains a complete NPC1a open reading frame along with contaminating bacterial sequence. To separate the NPC1a sequence from the contaminating bacterial sequence, PCR primers containing engineered NotI/KpnI sites were used to amplify and subclone the NPC1a coding sequence into pUAST. The specific primers used for PCR amplification were 5′-GTCTAGAGCGGCCGCCACAACCAAAATGAGTCCAAGATCCCCGC-3′ and 5′-CTAGTGGGTACCATTACAATACAACTATGGG-3′. The integrity of the isolated NPC1a sequence was subsequently verified by DNA sequencing. All of the transgenic constructs generated were injected into w1118 embryos using established methods to obtain transgenic lines (Ashburner 1989).

Lethal-phase analysis:

Homozygous first instar larvae (n > 30) for each of the NPC1a alleles were collected from parents heterozygous for a balancer chromosome marked with GFP using 488-nm light to select larvae that lack GFP expression. These larvae were placed onto standard food and developmental progress was noted. Larval instars were distinguished by spiracle and mouth hook development (Bate and Martinez Arias 1993).

Analysis of neuronal integrity:

The heads of age-matched flies were placed into Karnovsky's fix for at least 72 hr and prepared for thin sections. Briefly, heads were moved to 0.5× Karnovsky's fix in a vacuum oven overnight. Heads were rinsed three times in 0.1 m cacodylate before an Osmium post-fix overnight in a vacuum oven. Heads were rinsed in distilled water and then in 35% ethanol twice before placing in uranyl acetate for 1 hr. Heads were dehydrated in ethanol and then placed in propylene oxide. Samples were equilibrated with epon overnight before preparing 5-μm sections for analysis.

Preparation of Drosophila medium:

Cornmeal/molasses food was made with a standardized recipe: 72 g cornmeal, 24 g brewer's yeast, 9 g agar, and 106 ml molasses in 950 ml water were cooked to yield ∼1 liter of food. Briefly, molasses and 75% of the total water were heated in a double boiler apparatus and boiled for 15 min. A suspension of cornmeal, agar, and yeast in the remaining water was then added, and the food cooked at 95° for 30–45 min. Heat was reduced, and 14 ml nipagin (10% stock in 95% ethanol) was added once food reached 60°. Food was distributed to desired containers and allowed to cool overnight at room temperature before plugging and storing at 4°.

Cholesterol feeding assay:

Drosophila food containing cholesterol was made by adding cholesterol to standard food preparations after the addition of nipagin. Cholesterol (Sigma) was used from a 30-mg/ml stock solution in 100% ethanol to produce a final concentration of 200 ng/ml. To ensure adequate mixture and dispersion of cholesterol, food was mixed 100 ml at a time. Larvae were collected from grape juice agar plates within 2 hr of hatching and placed in glass vials containing either standard food or food plus cholesterol (n = 30 for each of three trials). Experiments were conducted blind, larval development was monitored at 25°, and the lethal phase was noted.

Parental rescue experiments were conducted by placing the NPC1a57A heterozygous stock into bottles containing 200 ng/ml cholesterol-containing food for 6 days. After this feeding period, the flies were transferred to a laying cage, and homozygous NPC1a57A first instar larvae were collected and aged on standard food lacking supplemental cholesterol poured in 35-mm petri dishes (n = 30 for each of three trials). The separate feeding of heterozygous NPC1a57A adults was done in bottles containing 200 ng/ml cholesterol-containing food for 6 days. After this feeding period, fed flies were mated with nonfed flies in a laying cage. The development of homozygous NPC1a57A larvae on standard food was monitored.

Ecdysone feeding assay:

Drosophila food containing 20-hydroxyecdysone (20HE) was made by adding 20HE (Sigma) to standard food preparations after the addition of nipagin. A 5-mg/ml stock in 100% ethanol was diluted to a final concentration of 120 μg/ml. Food for the ecdysone feeding assay was poured into 35-mm petri dishes (Falcon). To provide pulses of ecdysone that are required for proper development, larvae were transferred periodically between food lacking and food containing 20HE. Briefly, larvae were fed for 8 hr on standard food immediately after a molt and then moved onto food either with or without ecdysone treatment until the next molt. This cycle was repeated for each molt. First instar larvae were collected from grape juice agar plates within 2 hr of hatching, placed on standard food, and the plates were coded for blinding (n > 30 for three trials). The lethal phase was noted over the course of development.

Gal4 rescue experiments:

Flies containing the NPC1a57A allele in trans to a GFP-marked balancer chromosome and the UAS–NPC1a transgene were mated to flies heterozygous for the NPC1a57A allele and containing the desired Gal4 driver, and rescue was assessed at the third instar larval, pupal, and adult stages by identifying individuals lacking GFP expression. The genotype of rescued individuals was verified using a PCR reaction that simultaneously amplifies and distinguishes the NPC1a57A deletion from the wild-type NPC1a gene.

Sterol quantitation:

The Amplex Red cholesterol assay kit (Molecular Probes) was used to assess sterol content in NPC1a57A and NPC1a65A homozygous larvae as well as prepared Drosophila media. First instar larvae were collected from grape agar plates lacking yeast and placed on yeast paste for 8 hr to feed. Larvae were then washed from the yeast paste, rinsed of all yeast, and transferred to apple juice agar lacking yeast for 3 hr. Larvae were then weighed and homogenized in 150 mm NaCl, 50 mm Tris pH 7.5, 2 mm EGTA to make a 100-μg/μl larval homogenate. The homogenate was subjected to centrifugation at 5000 rpm for 5 min to pellet cuticle debris, and the supernatant was used to assay sterol content according to manufacturer instructions. Equal weights of standard and cholesterol-supplemented media were assayed with a similar protocol. Fluorescence was measured with a Packard FluoroCount fluorometer with a 530/590-nm filter set.

RESULTS

Identification of NPC1 homologs in the Drosophila genome:

To identify Drosophila homologs of the NPC1 gene the human NPC1 protein sequence was used to conduct a BLAST search of the translated Drosophila genome sequence. Two Drosophila NPC1 homologs, designated NPC1a and NPC1b, were identified from this search. The NPC1a gene maps to polytene region 31B1 of the second chromosome and encodes a 1287-amino acid polypeptide with 42% identity to NPC1. The NPC1b gene maps to polytene region 19E5 on the X-chromosome and encodes a 1254-amino acid polypeptide with 35% identity to NPC1. The NPC1a and NPC1b proteins exhibit even greater sequence conservation to NPC1 within defined protein domains, such as the sterol-sensing domain and the cysteine-rich domain (Figure 1). In addition to these conserved protein domains, the NPC1a, NPC1b, and NPC1 proteins exhibit similar hydrophobicity patterns suggesting that transmembrane topology is also conserved among this family of proteins (data not shown).

Figure 1.—

The Drosophila NPC1 homologs are highly conserved with the human NPC1 protein family. A diagram of the human NPC1 (hNPC1), human NPC1L1 (hNPC1L1), and Drosophila NPC1a and NPC1b protein primary structures with three functionally important domains: the NPC1 domain (NPC1), the sterol sensing domain (SSD), and the cysteine-rich domain (cys-rich), noted. The percentage identity to human NPC1 within each of the three domains is noted above the homolog sequences. The amino acid positions delineating these domains are noted below each sequence. A multiple protein alignment of the human NPC1 and NPC1L1 proteins and fruit fly NPC1a and NPC1b homologs was used to calculate percentage identity in these three domains (this alignment is included as supplemental data at http://www.genetics.org/supplemental/). Overall percentage identities to human NPC1 and NPC1L1 were calculated by constructing pairwise alignments of the homologs to human proteins.

Vertebrate genomes also encode a pair of closely related NPC1 paralogs that appear to possess distinct functions. The more recently identified protein, designated NPC1L1, has been found to function in sterol absorption in the intestine (Davies et al. 2000b; Altmann et al. 2004; Garcia-Calvo et al. 2005; Lammert and Wang 2005), while NPC1 appears to play a role in intracellular lipid and sterol trafficking in most cell types. To explore the possibility that the Drosophila NPC1a and NPC1b proteins are orthologous to specific vertebrate NPC1 family members we compared the sequences of the human NPC1 family members to the two Drosophila homologs. Results of this analysis indicate that the Drosophila NPC1a and NPC1b proteins are more similar both to human NPC1 and to each other (38% identity) than to NPC1L1 (Figure 1). NPC1a displays greater sequence similarity to both NPC1 and NPC1L1 than NPC1b does, raising the possibility that the gene duplication events that gave rise to multiple NPC1 family members in flies and humans occurred independently in these organisms. While these observations complicate efforts to define strict orthologous relationships within this gene family, several sequence features suggest that NPC1a and NPC1b may function equivalently to NPC1 and NPC1L1, respectively. In particular, the findings that NPC1a has greater sequence similarity to NPC1 and that NPC1a and NPC1 (but not NPC1b or NPC1L1) share a C-terminal lysosomal targeting motif suggest that NPC1a may play a functional role that more closely parallels that of NPC1 (Ioannou 2000; Scott et al. 2004). Furthermore, the NPC1L1 and NPC1b polypeptides share a putative YQRL Golgi retention signal (Humphrey et al. 1993) and no such sequence is found in NPC1 or NPC1a (see supplemental Figure 1 at http://www.genetics.org/supplemental/). In this report we focus on the function of the NPC1a gene. Our ongoing mutational analysis of the NPC1b gene will be described elsewhere.

Characterization of NPC1a expression:

To explore the expression pattern of NPC1a we performed Northern blot and in situ hybridization analyses (Figure 2). Northern blot analysis was conducted using poly(A)+ RNA isolated from the embryonic, third instar larval, and adult stages of development. This analysis indicates that the predicted 4.2-kb NPC1a transcript is expressed at constant levels throughout development. In situ hybridization staining of embryonic tissue indicates that the NPC1a transcript is expressed ubiquitously throughout this stage of development.

Figure 2.—

NPC1a is expressed at uniform levels throughout development and ubiquitously in the embryo. (A) Equal amounts of poly(A)+ RNA from embryo (E), third instar larval (L), and adult (A) stages was used in Northern blot analysis and probed with an antisense probe for NPC1a. A 4.2-kb transcript is expressed at similar levels at all developmental stages analyzed. (B) In situ hybridization in embryonic tissues using an antisense probe for NPC1a indicates that the transcript is present throughout the embryo at all stages. (C) A sense probe for NPC1a was used as a background control for the in situ hybridization and shows no staining at the embryonic stage. The embryonic stage of development is indicated.

Generation and characterization of NPC1a alleles:

To investigate the biological function of the NPC1a gene, we used a transposon mutagenesis approach to generate alleles of this locus. An exogenous source of transposase was used to induce transposition of the P{lacW}me31Bk06607 element, which lies ∼26 kb away from NPC1a, into the NPC1a gene. From this analysis a single insertion, designated Pros35LA1, was recovered (Figure 3). The Pros35LA1 insertion resides just upstream of the NPC1a gene in the predicted 5′ untranslated sequence of the adjacent Pros35 gene and confers a recessive first instar larval lethal phenotype as a homozygote and in trans to a chromosome that bears a deletion of the Pros35/NPC1a region, designated Df(2L)J2. The recessive lethal phenotype of the Pros35LA1 insertion appears to derive from inactivation of the Pros35 gene because Pros35 expression is severely attenuated by this insertion (data not shown), and the recessive lethal phenotype can be fully rescued by a genomic construct bearing an intact copy of the Pros35 gene (Figure 3A). A second P-element insertion, P{EPgy2}EY00751, was obtained more recently from a genome project consortium and resides near the predicted NPC1a transcriptional start site. The P{EPgy2}EY00751 insertion is fully viable as a homozygote with slightly reduced NPC1a transcript levels (Table 1).

Figure 3.—

The NPC1a genomic region and mutant alleles. (A) The relative locations of the NPC1a and Pros35 transcription units. The transcriptional start sites and intron/exon boundaries of the NPC1a and Pros35 genes were predicted using an algorithm and confirmed by direct sequencing of NPC1a and Pros35 cDNAs. The two P-element transposons, EY00751 and Pros35LA1, that reside upstream of the NPC1a and Pros35 genes are designated by triangles. The regions deleted in the NPC1a27A and NPC1a57A alleles derived from the EY00751 insertion are represented as horizontal lines above the transcript map. The NPC1a and Pros35 genomic transgenic rescue constructs are designated by horizontal lines below the transcript map. (B) Total RNA was isolated from larvae homozygous for the indicated NPC1a allele and probed for NPC1a, Pros35, and rp49 expression levels. NPC1a57A larvae lack expression of NPC1a, while NPC1a27A larvae show reduced NPC1a expression and altogether lack Pros35 expression. NPC1a65A is derived from a precise excision of the EY00751 transposon and is used as a wild-type control for experiments involving the NPC1a57A and NPC1a27A alleles. rp49 expression is used as a loading control.

View this table:
TABLE 1

Summary of NPC1a alleles

In an effort to identify more severe NPC1a alleles, the Pros35LA1 and P{EPgy2}EY00751 insertions were used in imprecise excision mutagenesis screens. Among >600 putative imprecise excision events, two deletions from the excision of P{EPgy2}EY00751 that extend into the NPC1a gene were identified (Figure 3A). The NPC1a27A deletion removes the first noncoding exon and part of the first intron of NPC1a, as well as the first third of the Pros35 transcript. Animals homozygous for the NPC1a27A deletion die early during the first instar larval stage. The NPC1a57A deletion extends from the P{EPgy2}EY00751 insertion site exclusively toward the NPC1a gene and removes the first noncoding exon, the first intron, and the first 25 bases of the second exon of NPC1a. This deletion includes the first seven bases of the NPC1a open reading frame. Homozygous NPC1a57A mutants die primarily as first instar larvae, but some NPC1a57A mutants (<10%) are able to survive until the second instar larval stage of development.

Several experiments were performed to test whether the recessive lethal phenotypes conferred by the NPC1a27A and NPC1a57A alleles derive from loss of NPC1a and/or Pros35 function. First, total RNA was isolated from homozygous larvae of these two lines and probed for NPC1a and Pros35 expression (Figure 3B). This analysis failed to detect NPC1a transcripts from NPC1a57A larvae but revealed normal or near normal abundance of Pros35 transcripts. NPC1a27A larvae have significantly attenuated NPC1a expression and completely lack Pros35 transcripts. Second, NPC1a and Pros35 transgenic constructs (Figure 3A) were used in an attempt to rescue the lethal phenotypes resulting from placement of the NPC1a27A and NPC1a57A deletions in trans to the Df(2L)J2 deletion chromosome. A genomic construct bearing an intact copy of the NPC1a gene was found to completely rescue the lethal phenotype of NPC1a57A hemizygotes, but failed to rescue the lethal phenotype of NPC1a27A hemizygotes. By contrast, a genomic construct bearing an intact copy of the Pros35 gene failed to rescue the lethal phenotype of NPC1a57A hemizygotes but was able to rescue the lethal phenotype of NPC1a27A hemizygotes. Third, the Pros35LA1 insertion, an allele of the Pros35 gene, was used in complementation analyses with the NPC1a57A and NPC1a27A mutations. The Pros35LA1 insertion fully complements the NPC1a57A allele but fails to complement the NPC1a27A deletion. Together these data indicate that the recessive lethal phenotype of NPC1a57A homozygotes derives from the loss of NPC1a function while the recessive lethal phenotype of NPC1a27A homozygotes results at least primarily from loss of Pros35 function.

Several lines of evidence indicate that the NPC1a57A deletion represents a null allele of NPC1a. NPC1a transcripts are undetectable in NPC1a57A homozygotes, and the lethal phase of an NPC1a57A homozygote is indistinguishable from that of animals bearing the NPC1a57A deletion in trans to the Df(2L)J2 deletion. Although the results of transgenic rescue experiments indicate that the recessive lethality of the NPC1a27A deletion derives at least primarily from loss of Pros35 function, NPC1a27A homozygotes also have significantly diminished NPC1a expression. Moreover, the NPC1a57A and NPC1a27A deletions do not fully complement one another, with adult escapers occurring at ∼7.5% of expected Mendelian ratios. These results indicate that the NPC1a27A deletion also represents a hypomorphic allele of NPC1a and that greatly reduced NPC1a expression is compatible with viability. The presence of low levels of NPC1a transcripts in NPC1a27A homozygotes raises the possibility that there may be a second uncharacterized transcriptional start site for NPC1a within the first intron of this gene.

To identify additional NPC1a alleles the NPC1a57A allele was used in complementation analysis with other lines recovered from the imprecise excision screen using the P{EPgy2}EY00751 insertion. Three lines, designated NPC1a14A, NPC1a25A, and NPC1a94A, fail to fully complement the NPC1a57A recessive lethal phenotype indicating that they represent NPC1a alleles. Molecular analyses indicate that each of these three lines retains a rearranged version of the P{EPgy2}EY00751 insertion with no substantial surrounding genomic deletion. NPC1a transcript abundance is significantly reduced in all three of these strains relative to a stock bearing a precise excision of the P{EPgy2}EY00751 insertion (designated NPC1a65A), but Pros35 transcript levels are unperturbed (data not shown). These findings indicate that the NPC1a14A, NPC1a25A, and NPC1a94A alleles represent hypomorphic alleles of the NPC1a gene. NPC1a94A larvae invariably die as prewandering third instar larvae, but both NPC1a14A and NPC1a25A exhibit lethality at a number of different stages of development and occasionally produce viable adults (Table 1).

The NPC1a mutants exhibit interesting and possibly informative developmental arrest phenotypes. While all NPC1a57A homozygous larvae die before molting into the third larval instar, they live for prolonged periods of time in their final larval stage. Ordinarily the first and second instar larval stages each last ∼24 hr in wild-type animals, but NPC1a57A larvae can live up to 10 days in their terminal stage of development. It appears that these animals die as a result of a failure to molt into the next larval instar. The long-lived larvae gradually acquire a swollen appearance apparently owing to the accumulation of fluid (data not shown). These phenotypes are also frequently observed in third instar larvae that are homozygous for the NPC1a14A and NPC1a25A hypomorphic alleles. These larvae frequently fail to pupariate, and among the larvae that attempt to enter the pupal stage a fraction of the animals appear to do so aberrantly (Figure 4). These malformed pupae arrest further development and die during this quasi-pupal stage. However, a small fraction of NPC1a14A and NPC1a25A homozygotes pupate normally and eclose. These adult flies appear to be morphologically and behaviorally normal and exhibit an apparently normal life span. Additionally, these adult escapers retain normal brain morphology indicating that these hypomorphic NPC1a alleles do not detectably influence neuronal integrity (data not shown).

Figure 4.—

NPC1a hypomorphic alleles display defects in puparium formation. Examples of NPC1a mutants that fail to fully execute pupal development are shown. Mutant puparia are frequently larger than normal and exhibit aberrant morphological phenotypes such as improper spiracles and altered cuticle morphology. NPC1a65A is a precise excision control chromosome, whereas NPC1a14A and NPC1a25A are hypomorphic NPC1a alleles. All images are to scale.

The lethal phase of NPC1a mutant animals is modified with ecdysone feeding:

The prolonged larval instars and the failure to molt and pupate properly are suggestive of defects in the processes influencing molting and metamorphosis in NPC1a mutant larvae. While many factors control the timing and execution of molting and metamorphosis in insects, one of the best characterized is the steroid hormone ecdysone (Garen et al. 1977). Ecdysone is synthesized from a dietary sterol precursor in the larval ring gland, a small endocrine organ, and released into the hemolymph in response to prothoracicotropic hormone, a neurosecretory peptide hormone secreted by the brain (Henrich et al. 1987; Kim et al. 1997; Zitnan et al. 1993).

To test whether the NPC1a phenotypes derive from reduced ecdysone production, homozygous NPC1a57A first instar larvae were collected and reared in either the absence or presence of 20-hydroxyecdysone (20HE), the active form of ecdysone. The addition of 20HE to the diet had a dramatic impact on the lethal phase of the NPC1a57A animals. When reared on food lacking 20HE, all NPC1a57A larvae died at the first or second instar stage. In contrast, 48% of the population survived past the second instar stage when the larvae were fed dietary 20HE (Figure 5). In rare cases larvae were able to pupate, although none of the pupae produced viable adults. The Pros35 null allele, NPC1a27A, was used as a negative control in this experiment, and no change in the lethal phase of these animals or in the development of precise excision genetic background control animals was observed upon feeding 20HE (Figure 5 and data not shown). These data indicate that the loss of NPC1a expression perturbs the ecdysone signaling pathway and that this defect is responsible for the developmental arrest seen in these animals.

Figure 5.—

The addition of 20HE to the diet of NPC1a null animals extends their life span. (A) NPC1a27A and NPC1a57A homozygous larvae fed a standard diet without 20HE all die before the third instar larval stage of development. The NPC1a27A larvae serve as a negative control for this experiment because these larvae die independently of an NPC1a perturbation. (B) NPC1a57A homozygous larvae exhibit an extended lethal phase when fed a diet containing 120 μg/ml 20HE. By contrast, the addition of 20HE to the diet of NPC1a27A larvae has no effect on the lethal phase of these mutants. Treatment of wild-type larvae with 20HE had no detectable effect on viability (not shown). Graphs display mean ± standard error of the population mean. n > 80 for all genotypes and treatments.

Excess dietary cholesterol can rescue NPC1a mutant animals to adulthood:

The synthesis of ecdysone in the prothoracic gland of the Drosophila ring gland is controlled by the activity of multiple enzymes that convert dietary sterols into ecdysone (Petryk et al. 2003; Warren et al. 2004). Because cholesterol is an upstream compound in the ecdysone synthesis pathway and vertebrate NPC1 putatively functions to traffic cholesterol, we asked whether the lethal-phase modification seen with dietary ecdysone can be explained by an upstream disruption of sterol trafficking or absorption. To test this idea we fed larvae lacking NPC1a expression excess cholesterol and monitored their development.

Homozygous first instar larvae were collected and continuously raised on food either with or without supplemental cholesterol. When NPC1a57A larvae are raised on standard food supplemented with 200 ng/ml cholesterol (1.2-fold higher sterol than standard food) their developmental progress is improved substantially, even to the adult stage (Figure 6). This result is specific to the NPC1a lesion present on the NPC1a57A chromosome because the lethal phase of NPC1a27A larvae was unchanged on high cholesterol food. Additionally, no effect was observed on the development of animals homozygous for a precise excision allele of the NPC1a gene (data not shown). The cholesterol-fed NPC1a null animals appear to be generally normal in morphology. Frequently, however, these animals arrest as wandering third instar larvae and fail to form a puparium. These larvae look strikingly similar to the hypomorphic NPC1a mutants. Cholesterol-rescued adults lacking NPC1a expression appear to be behaviorally normal although they are male-sterile. Additionally, the NPC1a null adults show a substantial reduction in life span with 50% of the population dead by day 29 vs. day 56 for precise excision controls. Interestingly, we find that the life span of the mutant adults is not significantly altered in the presence of excess dietary cholesterol (data not shown). The mechanisms underlying the shortened life span and male sterility of these adults are currently being investigated.

Figure 6.—

NPC1a mutants display an extended life span on a high cholesterol diet. (A) NPC1a57A and NPC1a27A homozygotes on a standard diet all die before the third instar larval stage. (B) NPC1a57A animals raised on standard food supplemented with 200 ng/ml cholesterol exhibited a dramatic extension in life span. The negative control NPC1a27A animals, however, were not affected by the change in diet. The inclusion of cholesterol in the diet of wild-type larvae does not affect their development (not shown). (C) When adult flies heterozygous for the NPC1a57A allele are fed a diet supplemented with 200 ng/ml cholesterol, their homozygous offspring display an extended lethal phase when reared on standard food lacking supplemental cholesterol. Graphs display mean ± standard error of the population mean. n > 80 for all genotypes and treatments.

To further explore the impact of cholesterol on the development of NPC1a mutant larvae we fed NPC1a57A heterozygous adults a high cholesterol diet and allowed their homozygous offspring to grow on food not supplemented with cholesterol. Interestingly, these homozygous offspring show an extended lethal phase and develop into pharate adults, a late pupal stage of development (Figure 6). To test whether this extension in lethal phase derives from enhanced maternal deposition of sterol into the developing egg or from an unknown paternal effect of a high sterol diet we separately fed heterozygous males and females a high cholesterol diet and mated them to heterozygotes that had been maintained on a standard diet. Interestingly, the most substantial rescue of the homozygous progeny occurs when both parental sexes have been fed a high sterol diet although significant, but less efficient, lethal-phase extension is seen upon feeding of only parental males or females (data not shown). From these data we conclude that the lethal-phase extension upon the feeding of a high-sterol diet to parents results from increased sterol deposition in the embryo. We hypothesize that the increased embryonic sterol content can be achieved by increased maternal loading of sterols into the developing egg or upon fertilization of eggs by sperm with a higher membrane cholesterol content. This higher embryonic sterol level may support higher levels of ecdysone synthesis required for developmental progression.

Our cholesterol and ecdysone feeding results correlate well with the hypothesis that NPC1a promotes intracellular sterol trafficking in the ring gland. However, an alternative hypothesis that is also consistent with our findings is that NPC1a promotes the absorption of dietary sterols in the midgut. To distinguish these possibilities we measured sterol levels in NPC1a null larvae relative to precise excision control larvae. If NPC1a functions in dietary sterol absorption we would expect to find a decrease in overall sterol content relative to controls, whereas a role for NPC1a in intracellular sterol trafficking would not be expected to affect sterol levels. Results of this analysis indicate that the loss of NPC1a function does not significantly affect the sterol content in mutant larvae (Figure 7). This result suggests that NPC1a is not required for the absorption of dietary sterols and supports a model in which NPC1a promotes the efficient intracellular trafficking of sterols in peripheral tissues following absorption.

Figure 7.—

The total cholesterol content of larvae is not detectably affected by mutations of NPC1a. The sterol content of NPC1a57A and NPC1a65A larvae was assessed using a commercial assay and found to be very similar. The NPC1a57A larvae may contain slightly higher levels of sterol, but this difference is not statistically significant. Three samples were measured in duplicate for each genotype; error bars display standard deviation.

NPC1a function is required in the ring gland:

Our results suggest that NPC1a functions to traffic sterols in Drosophila tissues and that a major consequence of the loss of NPC1a function is reduced ecdysone production in the larval ring gland. To test whether NPC1a function is primarily or exclusively required for proper trafficking of sterols in the ring gland we generated a UAS-responsive NPC1a cDNA transgene and investigated whether this transgene could rescue the NPC1a recessive lethal phenotype upon expression in the ring gland. Among a diverse set of Gal4 transgenes, only the ubiquitously expressed tubulin-Gal4 and ring gland expressed P0206–Gal4, 2-286–Gal4, and phantom–Gal4 drivers were able to rescue the NPC1a57A recessive lethal phenotype. No rescue was observed upon expression in other tissues including mesoderm (24B–GAL4 and twist–GAL4), post-mitotic neurons (elav–GAL4), midgut (mj33a–GAL4), and larval fat body (Lsp2–GAL4). The rescue of NPC1a57A larvae upon ectopic ring gland expression indicates that NPC1a expression in the ring gland is sufficient for viability, presumably because of recovery of the proper synthesis of ecdysone. The adults rescued upon ring gland expression appear to be morphologically normal, although they remain male-sterile. Experiments are currently under way to address the life span of these flies.

DISCUSSION

To explore the molecular mechanisms of sterol absorption and intracellular trafficking we have created loss-of-function mutations in the Drosophila NPC1a gene. We find that NPC1a null mutants die early in larval development and that hypomorphic alleles of the NPC1a gene confer defects in molting and metamorphosis. The lethal phase of NPC1a mutants can be extended by providing these mutants excess dietary cholesterol or ecdysone, a steroid hormone that triggers molting and metamorphosis, in the growth media. Additionally we find that the recessive lethal phenotype of NPC1a mutants is rescued by ectopically expressing wild-type NPC1a protein in the larval ring gland, the site of ecdysone synthesis. These results indicate that the Drosophila NPC1a gene is required for efficient utilization of sterols and proper steroid hormone metabolism and suggest that the core cellular function of NPC1a is conserved with the vertebrate NPC1 gene family. This functional conservation suggests that further genetic and molecular analyses of NPC1a function in Drosophila will facilitate our understanding of the molecular mechanisms of intracellular sterol trafficking and the precise functions of the NPC1 gene family.

Our work on NPC1a function allows us to revisit the question of whether NPC1a provides a function that is more similar to one or the other of the vertebrate NPC1 genes, NPC1 and NPC1L1. Several lines of evidence indicate that NPC1a more closely parallels the function of NPC1. First, the NPC1a gene, like NPC1, appears to be uniformly expressed, in contrast to NPC1L1 which appears to be expressed strongly only in the intestine (Altmann et al. 2004). Second, the finding that NPC1a expression in the ring gland but not midgut is able to rescue the NPC1a mutant phenotype indicates that this factor is not required for absorption of dietary sterols. Third, we find that NPC1a mutants are not depleted of sterols as would be predicted if NPC1a mutants are unable to absorb dietary sterols. In contrast, our recent work to explore the expression pattern and mutant phenotype of the Drosophila NPC1b gene suggests a close correspondence between the functions of NPC1b and NPC1L1 (our unpublished results). These experimental findings validate sequence features of the NPC1a and NPC1b proteins that suggest functional correspondence to NPC1 and NPC1L1, respectively.

Although our studies do not directly address the mechanistic nature of the sterol and ecdysone metabolic defects in NPC1a mutants, we believe that a model whereby NPC1a facilitates intracellular trafficking of sterols is the simplest interpretation of our data given knowledge of the functional roles of NPC1 homologs in other species. In the ring gland this trafficking defect may prevent delivery of sufficient amounts of sterol to sites of ecdysone synthesis, thus resulting in an ecdysone deficiency and the resulting molting and metamorphosis defects observed in NPC1a mutants. Our results further suggest that NPC1a is not absolutely required for intracellular sterol trafficking since the requirement for NPC1a can be bypassed by feeding flies excess cholesterol. Our ongoing studies of the NPC1b gene indicate that this factor is not responsible for the residual sterol trafficking in NPC1a mutants because NPC1b expression appears to be restricted to the midgut (data not shown). The residual intracellular sterol trafficking that remains in NPC1a mutants may explain why only the ring gland, the primary steroidogenic tissue in Drosophila, requires NPC1a function under normal growth conditions. Experiments are currently underway to directly test these hypotheses.

Our work on the Drosophila NPC1a gene correlates well with previously reported work in Caenorhabditis elegans. The worm genome also possesses two NPC1 homologs, ncr-1 and ncr-2 (Sym et al. 2000). The ncr-1; ncr-2 double mutants have reduced viability, gonadal migration defects, and constitutively form dauer larvae, a dormant life stage specialized for survival under harsh conditions (Sym et al. 2000; Li et al. 2004). The phenotypes of ncr-1; ncr-2 double mutants are similar to a class of ligand-binding domain mutants of daf-12, which encodes a nuclear steroid hormone receptor (Antebi et al. 1998; Antebi et al. 2000; Jia et al. 2002; Gerisch and Antebi 2004; Li et al. 2004). Li et al. (2004) recently demonstrated that feeding cholesterol to animals mutant for ncr-1 and ncr-2 is sufficient to suppress the dauer phenotype associated with the loss of both genes. Genetic epistasis analysis also indicates that the ncr genes function upstream of daf-12. These data suggest that the ncr genes deliver sterol precursors for the synthesis of the daf-12 steroid hormone ligand and closely parallel our results suggesting a perturbation in ecdysone synthesis in Drosophila NPC1a mutants.

While our work suggests a correspondence between the functions of NPC1 and NPC1a, we were unable to identify any substantial alterations in brain integrity in NPC1a hypomorphic animals even after aging individuals for as long as 40 days (data not shown). The lengthy life span and lack of detectable phenotypes associated with these adult escaper animals raise the possibility that NPC1a is required only for the efficient production of ecdysone in the larval ring gland. The widespread expression of NPC1a leaves open the possibility, however, that NPC1a is required for the integrity of neuronal and other tissues under conditions of limiting dietary sterol content. Furthermore, it remains possible that the sterol requirement of NPC1a mutants is epistatic to a primary defect in sphingolipid trafficking (Gondre-Lewis et al. 2003). These possibilities are currently under investigation.

In conclusion, we have developed a Drosophila model that may be useful for studying the function of NPC1, the mechanisms of intracellular sterol trafficking, and the effects of altered sterol trafficking on neuronal integrity. While significant progress has been made in our understanding of NPCD since the cloning of the NPC1 gene, including the determination of the membrane topology and subcellular distribution of NPC1 and the identification of at least some of the lipid mistrafficking and metabolic defects associated with loss of NPC1 function, many fundamental questions remain unanswered. In particular, little progress has been made in our understanding of the molecular mechanisms by which NPC1 promotes lipid trafficking, and we currently know nothing of the factors that regulate the expression, function, and distribution of NPC1. Moreover, the mechanism by which loss of NPC1 function results in neuronal death remains unclear. The use of a classical genetic approach in Drosophila to address these questions provides a powerful alternative to a purely biochemical or cell biological approach. The NPC1a allelic series described in this article, coupled with our finding that simple dietary factors can influence the associated phenotypes, should provide a useful starting point for this work.

Acknowledgments

The authors express their gratitude to Laurie Andrews and L. Katherine Moore for technical assistance in creating transgenic animals and identifying NPC1a alleles; Matt Scott and Xun Huang for discussions of their unpublished work; Kendal Broadie, Lynn Riddiford, Allen Shearn, James Truman, and the Bloomington stock center for fly stocks used in this work; the Electron Microscopy department employees, especially Franque Remington, at the Fred Hutchinson Cancer Research Center for help in preparing brain sections; and Xiaofeng Zhou for advice regarding the ecdysone feeding experiments. We also thank all members of the Pallanck lab for critical discussions and reading of the manuscript. Finally, special thanks go to the Jim Lambright Niemann-Pick Foundation and the Ara Parseghian Medical Research Foundation for their generous support of our work.

Note added in proof: The recent publication from Matt Scott and colleagues regarding their NPC1a Drosophila model (X. Huang, K. Suyama, J. Buchanan, A. J. Zhu and M. P. Scott, 2005, A Drosophila model of the Niemann-Pick type C lysosome storage disease: dnpc1a is required for molting and sterol homeostasis. Development 132: 5115–5124) is consistent with our data and supports our conclusions regarding NPC1a function in intracellular sterol trafficking.

Footnotes

  • Communicating editor: R. S. Hawley

  • Received June 7, 2005.
  • Accepted July 22, 2005.

References

View Abstract