Targeted mutation of the MLN64 START domain causes only modest alterations in cellular sterol metabolism.

The StAR-related lipid transfer (START) domain, first identified in the steroidogenic acute regulatory protein (StAR), is involved in the intracellular trafficking of lipids. Sixteen mammalian START domain-containing proteins have been identified to date. StAR, a protein targeted to mitochondria, stimulates the movement of cholesterol from the outer to the inner mitochondrial membranes, where it is metabolized into pregnenolone in steroidogenic cells. MLN64, the START domain protein most closely related to StAR, is localized to late endosomes along with other proteins involved in sterol trafficking, including NPC1 and NPC2, where it has been postulated to participate in sterol distribution to intracellular membranes. To investigate the role of MLN64 in sterol metabolism, we created mice with a targeted mutation in the Mln64 START domain, expecting to find a phenotype similar to that in humans and mice lacking NPC1 or NPC2 (progressive neurodegenerative symptoms, free cholesterol accumulation in lysosomes). Unexpectedly, mice homozygous for the Mln64 mutant allele were viable, neurologically intact, and fertile. No significant alterations in plasma lipid levels, liver lipid content and distribution, and expression of genes involved in sterol metabolism were observed, except for an increase in sterol ester storage in mutant mice fed a high fat diet. Embryonic fibroblast cells transfected with the cholesterol side-chain cleavage system and primary cultures of granulosa cells from Mln64 mutant mice showed defects in sterol trafficking as reflected in reduced conversion of endogenous cholesterol to steroid hormones. These observations suggest that the Mln64 START domain is largely dispensable for sterol metabolism in mice.

The mechanisms by which hydrophobic lipids like sterols are moved within cells and targeted to specific membranes is not well understood. Recent findings indicate that a family of proteins related to the steroidogenic acute regulatory protein (StAR) 1 perform critical functions in moving lipids within cells (1)(2)(3)(4). StAR, the prototype of the family, promotes the translocation of cholesterol from the outer to the inner mitochondrial membrane and cholesterol side-chain cleavage enzyme in steroidogenic cells (1,5,6). Mutations in the StAR gene cause congenital lipoid adrenal hyperplasia, a cholesterol storage disorder in which synthesis of all gonadal and adrenal cortical steroid hormones is severely impaired; the cholesterol that is not efficiently moved into the mitochondria accumulates in cytoplasmic lipid droplets (7,8). The C terminus of StAR possesses sterol transfer activity, and it has been named the StARrelated lipid transfer (START) domain (2,6,9). It consists of a 210-amino acid residue sequence that forms a compact ␣/␤ structure, a helix-grip fold, with a hydrophobic tunnel that can accommodate a sterol molecule (2, 10 -12). START domains can bind sterol, facilitate the transfer of cholesterol from sterol-rich unilammelar liposomes to acceptor membranes, and stimulate steroidogenesis when expressed in cells co-expressing the cholesterol side-chain cleavage system or when added to isolated steroidogenic mitochondria (5,6,10).
Sixteen mammalian START domain proteins have been identified to date (2). Of these, StAR (StarD1) and MLN64 (StarD3) consist of one subfamily. MLN64 was originally discovered as a gene amplified in breast and ovarian cancers. It contains a C-terminal domain with 37% amino acid identity and 60% amino acid similarity to the C-terminal domain of StAR (13,14). The N terminus of MLN64 contains a leader sequence and four putative transmembrane domains, distinguishing it from StAR and suggesting that MLN64 functions in a different subcellular compartment. The MLN64 START domain was subsequently found to have StAR-like activity in that it could promote steroidogenesis in a model cell system (14). Like the StAR START domain, the MLN64 START domain binds cholesterol, stimulates the movement of free cholesterol from sterol-rich donor vesicles to acceptor membranes, and augments steroid synthesis when added to isolated mitochondria (10,15).
MLN64 was localized to late endosomes, a cellular compartment involved in the trafficking of cholesterol (15,16). The predicted topology of MLN64 in this compartment has the START domain facing the cytoplasm, being anchored to the vesicle wall by the four transmembrane domains, which positions the START domain to interact with apposing organelles or possibly other START domain proteins in the cytoplasm (16). MLN64 is incorporated into this dynamic tabulating vesicular system that also contains NPC1, a protein with a sterol-sensing domain (17,18), and NPC2 (also known as HE1), a sterolbinding protein (19). Niemann-Pick type C disease, a lysosomal cholesterol storage disorder, is caused by deficiency of either NPC1 or NPC2. The intracellular cholesterol trafficking abnormalities in these Neimann-Pick type C mutant cells or druginduced phenocopies of the disorder have been well studied (20 -25). The presence of a START domain protein in a subcellular compartment known to be involved in vesicular sterol trafficking by virtue of the presence of NPC1 and NPC2, raised the possibility that MLN64 participates in the intracellular distribution of sterols. This notion was supported by the finding that expression of a truncated MLN64 protein lacking the START domain coupled to green fluorescent protein (GFP) in COS-1 and Chinese hamster ovary cells caused accumulation of free cholesterol in lysosomes, mimicking the cellular sterol trafficking defect of Niemann-Pick type C disease (15). It was proposed that the sterol trafficking abnormality was the result of a dominant negative action of the truncated MLN64-GFP fusion protein.
The present studies were conducted to probe the function of the MLN64 START domain using a gene targeting strategy with the goal of identifying the physiological role of Mln64 in sterol dynamics in the mouse. Based on observations summarized above, it was anticipated that disruption of the Mln64 START domain would result in a phenotype similar to that of spontaneous mutations in the Npc1 gene in mice, including a progressive neurodegenerative disorder (ataxia, progressive motor deficits, and later death) and, at a cellular level, accumulation of free cholesterol in lysosomes (26,27). Humans lacking functional NPC1 or NPC2 have a similar phenotype. Here we show that, unexpectedly, mice lacking the Mln64 START domain are healthy and display only minimal disturbances in sterol dynamics, indicating that the Mln64 START domain is largely dispensable for intracellular cholesterol trafficking.

Targeted Mutation of the Mln64 START Domain
We determined that the murine Mln64 protein is encoded by a gene containing 14 exons and spanning ϳ9 kb (Fig. 1). We disrupted the START domain in the Mln64 gene in murine embryonic stem cells by replacing exons 10 -13 (GenBank TM accession number AL591390) with the fusion gene ␤geo, which consists of an internal ribosome entry site-lacZ-Neo r fusion gene (28). This targeting construct removes the majority of the Mln64 START domain (from amino acid residues 267-380). Embryonic stem cells derived from 129/Sv mice were transfected with the linearized ␤geo targeting vector, selected in medium supplemented with G-418, and analyzed by Southern blotting to identify correctly targeted clones.
For Southern blotting, genomic DNA was digested with EcoRV, and the blots were probed with a 0.87-kb cDNA containing genomic sequence downstream from the targeted genomic sequence. Three correctly targeted embryonic stem cell clones were used to generate chimeric mice, which were crossed with C57BL/6J females to obtain heterozygous mutants. Mice used in these studies were the offspring of crosses between the F1 and/or F2 generations (129/SvJ/C57BL/6J genetic background). Mice were genotyped by PCR. Three set of primers were used in the PCR. One set of primers corresponded to a sequence in the Neo gene (5Ј-CACATCTGTAGAGGTTTTACTTGC-3Ј) and a sequence in exon 14 of the Mln64 gene (5Ј-CTACCTTGATTGAACCCCA-GAAC-3Ј). Another set of primers corresponded to a sequence in the deleted region of exon 13 of the Mln64 gene (5Ј-GGGACTTTGTGAAT-GTCCGACG-3Ј) and the above noted primer for exon 14 of the Mln64 gene.

Assessment of Fertility and Fecundity
The mice used in these experiments were maintained in a University Laboratory Animal Resource facility and all animal protocols were approved by the University of Pennsylvania's Animal Care and Use Committee in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. To assess fertility and fecundity, littermate males (Ͼ6 weeks old) were placed in cages with two mature wild-type females for 1 month or more. Littermate females were caged with a wild-type male for a similar period. The number of mice achieving a pregnancy and the number of offspring from each mating set or pregnancy were recorded.

Dietary Manipulations
Mice were fed ad libitum Purina Mouse Chow 5001 or one of three experimental diets (Teklad Research Diets) containing control diet (TD7001; Teklad), or 0.5% cholesterol-enriched diet (TD88137; Teklad) or 15.75% fat and 1.25% cholesterol (high fat) diet (TD90221; Teklad). The high fat diet (TD90221; Teklad) contained (by weight) 75% Purina Mouse Chow, 15.75% fat, 1.25% cholesterol, and 0.5% sodium cholate. Mice over 6 weeks of age were housed individually and fed the specified diet for 1 week before harvesting tissue and blood. Mice were weighed and anesthetized with isoflurane (Forane; Baxter), and a syringe was used to draw blood from the heart. The mice were then killed by cervical dislocation, and organs were harvested, weighed, and flash-frozen in dry ice. Serum was generated from the blood following coagulation.
To assess ovarian steroidogenic function, six immature female mice (38 -45 days old) were injected intraperitoneal with 5 IU of pregnant mare's serum gonadotropin (Calbiochem). After 44 -48 h, all mice were injected with 5 IU of human chorionic gonadotropin (Sigma). Mice were anesthetized, and blood was collected by cardiac puncture. The mice were killed by cervical dislocation, and the ovaries and uterus were removed, cleaned, and weighed.

Measurement of Serum, Bile, and Hepatic Lipids
Cholesterol and triglyceride were quantified in serum and bile using cholesterol reagent and triglyceride reagent (2350 and TR22321, respectively (Thermo DMA). Liver tissue (1 g) was homogenized in 10 ml of ice-cold buffer (phosphate-buffered saline, pH 7.4, 1 mM EDTA, 1% Nonidet P-40, 1 g/ml aprotinin, 1 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride). Aliquots of liver homogenate were extracted with hexane/isopropyl alcohol (3:2, v/v) at 4°C, and the solvent was evaporated under nitrogen gas. Hepatic total and free cholesterol and triglyceride in the dried extract were quantified using cholesterol reagent, triglyceride reagent, and free cholesterol reagent (401-100P and TR22321, Thermo DMA; and 274-47109; Wako Chemicals, respectively). The level of cholesterol ester was calculated by subtracting the free cholesterol content from the total cholesterol content. Hepatic and bile phospholipids levels were quantified by a modification of the Bartlett procedure (29). Serum nonesterified free fatty acid and glucose were quantified using enzymatic colorimetric assays (Wako Chemicals and Stanbio Laboratories, respectively).

RNA Isolation, Real Time PCR Analysis, and Northern Blot Analysis
Total RNA was isolated using TRIzol reagent (Invitrogen). Real time PCR analysis was performed on an ABI PRISM 7900HT sequence detection system with target-specific probes and primers designed with Primer Express (PerkinElmer Life Sciences). Total RNA (5 g) was treated with DNase I (Promega) followed by cDNA synthesis using the Moloney murine leukemia virus reverse transcriptase (Promega) and oligo(dT) primer (Promega). The resulting cDNAs were diluted 1:10 in sterile water, and 1-l aliquots were used in the quantitative real time PCR. The primers used to quantitate mRNA are listed in Table I. In order to account for differences in starting material, glyceraldehyde-3phosphate dehydrogenase primers and probe reagents from Applied Biosystems were used as described by the manufacturer. In order to quantify differences, the samples were compared with standard curves for each gene and glyceraldehyde-3-phosphate dehydrogenase, and the average value for the triplicate determinations was used in all subsequent calculations.
For Northern blot studies, the total RNA samples (50 g/lane) were separated in 1.0% agarose gels containing formaldehyde and transferred to nylon membrane (Hybond N; Amersham Biosciences) according to the manufacturer's recommendations. Blots were subsequently hybridized with [␣-32 P]dCTP-labeled probes prepared by reverse transcriptase PCR using primers listed in Table I and wild-type mouse liver template. Hybridization was performed with Quikhyb (Stratagene). Membranes were exposed to X-AR films (Eastman Kodak Co.).

DNA Microarray Analysis
Target Preparation and Hybridization-All protocols were conducted as described in the Affymetrix GeneChip Expression Analysis technical manual. Briefly, 7 g of total RNA was converted to first strand cDNA using Superscript II reverse transcriptase primed by a poly(T) oligomer that incorporated the T7 promoter. Second strand cDNA synthesis was followed by in vitro transcription for linear amplification of each transcript and incorporation of biotinylated CTP and UTP. The cRNA products were fragmented to 200 nucleotides or less, heated at 99°C for 5 min, and hybridized for 16 h at 45°C to mouse MOE 430A Affymetrix microarrays. The microarrays were then washed at low (6ϫ SSPE) and high (100 mM MES, 0.1 M NaCl) stringency and stained with streptavidin-phycoerythrin. Fluorescence was amplified by adding biotinylated anti-streptavidin and an additional aliquot of streptavidin-phycoerythrin stain. A confocal scanner was used to collect fluorescence signal at 3-m resolution after excitation at 570 nm. The average signal from two sequential scans was calculated for each microarray feature.
Initial Data Analysis-Affymetrix Microarray Suite 5.0 was used to quantitate expression levels for targeted genes; default values provided by Affymetrix were applied to all analysis parameters. Border pixels were removed, and the average intensity of pixels within the 75th percentile was computed for each probe. The average of the lowest 2% of probe intensities occurring in each of 16 microarray sectors was set as background and subtracted from all features in that sector. Probe pairs were scored positive or negative for detection of the targeted sequence by comparing signals from the perfect match and mismatch probe features. The number of probe pairs meeting the default discrimination threshold ( ϭ 0.015) was used to assign a call of absent, present, or marginal for each assayed gene, and a p value was calculated to reflect confidence in the detection call. A weighted mean of probe fluorescence (corrected for nonspecific signal by subtracting the mismatch probe value) was calculated using the one-step Tukey's biweight estimate. This signal value, a relative measure of the expression level, was computed for each assayed gene. Global scaling was applied to allow comparison of gene signals across multiple microarrays; after exclusion of the highest and lowest 2%, the average total chip signal was calculated and used to determine what scaling factor was required to adjust the chip average to an arbitrary target of 150. All signal values from one microarray were then multiplied by the appropriate scaling factor (Genespring version 6.1 (Silicon Genetics Software)).

Western Blot Analysis
Equal amounts of liver protein (50 g/lane) were loaded onto 10% SDS-polyacrylamide gels and then transferred to Immobilon polyvinylidene difluoride membranes (Millipore Corp.). Western blot analysis was carried out using a rabbit anti-recMLN64 N-terminal (amino acid residues 1-52) and anti-C-terminal MLN64 antibodies recognizing either the START domain (amino acids 216 -440) or a peptide in the START domain (amino acid residues 370 -385) as previously described (14,15) For the detection of antibody protein complexes, the SuperSignal West Pico or Femto Kit (Pierce) was used according to the manufacturer's instructions.

Mouse Embryonic Fibroblasts Preparation
Mouse embryonic fibroblasts were prepared from embryonic day 14.5 embryos from the mated Mln64 Ϫ/ϩ littermates (30). DNA extracted from embryos was genotyped by PCR and Southern blot using standard techniques to identify wild-type and homozygous mutants.

Primary Granulosa Cell Culture
Mouse granulosa cells were collected from periovulatory follicles from mature female mice given a follicular stimulation protocol. Three wild type and three nullizygous female mice were injected intraperitoneal with 5 IU of pregnant mare's serum gonadotropin (Calbiochem). After 44 -48 h, the mice were killed by cervical dislocation, and the ovaries were carefully dissected away and placed into 4°C Dulbecco's phosphate-buffered saline (PBS; Invitrogen) containing 0.1% bovine serum albumin (Sigma). To release the granulosa cells from the folli- cles, the ovaries were repeatedly punctured with a 30-gauge needle. The expressed granulosa cells and minor contaminants (i.e. oocytes, red blood cells, and stromal cells) were collected and centrifuged at 770 ϫ g to pellet the cells in a 1.5-ml Eppendorf tube. The cells were cultured in Dulbecco's minimal essential medium/F-12 supplemented with 10% fetal calf serum and 100 g of penicillin-streptomycin/ml in 5% CO 2 , 95% air at 37°C. Dishes were coated with fibronectin from human plasma (Sigma) for at least 30 min prior to plating. Cells were plated at 100,000 cells/well in 12-well plates on day 0. On day 1, after washing with PBS, medium was changed to Dulbecco's minimal essential medium/F-12 containing 10% fetal calf serum and 100 g of penicillinstreptomycin/ml without or with 5 g/ml 22(R)-hydroxycholesterol, 1 mM 8-bromo-cyclic AMP (8-Br-cAMP), or ethanol (added in the same volume as for 22(R)-hydroxycholesterol). Each treatment group consisted of triplicate wells. After 24 h, medium was collected and frozen at Ϫ20°C until progesterone was assayed. Cells were scraped into ice-cold buffer (PBS, pH 7.4, 1 mM EDTA, 1% Nonidet P-40, 1 g/ml aprotinin, 1 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride). The protein concentrations were determined with the Bio-Rad protein assay reagents. These experiments were conducted on three separate occasions.
Progesterone secretion in the basal and 8-Br-cAMP-stimulated conditions was normalized to steroidogenic potential of the cells using progesterone production in the presence of 22(R)-hydroxycholesterol.

Histology and Transmission Electron Microscopy
Mouse livers were perfusion fixed through the heart with ice-cold 4% paraformaldehyde in PBS from homozygous mutant Mln64 mice and wild-type mice. The livers were removed and immersed in the same fixative for an additional 2 days at 4°C. Following perfusion fixation, 10-m frozen sections were cut from the whole left, right, and medial lobes of each liver. The sections were stained with Nile Red (0.1 g/ml PBS) for 15 min at room temperature, washed three times in PBS, and mounted in 90% glycerol in PBS containing 1 mg/ml n-propyl gallate as an antifade agent. Images were taken on a Zeiss LSM410 confocal microscope equipped with a Melles Griot Omnichrome krypton argon laser using a 488-nm excitation line and 515-540 band pass emission filter. Transmission electron microscopy was carried out on thin sections of liver as previously described (31) Immunocytochemistry Brain tissue was stained with filipin and an anti-GM3 ganglioside antibody as described by Zervas et al. (32). Mouse embryonic fibroblasts were grown on Nunc LabTek two-chambered glass slides. They were fixed at room temperature for 30 min in 3% paraformaldehyde in PBS, pH 7.4, washed 3 ϫ 5 min in PBS, and stained with Nile Red, mounted, and visualized as described above for liver sections. Cells were grown and fixed as above and stained overnight in 50 g/ml filipin in PBS, washed three times for 5 min each in PBS, and mounted as for the liver sections. They were imaged on a Zeiss LSM410 confocal microscope equipped with a Coherent Enterprise argon laser using a 364-nm excitation line and 400 -435 band pass emission filter.

Steroidogenic Activity in Transfected Embryonic Fibroblasts as a Probe of Cellular Sterol Trafficking
For assays of steroidogenic activity (15), embryonic fibroblasts were cultured in Dulbecco's minimal essential medium supplemented with 10% fetal calf serum and 50 g of gentamycin/ml in 5% CO 2 , 95% air at 37°C for cell propagation and plating. Cells were plated at 30,000   (14). Each experiment was performed with triplicate cultures for each cell preparation. After 24 h, medium was collected and frozen at Ϫ20°C until progesterone was assayed. The results were expressed as the ratio of progesterone production from endogenous sterol/progesterone production in the presence of 22(R)-hydroxycholesterol.

Statistical Analysis
Statistical tests were performed using the JMP 3.1.5 computer program (SAS Institute Inc.). Data were tested for heterogeneity of variance, and when detected, data were log-transformed. Tukey-Kramer mean separation, Wilcoxon, or Mann-Whitney tests were performed for comparison between the means with p Ͻ 0.05 taken as the level of significance.

RESULTS
Targeted Disruption of the Mln64 START Domain-We disrupted the START domain of the Mln64 gene in murine embryonic stem cells by replacing exons 10 -13 with the fusion gene, ␤geo (Fig. 1). To generate chimeras, embryonic stem cells carrying a mutant allele of the Mln64 gene ( Fig. 2A) were injected into blastocysts, which were then implanted into pseudopregnant mice. Mutant mice were produced from the chimeric offspring. Disruption of the Mln64 gene was confirmed by PCR analysis and Southern blotting. Northern blot analysis on liver RNA demonstrated the absence of the 3-kb Mln64 transcript when probed with a cDNA encompassing exons 10 -13 (Fig. 2B). This reflects deletion of 114 amino acids of the START domain, which prevents formation of the helix-grip fold and the hydrophobic tunnel that is essential for START domain function (10). When the same blot was reprobed with a probe representing exons 1-9, a truncated transcript of 2.4 kb was detected (Fig. 2C). This transcript represents exons 1-9 fused to exon 14, which encodes a 292-amino acid protein containing the N terminus of Mln64 and, as a result of a frameshift, a new C-terminal sequence (Fig. 2G). The presence of this mutant mRNA was verified by PCR and DNA sequence analysis. Western blot analysis of liver homogenates demonstrated the absence of intact 50-kDa MLN64 protein (Fig. 2E) but the presence of a 35-kDa truncated protein encoded by exons 1-9 fused to exon 14 (Fig. 2F).
Viability, Fertility, and Fecundity in Mln64 START Domain Mutant Mice-The proportion of wild-type (51 of 174; 29.3%), heterozygous (81 of 174; 46.6%), and nullizygous (36 of 174; 24.1%) offspring from mating of heterozygous males and females was not significantly different from the expected Mendelian pattern of inheritance ( 2 test, p Ͼ 0.05). The mice were viable and grew normally, and there was no evidence of neurological dysfunction over a more than 8-month period of observation. By comparison, mice homozygous for Npc1 mutations develop neurological symptoms by 42 days of age. Moreover, histological examination of the brains of 4-month-old homozygous mutant mice revealed no significant pathology, including in the cerebellum, which shows profound changes in mutant Npc1 mice. Immunocytochemical staining with filipin for free cholesterol storage and an antibody against ganglioside GM3, which characteristically accumulates in the neurons of Npc1 mutant mice (21), showed no differences from wild-type brain sections (Fig. 3). Both female and male heterozygous and homozygous mutant mice were fertile and had similar litter sizes to wild-type mice (Table II).
Serum Cholesterol, Triglyceride, Nonesterified Free Fatty Acid, Glucose, and Steroid Levels-There were no significant differences in serum cholesterol, triglycerides, nonesterified free fatty acid, and glucose levels among wild-type and homozygous mutant mice fed various diets (Fig. 4). Testosterone and corticosterone levels in homozygous mutant male mice fed a regular diet were not statistically different from those found in wild-type mice, although the testosterone levels tended to be lower (Fig. 4). Immature female homozygous mutant mice (n ϭ 3) and wild-type mice (n ϭ 3) stimulated with pregnant mare's serum gonadotropin-human chorionic gonadotropin had similar mean ovarian weights/body weights Ϯ S.E. Hepatic Free Cholesterol, Cholesterol Ester, Triglyceride, and Phospholipid Levels-Hepatic free cholesterol, cholesterol ester, triglyceride, and phospholipid levels were similar in homozygous mutant male and wild-type mice fed a control diet or a 0.5% cholesterol-enriched diet for 1 week (Fig. 5). However, there was significantly greater cholesterol ester storage in homozygous mutant mice fed a high fat diet (Fig. 5).
We measured the bile content of cholesterol and phospholipids and the molar ratio of cholesterol/phospholipids in four wild-type and four nullizygous male mice fed the high fat diet. Although the cholesterol content and molar ratio of cholesterol/ phospholipid were lower in nullizygous mice, the values were not significantly different from wild-type bile (p Ͼ 0.05): wildtype bile cholesterol (means Ϯ S.E.), 0.60 Ϯ 0.20 mM; nullizygous bile cholesterol, 0.51 Ϯ 0.01 mM; wild-type bile phospholipids 1.41 Ϯ 0.40 mM; nullizygous bile phospholipids, 1.77 Ϯ 0.20 mM; wild-type cholesterol/phospholipid molar ratio 0.44 Ϯ 0.12; nullizygous cholesterol phosopholipid molar ratio, 0.28 Ϯ 0.05.

Histochemical Analysis of Hepatic Neutral Lipid Distribution and Transmission Electron Microscopy-Nile
Red staining for neutral lipid and filipin staining for free cholesterol revealed no differences in lipid distribution in hepatocytes from male homozygous mutant and wild-type mice fed a regular chow diet (Fig. 6). Transmission electron microscopy revealed no significant differences in ultrastructure between wild-type and Mln64 START domain-deficient mice fed a regular chow diet or cholesterol-enriched diet for 7 days (Fig. 7). Notably, endosomes and lysosome-like structures appeared to be similar in size and distribution, and enlarged lysosomal structures, characteristic of the Niemann-Pick type C cells, were absent from the Mln64 START domain-deficient mice. These observations collectively rule out a sterol trafficking defect similar to that observed in Niemann-Pick type C disease in mice eating a normal chow diet.
Northern Blots and Quantitative Real Time PCR Analysis of Transcripts-We analyzed levels of mRNAs for a variety of sterol-sensitive transcripts (HMG-CoA synthase transcript suppressed with cholesterol accumulation and Cyp7a, Abcg5, and Abcg8 transcripts increased with cholesterol accumulation). These were not significantly different in homozygous mutant males and wild-type mice fed a standard chow diet or 0.5% cholesterol-enriched or high fat diets for 1 week compared with wild-type males fed similar diets (Fig. 8). The HMG-CoA synthase mRNA levels tended to be lower in the mutant mouse liver, and the Cyp7a mRNA levels were slightly higher. Taking a ratio of the HMG-CoA synthase mRNA levels to Cyp7a mRNA levels, there was a statistically significantly lower ratio for the mutant mice compared with wild-type mice fed a regular chow diet (0.59 Ϯ 0.13 for wild-type liver versus 1.56 Ϯ 0.33 for Mln64 START domain-deficient mice p Ͻ 0.03). As expected, HMG-CoA synthase mRNA was reduced in animals fed cholesterol-enriched and high fat diets. The cholesterol-enriched diet increased Cyp7a, Abcg5, and Abcg8 mRNA levels in both wild-type and homozygous mutant livers, whereas the high fat diet suppressed expression, again as anticipated (33).
Expression of other genes implicated in intracellular sterol transport (Npc1 and Npc2) as well as two other START domain protein transcripts (StarD4 and StarD5) and a transcript encoding a protein homologous to the MLN64 N-terminal domain (MENTHO) were not altered in Mln64 START domain-deficient mice fed a normal chow diet as assessed by Northern blotting (Fig. 9).
Global Transcript Profiling of Liver-Affymetrix genechips were used to analyze the profile of transcripts in RNA extracted from the livers of three different homozygous mutant and three different wild-type male mice. Of the 9165 transcripts detectable in this analysis, the only transcript showing a Ͼ2-fold and statistically significant difference by the Welch t test (p Ͻ 0.01) was Map3k7. These observations corroborate the lack of significant differences in the abundance of mRNAs related to sterol metabolism. However, quantitative real-time PCR analysis of 10 different samples from wild-type and mutant mouse livers did not reveal a significant difference on Map3k7 mRNA levels between wild-type and mutant mice, indicating that the microarray analysis yielded a false positive result.
Analysis of Embryonic Fibroblasts-Cytochemical staining of cultured embryonic fibroblasts with Nile Red for neutral lipids and filipin for free cholesterol demonstrated no significant differences in cells cultured in medium enriched with human low density lipoprotein (Fig. 10). However, treatment of the wild-type embryonic fibroblasts with progesterone (50 g/ ml) caused the formation of filipin-laden lysosomes, demonstrating that a pharmacological manipulation known to produce a phenocopy of the Niemann-Pick type C sterol trafficking defect could be detected (data not shown) (23). When cells were stained with Nile Red to analyze the distribution of neutral lipids, no differences were observed between wild-type and knockout cells (Fig. 10).
Mln64 START Domain-deficient Cells Reveal a Defect in Cholesterol Trafficking for Steroidogenesis-Our previous stud-ies suggested that MLN64 might participate in the movement of cholesterol to mitochondria, a process that is essential for steroidogenesis (15). We therefore evaluated this process using embryonic fibroblasts prepared from wild-type and homozygous mutant embryos. Using five different embryonic fibroblast cultures, we found a significantly reduced production of progesterone from endogenous substrate, suggesting the existence of a sterol trafficking abnormality related to the movement of cholesterol to mitochondria (Table III).
To further explore the process of steroidogenesis, we examined progesterone secretion by primary cultures of granulosa cells collected from the ovaries of gonadotropin-primed wildtype and nullizygous mice in the basal state and when stimulated with 8-Br-cAMP. The maximal steroidogenic activity of the cultured cells was assessed by incubating them with 22hydroxycholesterol, a more soluble substrate that moves into mitochondria without the need for StAR and thus bypasses the rate-limiting step in steroidogenesis and allows expression of maximal steroidogenic activity. This value was used to normalize steroid synthesis from endogenous cholesterol to the overall steroidogenic capacity of the cells. Although granulosa cells from nullizygous mice produced significantly more progesterone in the basal state than wild-type cells (possibly reflecting the impact of compensatory responses in the nullizygous animals to endogenous gonadotropins), when they were stimulated with 8-Br-cAMP, the nullizygous cells had a significantly smaller increase in steroidogenesis compared with wild-type cells (Table IV), indicating that relative to their total steroidogenic capacity, the nullizygous cells were less efficient in converting endogenous substrate into progesterone than wild-type granulosa cells. These observations parallel the results obtained with transfected embryonic fibroblasts. DISCUSSION START domain proteins have been implicated from genetic studies in the movement of substrate cholesterol to mitochondria of steroidogenic cells (7,8) and in the intracellular trafficking of ceramide (3). The present studies were undertaken to use targeted mutation in the mouse to probe the function of the Mln64 START domain. Based on the close homology between StAR and MLN64, it was anticipated that MLN64 has a role in intracellular cholesterol movement. Moreover, the localization of MLN64 to late endosomes and its colocalization with other proteins involved in vesicular sterol trafficking (NPC1 and NPC2) led us to postulate that loss of the START domain would result in a phenotype similar to the Npc1 mutant mouse and cholesterol trafficking defect not unlike that of Niemann-Pick type C disease, the accumulation of free cholesterol in late endosomes/lysosomes. However, mice homozygous for a mutation disrupting the Mln64 START domain were generally healthy and displayed only modest abnormalities in cholesterol dynamics and no evidence for accumulation of cholesterol in the liver, a characteristic of the Npc1 mutant mouse, particularly accumulation of free cholesterol in the late endosomes/ lysosomes and the attendant compensatory alterations in expression of genes involved in cholesterol synthesis and metabolism that are characteristic of mice lacking functional NPC1. Dietary manipulations, as performed in our study, exacerbate the hepatic sterol metabolism defects in Npc1 mutant mice. If anything, the opposite was observed in the Mln64 mutant mice, with HMG-CoA synthase mRNA levels tending to be lower. Moreover, when Mln64 mutant mice were fed a high fat diet, cholesterol ester, not free cholesterol, accumulated in the liver.
The mechanism underlying the build up of sterol esters is not yet known, although the recent discovery that StAR binds to  TABLE IV Normalized progesterone production by primary cultures of ovarian granulosa cells Mouse granulosa cells were collected from periovulatory follicles from mature female mice given a follicular stimulation protocol. On day 1, after washing with PBS, medium was changed to Dulbecco's minimum essential medium/F-12 containing 10% fetal calf serum and 100 g of penicillin-streptomycin/ml without or with 5 g/ml 22(R)-hydroxycholesterol, 1 mM 8-Br-cAMP, or ethanol (added in the same volume as for 22(R)-hydroxycholesterol). Each treatment group consisted of triplicate wells. After 24 h, medium was collected for progesterone assays, and cell protein was assayed. Progesterone secretion in the basal and 8-Br-cAMP-stimulated conditions was normalized to steroidogenic potential of the cells using progesterone production in the presence of 22(R)hydroxycholesterol. The -fold response to 8-Br-cAMP stimulation was calculated as the ratio of the normalized 8-Br-cAMP to basal progesterone production. The results presented are means Ϯ S.E. from three separate experiments, each conducted with triplicate wells for each treatment. hormone-sensitive lipase raises the possibility that in the absence of the MLN64 START domain, mobilization of sterol esters by this sterol esterase is impaired (34). It is interesting to note that using 6-dansyl-cholesterol as a probe, Wiegand et al. (35) found that the probe accumulated in larger cytoplasmic lipid droplets in NPC1 mutant cells, implicating the NPC1 compartment in the turnover of lipids in cytoplasmic droplets (35). Fibroblasts derived from mutant embryos and granulosa cells stimulated with 8-Br-cAMP showed deficits in the utilization of endogenous cholesterol for steroidogenesis, which presumably reflects an abnormality in sterol trafficking of substrate to mitochondria. A similar type of defect was observed in COS-1 cells transfected with a truncated MLN64-GFP fusion protein, which was presumed to disrupt MLN64 function by acting as a dominant negative (14). In the latter experiments, the truncated MLN64-GFP fusion protein also caused accumulation of free cholesterol in lysosomes, which was not observed in the mutant embryonic fibroblasts. This discrepancy suggests that the truncated MLN64-GFP protein may have had effects on other cellular functions in addition to any action directly related to MLN64. Indeed, transfection of cells with MENTHO, which encodes a protein related to the MLN64 N-terminal domain, caused an abnormality in vesicular trafficking (36). Given the similarities between the truncated MLN64-GFP fusion and MENTHO, it is possible that the results of previous studies reflected a combination of effects on MENTHO and MLN64.
The deficit in utilization of endogenous cholesterol by mutant embryonic fibroblasts and cultured granulosa cells for steroidogenesis appears to be accentuated in vitro, since there were no significant differences in serum steroid levels in mutant mice, although testosterone levels tended to be lower in males, and progesterone levels tended to be lower in gonadotropinstimulated females. This may be a reflection of differential use of substrate pools for steroidogenesis by steroidogenic cells in vivo. Our previous observations of a close association of the tubulo-vesicular compartment containing MLN64 with mitochondria suggest that the transfer of cholesterol out of the late endosomes to the mitochondrial outer membrane, a necessary step in the acquisition of potential steroidogenic substrate, is impaired in the absence of a functional Mln64 START domain.
The lack of a profound phenotype in sterol dynamics in the Mln64 START domain mutant mice may be the result of redundant actions of other START domain proteins. Having said this, we cannot rule out the possibility that a dietary manipulation not explored in our experiments might bring out a more profound alteration in lipid metabolism in the nullizygous mice. It should be noted that targeted mutation in another START domain protein, phosphatidylcholine transfer protein (StarD2), did not result in a demonstrable hepatic phenotype in phospholipid metabolism (37). Alternatively, it may be that the MLN64 N-terminal domain, which is retained in our mutants, plays an unexpected role in sterol trafficking. If this is indeed the case, the concept of START domain action based on structural studies and site-directed mutagenesis would require serious reevaluation. It is also possible that despite its close structural relationship to StAR, MLN64 functions in the trafficking of lipids other than cholesterol.
In conclusion, mutation in the START domain of MLN64, a protein closely related to StAR and thought to play a role in intracellular sterol trafficking, does not markedly impair cholesterol dynamics but may have a minor facilitating role in moving cholesterol to the steroidogenic machinery. These findings suggest that proteins with redundant functions exist that subserve the role of the Mln64 START domain and that the Mln64 START domain is largely dispensable with respect to cholesterol dynamics.