Steroidogenic acute regulatory protein-binding protein cloned by a yeast two-hybrid system.

Steroidogenic acute regulatory (StAR) protein plays a key role in the transport of cholesterol from the outer mitochondrial membrane to the inner membrane. A StAR mutant protein lacking the first 62 amino acids (N-62 StAR protein) has been reported to be as effective as wild-type StAR protein. In the present study, we examined the mechanism by which StAR protein stimulates steroidogenesis. A Gal4-based yeast two-hybrid system was used to identify proteins interacting with N-62 StAR protein. Nine positive clones were obtained from screening 1 x 106 clones. The results of pull-down assays and mammalian two-hybrid assays confirmed interaction between N-62 StAR protein and the clone 4 translated product. The clone 4 translated product was named StAR-binding protein (SBP). We prepared an expression plasmid (pSBP) by inserting SBP cDNA into the pTarget vector. After cotransfection with the human cytochrome P450scc system, StAR expression vector, and pSBP, the amount of pregnenolone produced by COS-1 cells was increased. The amount of steroid hormones produced by steroidogenic cells subjected to small interfering RNA treatment was less than that produced by control cells. In conclusion, SBP binds StAR protein in cells and enhances the ability of StAR protein to promote syntheses of steroid hormones.

The first step in the biosynthesis of steroid hormones is the conversion of cholesterol into pregnenolone. The rate-limiting process is the transport of cholesterol from the outer mitochondrial membrane to the inner membrane, where cytochrome P450 side-chain cleavage (P450scc) 1 enzyme resides. Steroidogenic acute regulatory (StAR) protein plays a key role in the intra-mitochondrial movement of cholesterol (1). Mutations in the StAR gene cause congenital lipoid adrenal hyperplasia, a condition in which cholesterol accumulates in cytoplasmic lipid droplets, and adrenal and gonadal steroidogenesis is severely impaired (2,3). StAR gene knockout mice have the same phenotype as that of humans with congenital lipoid adrenal hyperplasia, that is cholesterol accumulation predominantly in the adrenal gland and markedly reduced steroid hormone secretion (4,5).
The tropic hormones ACTH, luteinizing hormone, and follicle-stimulating hormone stimulate steroid hormone production in the adrenal glands and gonads through a cAMP-dependent pathway. StAR gene expression increased rapidly in response to cAMP stimulation (6,7). Although the human StAR gene promoter lacks cAMP-responsive elements, steroidogenic factor-1 (8), CCAAT/enhancer-binding proteins, and GATA-4 confer cAMP-dependent StAR gene expression (9 -11). The factors SREBPs, Sp1, insulin-like growth factors, transforming growth factor-␤, tumor necrosis factor-␣, DAX-1, and SIK have also been shown to regulate StAR gene expressions (12)(13)(14)(15)(16)(17)(18)(19). Increased progesterone synthesis prior to an increase in the level of StAR mRNA expression in response to cAMP induction has been revealed by analysis of hyperacetylation in the StAR gene promoter (20). Although this increased steroidogenesis is thought to be the result of phosphorylation of the StAR protein by protein kinase A (8), the mechanism is still not clear.
Following synthesis, the 37-kDa StAR pre-protein is imported into mitochondria with subsequent cleavage of the mitochondria targeting sequence, yielding a 30-kDa mature StAR protein (21,22). StAR mutant protein lacking the first 62 amino acids (N-62 StAR protein), which contain the mitochondrial targeting sequence, has been reported to be as effective as wild-type StAR protein in stimulating steroidogenesis (23). This led to the conclusion that the C terminus of StAR protein encodes its biological function for steroidogenesis and that the StAR protein acts on the outer mitochondrial membrane to stimulate cholesterol translocation. The proposed site of action of StAR protein on the cytoplasmic face of the outer mitochondrial membrane raised questions about the mechanisms by which StAR protein trades with mitochondria and expresses its action in the cytoplasm. In the present study, we used a yeasttwo hybrid system to screen for proteins that interact with StAR protein and modulate its steroidogenic action.

MATERIALS AND METHODS
Plasmid Constructs-A plasmid expressing a GAL4-N-62-StAR fusion lacking 62 amino acid-terminal residues was constructed by inserting an EcoRI fragment, prepared by PCR using human StAR cDNA as a template, into a pACT2 vector, which has a GAL4-activating domain (GAD) (Clontech Laboratories, Inc., Palo Alto, CA). Plasmids expressing GAL4-StAR mutants (GAL4-R193X, GAL4-Q253X, and GAL4-frameshift) were also constructed by inserting an EcoRI fragment prepared by PCR from cDNA of human StAR mutants, which were previously constructed (2). A plasmid expressing a GAD-N-62-StAR fusion was * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ To whom correspondence should be addressed: Dept. of Biochemistry, Hokkaido University Graduate School of Medicine, Kita-ku, Kita 15, Nishi 7, Sapporo 060-8638, Japan. Tel.: 81-11-706-5047; Fax: 81-11-727-6006; E-mail: terusuga@med.hokudai.ac.jp. 1 The abbreviations used are: P450scc, P450 side-chain cleavage; StAR, steroidogenic acute regulatory; SBP, StAR-binding protein; DNA-BD, DNA-binding domain; GAD, GAL4-activating domain; AD, activation domain; RACE, rapid amplification of cDNA ends; siRNA, small interfering RNA; RT, reverse transcriptase; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; DMEM, Dulbecco's modified Eagle's medium; GFP, green fluorescent protein; RFP, red fluorescent protein; PBR, peripheral-type benzodiazepine receptor. also constructed by inserting an EcoRI fragment prepared by PCR and cloned into a pACT2 vector, which has a GAD (Clontech). Plasmid pVP16-StAR was constructed by inserting the EcoRI fragment from human N-62 StAR cDNA into the pVP16 vector, which has an activation domain (AD) derived from the VP16 protein of herpes simplex virus. The clone 4 translated product was named StAR-binding protein (SBP). We produced plasmid pM-SBP by inserting the EcoRI/BamHI fragment from SBP cDNA into the pM vector, which has a GAL4 DNA-binding domain (DNA-BD). Reverse combinations were also prepared; pVP16-SBP was constructed by inserting the EcoRI/BamHI fragment of SBP into the pVP16 vector. To produce plasmid pM-StAR, the EcoRI fragment from N-62 StAR was cloned into the pM vector. pG5luc (Promega Corp., Madison, WI) contains the chloramphenicol acetyltransferase gene or the luciferase gene as a reporter. From RACE results, we prepared an expression plasmid (pSBP) by inserting the EcoRI fragment of the entire coding region amplified by PCR from testis cDNA into the pTarget vector (Promega). A plasmid expressing an SBP-GFP fusion protein was constructed with the N-terminal green fluorescent protein (GFP) (pSBP-GFP) by inserting an EcoRI fragment, prepared by PCR using SBP cDNA as a template, into pEGFP-N1 (Clontech). A StAR-RFP fusion protein was also constructed with the N terminus of red fluorescent protein (pStAR-REP) by inserting an EcoRI fragment prepared by PCR from the full-length cDNA of human StAR cDNA into pDsRed2-N1 (Clontech). The plasmids were prepared for transfection studies using a Qiagen Maxiprep system (Qiagen, Hilden, Germany).
Cell Culture-Mouse Y-1 adrenal tumor cells, COS-1 cells, and human Hep G2 cells were obtained from RIKEN Cell Bank (Tsukuba, Japan). Human adrenocortical carcinoma H295R cells and mouse MA-10 Leydig cells were a gift from Dr. Mitsuhiro Okamoto, Osaka University Medical School (Osaka, Japan). Human MCF-7 breast cancer cells were obtained from ATCC (Manassas, VA). Human granulosalike tumor KGN cells were a gift from Dr. Yoshihiro Nishi, Graduate School of Medical Sciences, Kyushu University (24). The Y-1 cells and COS-1 cells were grown in 35-mm plastic dishes. The Y-1 cells and COS-1 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum and 50 g/ml gentamycin. The KGN cells were grown in DMEM/F-12 containing 10% fetal calf serum and 50 g/ml gentamycin. H295R cells were grown in DMEM/F-12 containing 2% ULTROSER G (BioSepra, Cergy-Pontoise, France) and 1% ITS Premix (BD Biosciences).
Pull-down Experiment-A plasmid expression clone was constructed by inserting an EcoRI fragment prepared by PCR and cloned into a pCI vector (Promega). Translated protein was synthesized in vitro using a T7 RNA polymerase-based TNT-coupled reticulocyte lysate system (Promega). A plasmid expressing a His-tagged CBD-N-62-StAR fusion protein lacking 62 amino acid-terminal residues was constructed by inserting an EcoRI fragment prepared by PCR using a human StAR cDNA as a template into a pET38b vector, which has a C-terminal His tag (Novagen, San Diego). CBD-N-62-StAR protein was expressed in bac-teria according to the manual of the manufacturer and was used for pull-down assays. His-tagged N-62-StAR fusion protein bound to His-Bind Resin (Novagen) was incubated for 3 h with 50 l in vitro translated [ 35 S]methionine-labeled translated clone in a total volume of 250 l of incubation buffer (50 mM potassium phosphate, pH 7.4, 150 mM KCl, 1 mM MgCl 2 , 10% glycerol, 0.1% Triton X-100). Resin was collected by microcentrifugation and washed three times. Washed beads were resuspended in 20 l of 2ϫ SDS sample buffer, heated for 5 min, and pelleted in a microcentrifuge, and the supernatant was subjected to SDS-PAGE and autoradiography.
Mammalian Two-hybrid Assay-Culture cells at 40 -60% confluence were transfected with plasmids with 0.1 g of the pG5luc reporter plasmid, 0.5 g of a pM-N-62-StAR protein (GAL4-StAR protein) expression plasmid, 0.5 g of a pVP16-StAR-binding protein (VP16-SBP) expression plasmid, and 0.1 g of pCH110 using FuGENE 6 (3 l/g DNA) (Roche Applied Science). The cells were cultured for 48 h after transfection and then harvested. Luciferase assays were performed using a Luciferase Assay System (Promega). The assay results were normalized to ␤-galactosidase activity to compensate for variation in transfection efficiency. Each experiment was repeated at least three times.
Northern Blot Analysis-Northern blots each containing 2 g of poly(A) ϩ RNA from various human tissues were purchased from Clontech and probed with 2.3-kb of human clone 4 cDNA and ␤-actin cDNA according to the supplier's protocol. Northern blotting and hybridization using human clone 4 were carried out as described previously (25).
RACE Analysis of the 5Ј-Site of cDNA of SBP-Marathon-Ready cDNA, 5Ј-stretched human testis cDNA (Clontech), was amplified using primers. The sense primer was Adapter Primer 1 (AP1, 5Ј-CCATCCTA-ATACGACTCACTATAGGGC), and the antisense primer was 5Ј-CAAACTGGAAAGAGCCTCCTCGTGAG-3Ј, which was 198 bp downstream from the translation start site identified in clone 4. PCR was performed according to the supplier's protocol. The PCR products were electrophoresed, and DNA fragments were cut from the gels. The fragments were ligated to PCR 2.1 vectors using a manual protocol (Invitrogen). PCR products were sequenced, and the sequences were compared with the human genome sequence.
Transfection and Pregnenolone Immunoassay-COS-1 cells were transfected with F2, a vector for the cytochrome P450 cholesterol sidechain cleavage system (called the F2 system) kindly provided by Dr. Walter L. Miller of the University of California, San Francisco (26), pStAR (pSPORT StAR cDNA) (2), and pSBP using FuGENE 6. The cells were incubated for 48 h following transfection. Some dishes were treated with 22(R)-hydroxycholesterol (1 g/ml) during the final 24 h of culture. Forty eight hours after transfection, the medium was collected for radioimmunoassay of pregnenolone. The assay results were normalized by serum pregnenolone concentrations produced by cultures with 22(R)-hydroxycholesterol to compensate for variation in transfection efficiency. Each experiment included triplicate cultures for each treatment group and was repeated at least three times.
Inhibition of SBP Expression by Small Interfering RNA (siRNA)-Cultures of sub-confluent (40 -50% confluent) H295R cells and KGN cells were plated so that 35-mm plastic dishes received equal numbers of cells. Endogenous SBP mRNA was targeted in cells with transfection by the addition of 19-nucleotide duplex (siRNA-SBP-I) and 21-nucleotide duplex (siRNA-SBP-II) (Dharmacon, Inc., Lafayette, CO). These duplex RNAs target 106 and 393 nucleotides downstream of the start codon of SBP, respectively. siRNA were constructed using the ribooligonucleotide pairs SBP-I and SBP-II with the following sequences: 5Ј-CGGGAUGUUUCCAGUGACAdTdT-3Ј and 5Ј-UGUCACUGGAAA-CAUCCCGdTdT-3Ј (SBP-I) and 5Ј-GAACUUGGAAGAGGGGAGGC-AdTdT-3Ј, 5Ј-UGCCUCCCCUCUUCCAAGUUCdTdT-3Ј (SBP-II). As a control for the specificity of these duplexes, we used a scramble ribooligonucleotide pair (siRNA-Scramble) with the following sequences: and 5Ј-CGAAUCCUA-CAAAGCGCGCdTdT-3Ј. The oligonucleotides were annealed according to the Dharmacon protocol. Three hundred pmol of each duplex was introduced into cells using 15 l of metafectene (Biontex Laboratories GmbH, Munich, Germany) as recommended by the manufacturer. Dishes for H295R cells were treated with 20 M trilostane (provided by Mochida Pharmaceutical Co., Ltd., Tokyo, Japan) 6 h after transfection to inhibit the enzyme activity of 3␤-hydroxysteroid dehydrogenase, which transforms pregnenolone into progesterone (27). 48 h after transfection, cells were collected for radioimmunoassay of steroid hormones. In some experiments, total RNA was extracted, and RT-PCR for SBP (27 cycles) and for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (24 cycles) as a control was performed using the following primers for GAPDH: sense, 5Ј-TGCCGTCTAGAAAAACCTGC-3Ј; antisense, 5Ј-ACCCTGTTGCTGTAGCCAAA-3Ј.
Fluorescence Microscopy-COS-1 cells were used for microscopic analysis. Cells were maintained in DMEM containing 10% fetal bovine serum. Before transfection, cells were seeded on coverslips in culture dishes. Cells were transfected using FuGENE 6 according to the manufacturer's instructions. After transfection, cells were incubated for 24 h. Transfected cells were fixed with 4% paraformaldehyde in phosphate-buffered saline for 30 min. After washing twice in phosphatebuffered saline, each coverslip was mounted onto a glass microscope slide. Observations were made with a fluorescence microscope equipped with a mercury lamp with excitation wavelengths of 450 -490 and 546 nm for GFP and DsRed, respectively (Axiophot, Carl Zeiss Inc., Oberkochen, Germany). Emission was used with a 515-565-nm bandpass filter for GFP and a 590-nm longpass filter for DsRed RFP. A digital camera (DXM 1200, Nikon, Tokyo, Japan) was attached to the microscope. The soft view system (ACT-1) provided by Nikon was used for image capture. Adobe Photoshop 5.0 (Adobe System Inc., San Jose, CA) was used for image processing.

Yeast Two-hybrid Screening of Proteins Interacting with
StAR Protein-A Gal4-based yeast two-hybrid system was used to identify proteins interacting with StAR protein. A plasmid expressing a GAL-StAR fusion protein, containing human N-62 StAR protein lacking 62 amino acid-terminal residues, was used to screen a human testis cDNA library in yeast strain CG-1945. Nine clones were obtained from screening 1 ϫ 10 6 clones. DNA sequence and data base analysis revealed that these clones could be divided into three groups: ␣-helix coiledcoil rod homologue (HCR, GenBank TM accession number NM 019052), rabaptin-5 (RAB5EP, accession number NM 004703), and nucleobindin 2 (NUCB2, accession number NM 005013) ( Table I). Both HCR and NUCB2 expressed the LacZ phenotype. From the expected intracellular localization of proteins encoded in the clones, we chose clone 4, which contained a 2.3-kb insert encoding a putative cytoplasmic protein, for analysis. Clone 4 contained an open reading frame of 1971 nucleotides encoding a 657-amino acid protein and a 62-nucleotide 3Ј-untranslated sequence that ended in a poly(A) ϩ tail preceded 21 nucleotides upstream by an AATAAA sequence. Other clones were not examined in this study.
Interactions of StAR Protein and Clone 4 in the Yeast Twohybrid Assay-To determine the interaction between clone 4 and StAR protein in vivo, a plasmid expressing a GAD-clone 4 fusion and a plasmid expressing GAL4-N-62-StAR and GAL4-StAR mutant fusions were examined. Transfections were also performed with the reverse combinations of clone 4 fused to GAL4 and N-62 StAR fused to GAD. The yeast transfected with StAR, and clone 4 hybrid vector expressed the LacZ phenotype, but the yeast transfected with a StAR mutant and clone 4 did not express the LacZ phenotype (Table II), confirming that N-62 StAR interacts with clone 4 in the yeast.
Direct Interaction between StAR Protein and Clone 4 in Vitro-Pull-down assays were performed to examine the direct interaction between clone 4 and StAR protein. In vitro translated clone 4 (50 l) was mixed with CBD or CBD-N-62-StAR fusion protein (100 ng) and then subjected to SDS-PAGE and autoradiography. These experiments revealed that the clone 4 translated protein interacted with CBD-N-62-StAR protein but not with CBD (Fig. 1). We therefore named the clone 4 protein (HCR) StAR-binding protein (SBP).
Interaction between SBP and StAR Protein in Cells-To confirm further interaction between SBP and StAR protein, we used two-hybrid assays in Y-1 cells, COS-1 cells, and MA-10 cells. The empty vectors, GAL4 DNA-BD (pM) and AD (pVP16), did not activate the reporter genes. Cotransfection of GAL4 DNA-BD (pM) and AD-StAR (pVP16-StAR), pM and AD-SBP (pVP16-SBP), GAL4 DNA-BD-SBP (pM-SBP) and AD (pVP16), and GAL4 DNA-BD-StAR (pM-StAR) and AD (pVP16) did not increase reporter activity. However, the GAL4DNA-BD-StAR (pM-StAR) and AD-SBP (pVP16-SBP) fusion proteins induced 100-fold greater activation of the promoter compared with the activity observed when the pM and pVP16 vectors were cotransfected into Y-1 cells. The switched domain constructs, pM-SBP and pVP16-StAR, also activated the reporter in Y-1 cells (Fig. 2A). Cotransfection with pM-StAR and pVP16-SBP

Analysis of interaction between clone 4 and a StAR
protein mutant by a yeast two-hybrid system Interaction between clone 4 and a StAR protein mutant was analyzed by using a yeast two-hybrid system. Yeast strain Y187 was transfected with GAL4-N-62-StAR hybrid expression vector and GAD-clone 4 hybrid expression vector. The yeast transfected with StAR protein and clone 4 hybrid vector expressed the LacZ phenotype, but the yeast transfected with the StAR protein mutant and clone 4 did not express the Lac phenotype. 2B) and MA-10 cells (Fig. 2C). These findings reflect the interaction between SBP and StAR protein in vivo. SBP mRNA Expression in Human Tissues-Northern blot analysis was performed to examine the expression of SBP. Northern blots that each contained 2 g of poly(A) ϩ RNA isolated from the indicated tissues were probed sequentially with SBP and ␤-actin cDNAs. Expression of the SBP gene was detected in all tissues examined. Its expression level was particularly high in the testis, where the sizes of transcripts were 2.4 and 3.8 kb (Fig. 3). To examine the expression of SBP in various cell lines, we performed RT-PCR using mRNA ex-tracted from Hep G2 cells, KGN cells, H295R cells, and MCF-7 cells. The expected amplification products (400 bp) were obtained in all of the cell lines examined (Fig. 4).
Identification of the Transcription Start Sites of SBP Gene Using Rapid Amplification of cDNA Ends (RACE)-For determination of the transcription start sites of the SBP gene, we employed a method based on anchored PCR using uncloned single-strand human testis cDNA with an anchor primer and a second primer that is specific to SBP. RACE products were electrophoresed, and DNA fragments were subcloned into PCR2 vectors for sequence analysis (Fig. 5A). Fifteen PCR

FIG. 3. Expressions of SBP mRNA in various human tissues.
Northern blots each containing 2 g of poly(A) ϩ RNA isolated from the indicated tissues were probed sequentially with SBP and ␤-actin cDNAs. The autoradiogram for the SBP probe was exposed for 48 h. The blots were exposed for 2 h for actin. products were sequenced, and the sequences were compared with the human SBP genome sequence (Fig. 5B). The SBP gene has two transcription start sites from alternative splicing and several minor transcription start sites. Although SBP has exon 1a and exon 1b, both 5Ј-sites interact with the same 3Ј-splice acceptors (Table III). The main transcription start sites are located 79 (exon 1a) and 430 bp (exon 1b) upstream of the translation start site (ATG) in exon 2. The first translation start site ATG in clone 4 is present in exon 4 in SBP in the testis. Clone 4 encodes an N-terminally truncated 656-amino acid region of the HCR protein (757 amino acids) (28) (Fig. 5C). From the RACE results, we prepared an expression plasmid (pSBP) by inserting the EcoRI fragment of the coding region of 757 amino acids amplified by PCR from testis cDNA into the pTarget vector (Promega).
Effect of SBP on Steroid Hormone Production-The SBP gene was found to be expressed in steroid hormone-producing cells, including H295R cells and KGN cells. To examine the effect of SBP on the action of StAR protein, COS-1 cells were cotransfected with F2, StAR, and SBP expression plasmids (pSBP). The amount of pregnenolone produced by COS-1 cells was increased by 138% compared with that produced by cells transfected by F2, pStAR, and an empty vector (Fig. 6).
siRNA Treatment Decreases the Production of Steroid Hormones-To examine further the role of SBP in the production of steroid hormones in cells, we performed an experiment to elim-inate SBP expression using siRNA (29) and assayed steroid hormones. Two target sequences of the SBP gene were selected for gene silencing. The effects of siRNAs were assayed by RT-PCR. SBP gene expression levels were reduced by both siRNA-SBP-I and siRNA-SBP-II. After treatment of cells transfected with the double-stranded RNAs, the amounts of pregnenolone produced by H295R cells (85 Ϯ 5.0 and 66 Ϯ 8.2 ng/dish when treated with siRNA-SBP-I and siRNA-SBP-II, respectively) were significantly less than (p Ͻ 0.05) that produced by H295R cells transfected with scramble siRNA (Fig. 7A). The amounts of pregnenolone produced by H295R cells treated with SBP-I and SBP-II siRNA were decreased by 56.5 and 37.5%, respectively, compared with the amount produced by H295R cells treated with scramble siRNA. The amounts of progesterone produced by KGN cells treated with SBP-I siRNA and SBP-II siRNA were also decreased by 71 and 55%, respectively, compared with the amount produced by KGN cells treated with scramble siRNA (Fig. 7B). Reducing the level of SBP gene expression by siRNA treatment of targeted SBP gene sequences resulted in a decrease in the production of steroid hormones.
Determination of SBP Localization in Cells-Fluorescent microscopic observations were performed to determine the subcellular localization of SBP in COS-1 cells that had been cotransfected with pSBP-GFP and pStAR-REP. Punctate cytoplasmic staining of SBP-GFR fusion protein was observed in cells. The size of punctate signals varied from one vesicle to another (Fig. 8A). StAR-REP fusion protein exhibited ovoidshaped signals in a reticulum pattern. These signal patterns are consistent with the characteristics of mitochondria (Fig.  8B). SBP-GFP fusion protein signals were partially overlapped with StAR-REP fusion protein signals (Fig. 8C). The findings indicate that SBP is localized in organelles, including endosomes and mitochondria, in the cytoplasm. DISCUSSION We screened a human testis cDNA library to identify proteins that interact with the StAR protein in the cytoplasm or outer mitochondrial membrane. Three clones, RAB5EP, NUCB2, and HCR, were identified by the screening. NUCB2, which functions as a DNA-binding protein, is a novel DNAbinding/EF-hand/leucine zipper (NEFA) (30). StAR homology domains are known as StAR-related lipid transfer (START) domains (31). START domains have been found in many kinds of proteins, including transcription factors (31). NUCB2 may be a transcription factor that interacts with the START domain that StAR-like transcription factors have. Rab5, which is a protein belonging to the family of Rab GTPases, regulates sequential transport steps along the endocytic and recycling pathway on early endosomes in cells (32). The secondary structure of SBP was predicted to consist of ␣-helical coils, which are features of the cytoskeleton, and it was therefore named ␣-helical coiled-coil rod homologue (HCR) (28). A BLAST search revealed that the amino acid sequence of SBP has little homology with those of known proteins. SBP did not interact with a StAR protein mutant that lacked the biologically active C terminus of the StAR protein (33). Thus, SBP protein interacts with the START domain of the C terminus of the StAR protein in cells. In mammalian two-hybrid assays, the interaction between StAR protein and SBP in cells transfected with plasmids of reverse orientation (pM-SBP and pVP16-N-62-StAR) was less than that in cells cotransfected with the expression vectors (pM-N-62-StAR and pVP16-SBP). These fusion protein structures may restrict interaction between SBP and StAR protein in cells.
SBP was found to be expressed in all tissues and in all cell lines examined, including steroid-producing cells. However, the expression level was not high except in the testis. After cotransfection with F2, pStAR, and pSBP, the activity of steroidogenesis was increased by 138%. The results of siRNA experiments also support the function of SBP, which have the effect of steroidogenesis in cells. Although the function of the larger transcription of SBP is not known, RACE results showed that SBP has several transcriptions start sites. Alternative transcription of the 5Ј-untranslated region can be regulated in a tissue by alternative usage of promoters. SBP may have a larger translated protein that acts as a pro-protein of SBP or may translate another protein that has a different function in the testis. Although many two-hybrid libraries contain partial rather than full-length cDNAs, small cDNA molecules may improve the sensitivity of the selection process (34). Clone 4 encodes an N-terminally truncated 656-amino acid region of SBP. Although clone 4 is supposed to be a partial cDNA, truncated SBP may function in the testis. Recently, StarD4, StarD5, and StarD6 proteins, which each contain a START domain, have been identified using cDNA microarrays. StarD4 and StarD5 are ubiquitously expressed, whereas StarD6 expression is limited to the testis. These proteins are thought to function in the intracellular shuttling of sterols or other lipids (35). It is possible that StAR protein and SBP not only function in steroidogenesis but also function together for the transport of cholesterol to regulate cellular metabolic processes, because StAR protein is present not only in Leydig cells but also in Sertoli cells in the testis (36).
Although steroidogenesis requires the continuous synthesis of new StAR proteins and is associated with the 37-kDa preprotein (37), an N-62-StAR recombinant protein stimulates the transfer of cholesterol from sterol-rich liposomes to mitochondria (38). Based on its physical characterization, StAR protein changes its conformation into a molten globule to associate  7. siRNA treatment decreased the production of steroid hormones in cells. Endogenous SBP gene expression was partially eliminated in H295R cells (A) and KGN cells (B) by transfection of a 19-nucleotide RNA duplex (siRNA-SBP-I) and 21-nucleotide RNA duplex (siRNA-SBP-II) directed to the SBP sequence. As a control, scramble sequences were introduced into the cells. The media were collected 48 h after treatment, and the amounts of pregnenolone (A) and progesterone were determined by radioimmunoassays. Values presented are means Ϯ S.E. of concentrations in the media in experiments in which each treatment group contained three replicate cultures. The effects of siRNAs were assayed by reversed RT-PCR for SBP (400 bp) and GAPDH gene (231 bp). * and **, significant difference in pregnenolone or progesterone production compared with that by cells transfected with scramble siRNA and that by cells treated with siRNA-SBP-I and siRNA-SBP-II; p Ͻ 0.05; p Ͻ 0.01. Each experiment was replicated on three separate occasions.
with the outer mitochondrial membrane and causes the transfer of cholesterol to the inner mitochondrial membrane (39 -41). These findings suggest that StAR protein functional sites of action are outside mitochondria and transfer cholesterol from the outer membrane to the inner membrane. The facility of N-62 StAR protein to transport cholesterol between membranes in a reconstitute system has been reported to be 1.8 molecules of cholesterol/molecule of the StAR protein (41). The facility in a reconstitute system is smaller than that in Y-1 cells, in which StAR protein is able to transfer up to 400 molecules of cholesterol (37). It is likely that another protein in the cytoplasm is required for effective cholesterol transport. The peripheral-type benzodiazepine receptor (PBR) has been shown to function in steroidogenesis, mitochondrial respiration, apoptosis, and cell proliferation (42,43). Because the PBR has been found in the outer mitochondrial membrane, where StAR protein is thought to act (42), StAR protein and SBP may work together in steroidogenesis with supporting functions of PBR. Alternatively, SBP may bind StAR protein and sustain StAR protein outside of the mitochondrial membrane to stimulate steroid hormone production, because the function of StAR protein on the outer mitochondrial membrane is regulated by its rate of import into mitochondria (44).
Lipid droplets in the cytoplasm, in which steroidogenic cholesterol is stored, are tightly attached to intermediate filaments. Mitochondria are also attached to intermediate filaments (45). The transport of cholesterol to mitochondria is associated with filaments and regulates steroidogenesis (46). MLN64, which resides in late endosomes and lysosomes, is proteolytically cleaved and includes START domains (47)(48)(49)(50). MLN64 has been reported to associate with the mobilization of low density lipoprotein-derived cholesterol through late endosomes to mitochondria. Observation under a confocal microscope revealed that MLN64 in late endosomal tubules and StAR protein in the microtubule-dependent mitochondrial matrix exist in close proximity but are not completely overlapped in cells (51). We have shown that signals of SBP, which is localized in organelles, including endosomes and mitochondria, in the cytoplasm are overlapped with StAR protein signals. SBP may interact with StAR protein in the cytoplasm. SBP has important functions with StAR protein in cells: one is steroidogenesis, as is well known, and the others are unknown. Further studies, including studies using mouse knockout models, are needed to determine the functions of SBP.