Complex Role of the Mitochondrial Targeting Signal in the Function of Steroidogenic Acute Regulatory Protein Revealed by Bacterial Artificial Chromosome Transgenesis in Vivo

Abstract

The steroidogenic acute regulatory protein (StAR) stimulates the regulated production of steroid hormones in the adrenal cortex and gonads by facilitating the delivery of cholesterol to the inner mitochondrial membrane. To explore key aspects of StAR function within bona fide steroidogenic cells, we used a transgenic mouse model to explore the function of StAR proteins in vivo. We first validated this transgenic bacterial artificial chromosome reconstitution system by targeting enhanced green fluorescent protein to steroidogenic cells of the adrenal cortex and gonads. Thereafter, we targeted expression of either wild-type StAR (WT-StAR) or a mutated StAR protein lacking the mitochondrial targeting signal (N47-StAR). In the context of mice homozygous for a StAR knockout allele (StAR^−/−), all StAR activity derived from the StAR transgenes, allowing us to examine the function of the proteins that they encode. The WT-StAR transgene consistently restored viability and steroidogenic function to StAR^−/− mice. Although the N47-StAR protein was reportedly active in transfected COS cells and mitochondrial reconstitution experiments, the N47-StAR transgene rescued viability in only 40% of StAR^−/− mice. Analysis of lipid deposits in the primary steroidogenic tissues revealed a hierarchy of StAR function provided by N47-StAR: florid lipid deposits were seen in the adrenal cortex and ovarian theca region, with milder deposits in the Leydig cells. Our results confirm the ability of StAR lacking its mitochondrial targeting signal to perform some essential functions in vivo but also demonstrate important functional defects that differ from in vitro studies obtained in nonsteroidogenic cells.

PITUITARY TROPHIC HORMONES, ACTH for the adrenal cortex and FSH and LH for the gonad, are the major physiological regulators of steroidogenesis. Temporally, trophic hormone stimulation of steroidogenesis can be divided into two phases. The acute phase occurs within seconds to minutes and involves enhanced mobilization and delivery of cholesterol to the cholesterol side-chain cleavage enzyme (Cyp11a1) inside the mitochondria (1); the chronic phase, in contrast, occurs over hours to days and reflects increased transcription of many constituents of the steroidogenic pathway (2). The steroidogenic acute regulatory protein (StAR) is essential for the acute response, facilitating cholesterol delivery to the inner mitochondrial membrane where Cyp11a1 resides (3,4,5,6,7). The pivotal role of StAR in steroid hormone synthesis was dramatically revealed when it was shown that StAR mutations in humans cause congenital lipoid adrenal hyperplasia (lipoid CAH; MIM 201710), an autosomal recessive disorder characterized by impaired steroid hormone synthesis and abundant lipid deposits in the adrenal cortex and gonads (8,9,10). Subsequent studies showed that knockout (KO) mice lacking StAR (Star^−/−) had a very similar phenotype to that seen in humans with lipoid CAH (11,12,13).

Despite the unequivocal evidence for the essential role of StAR in normal steroidogenesis, several puzzling features remain. First, nonsteroidogenic cells transfected with the other components needed to convert cholesterol to pregnenolone performed StAR-independent steroidogenesis at approximately 10% of the level seen when StAR was present (14,15); the basis for this StAR-independent steroidogenesis is unknown. In addition, StAR-mediated cholesterol delivery is strongly regulated by trophic hormones that enhance both StAR transcription and phosphorylation in response to cAMP (4); this process is markedly blocked by the protein synthesis inhibitor cycloheximide (16), suggesting that the only newly synthesized StAR is active in cholesterol transfer. StAR then undergoes mitochondrial entry and processing through the mitochondrial targeting signal and protein complex, which involves protein traffic across the membranes, and the mitochondrial processing peptidase (4,6); pregnenolone synthesis in mouse adrenocortical Y1 cells or Leydig MA-10 cells was markedly inhibited by some agents that disrupt the mitochondria hydrogen ion gradient (CCCP) or that inhibit mitochondrial electron transport (antimycin A), ATP synthesis (oligomycin), pH gradient (nigericin), or the mitochondrial processing peptidase inhibitor (orthophenanthrolene) (17,18); these results suggest that both mitochondrial potential and intra-mitochondrial processing of StAR are essential for its activity in cholesterol transport. Kinetic studies in mouse Y1 adrenocortical cells have suggested that each molecule of newly synthesized StAR is responsible for the mitochondrial entry of 400 molecules of cholesterol per minute (17). However, there has been considerable debate about the mechanism by which StAR facilitates steroidogenesis (5,6,7). The crystal structure of the StAR-related lipid transfer domain of MLN64, which has considerable sequence homology with StAR, revealed a hydrophobic tunnel large enough to accommodate only one molecule of cholesterol (19). Based on this finding, the authors proposed that StAR acts in the space between the outer and inner mitochondrial membranes to shuttle cholesterol across the aqueous environment via this tunnel.

An alternative line of investigation proposed that StAR acts exclusively at the outer mitochondrial membrane (7), where effects of protonated phospholipid head groups from the membrane induce the protein to assume a molten globule state that contributes to cholesterol delivery (20,21,22). Important components of this model include the finding that StAR tethered to the outer mitochondrial membrane was fully active in stimulating cholesterol delivery and that a more rapid rate of StAR entry into the mitochondria was associated with less efficient steroidogenesis (23). In addition, these investigators found that expression of a truncated StAR protein lacking its mitochondrial targeting signal stimulated steroidogenesis at protein levels comparable to those seen normally in steroidogenic cells, suggesting a physiological rather than pharmacological action of this truncated StAR protein.

An important tenet of the latter model is that recombinant expression of truncated StAR proteins in nonsteroidogenic cells provides valid insights into events in authentic steroidogenic cells. To explore this model further, we used bacterial artificial chromosome (BAC) transgenesis to express either wild-type StAR (WT-StAR) or StAR lacking the amino-terminal mitochondrial targeting signal (N47-StAR) in the steroidogenic cells of the adrenal cortex and gonads. Despite achieving expression levels comparable to that of endogenous StAR, we demonstrate tissue-specific differences in the function of the WT- and N47-StAR proteins in vivo.

RESULTS

Targeting Expression of StAR/Enhanced Green Fluorescent Protein (eGFP) BAC Transgene to Steroidogenic Cells that Express StAR

To permit us to explore structure-function aspects of StAR within steroidogenic cells, we sought to develop a system to drive StAR expression within steroidogenic cells of the adrenal cortex and gonads in vivo. BAC transgenesis is a powerful tool for transgenic expression that is copy-number dependent and position independent (24,25). To this end, we identified a StAR BAC clone that included 47 kb upstream of the transcription initiation codon, the entire Star structural gene, and 62 kb downstream of the termination codon of the gene (Fig. 1, A and B). To examine the ability of these genomic sequences to target gene expression correctly, we first inserted a cassette encoding eGFP and a polyadenylation signal from bovine GH (bGH) into the BAC clone at the normal initiator methionine (Fig. 1C). After BAC modification by homologous recombination in Escherichia coli, independent transgenic lines with supercoiled BAC DNA were generated and analyzed as described in Materials and Methods.As shown in Fig. 2, the StAR/eGFP transgene was expressed at high levels throughout the adrenal cortex, in Leydig cells in the testes, and in theca and interstitial cells in the ovary. Immunohistochemical analysis with either polyclonal antibodies against mouse StAR protein or monoclonal antibodies against eGFP protein confirmed the expression of eGFP in steroidogenic cells of the adrenal glands and gonads (data not shown). We did not observe fluorescence or immunological activity of the eGFP in other tissues, including brain, thymus, heart, and kidney (data not shown).

Figure 1.

Strategy for Generating the StAR/eGFP and N47-StAR BAC Transgenes

A, Structure of the WT-StAR BAC clone. The mouse Star gene was cloned into the HindIII and BamHI sites of pBeloBAC11. The 121-kb StAR BAC clone contains the Star structural gene, 47 kb of 5′-flanking region, and 62 kb of 3′-flanking region. B, Restriction map of the WT-StAR BAC clone. BAC DNA was digested with eight restriction enzymes (1–8) and separated by pulsed-field gel electrophoresis. Southern blot analysis with a StAR probe revealed the unarranged Star locus. Digestion with NdeI distinguished between N47-StAR and WT-StAR BACs. C, The StAR/eGFP BAC contains the eGFP cDNA and polyadenylation (polyA) signal from bGH inserted at the StAR initiator methionine. D, The N47-StAR BAC encodes a truncated StAR protein with deletion of the first 47 amino acids.

Figure 2.

Expression of the StAR/eGFP Transgene in Steroidogenic Tissues of Adult Mice

Tissues were harvested from StAR/eGFP transgenic mice at 8 wk of age, and expression of eGFP was determined by fluorescence photomicrography of adrenal gland (×100 magnification), testis (×100 magnification), and ovary (×40 magnification), respectively. AF, Antral follicle; CL, corpus luteum; ST, seminiferous tubule.

Targeting Expression of WT-StAR or N47-StAR BAC Transgene to Steroidogenic Cells in Adrenal Glands and Gonads of StAR KO Mice

Having established the utility of the StAR BAC to direct appropriate transgenic expression, we next compared the function of WT-StAR and N47-StAR. The latter construct was prepared by converting three adenine residues that comprise potential initiator ATGs to thymidine residues and by modifying a thymidine residue three nucleotides 5′ of the new initiator methionine to an adenine residue to better match the Kozak consensus (Fig. 1D). The WT- and N47-StAR BACs were then used to generate independent transgenic mouse lines as described in Materials and Methods.

Three independent lines with comparable copy numbers for each transgene were selected for further study, designated WTTg(a), WTTg(b), WTTg(c), N47Tg(a), N47Tg(b), and N47Tg(c), respectively (Fig. 3A). At both the mRNA and protein levels, the WTTg and N47Tg lines directed comparable expression of StAR in the primary steroidogenic tissues of newborn (Fig. 3B) and adult (Fig. 3C) mice, although lower expression of N47-StAR protein was found in adult testes. In a substantial fraction of these N47-StAR transgenic mice with homozygosity for the StAR KO allele, the testes were not fully descended (see below), possibly confounding analysis of their function. In addition to basal expression at levels comparable to endogenous StAR, the BAC-driven StAR expression responded appropriately to trophic hormones that normally regulate StAR expression and the acute induction of steroidogenesis (Fig. 3C). Collectively, these studies indicate that the BAC transgenes provide comparable amounts of basal and hormone-induced StAR expression in the desired tissues, thereby enabling us to examine the relative efficiencies of the WT- and N47-StAR proteins in steroidogenesis. Based on the comparable levels of StAR expression at both the mRNA and protein levels (Fig. 3) and the comparable phenotypes conferred by each of three independently derived StAR transgenes in mice homozygous for the StAR KO allele (data not shown), we focused further studies on the WTTg(a) and N47Tg(a) BAC transgenic lines, as described below.

Figure 3.

Generation and Characterization of Star^−/− Mice Carrying the WT-StAR (Star^−/−WTTg) or N47-StAR (Star^−/−N47Tg) Transgene

The WT-StAR and N47-StAR BAC transgenic mice were generated as described in Materials and Methods; three independent transgenic lines for each transgene (designated a, b, and c) were examined. All transgenic mice were maintained as heterozygotes and bred with Star^+/− mice to produce Star^−/−WTTg or Star^−/−N47Tg mice. A, Quantitative Southern blotting with a ³²P-labeled StAR probe demonstrates comparable copy numbers of the WT- and N47-StAR transgenes. Star^−/−WTTg and Star^−/−N47Tg mice were genotyped as having both two disrupted Star genes (KO) and the WT- or N47-StAR transgene, respectively. Relative copy numbers of the StAR transgenes are shown for each independent line of WT-StAR BAC or N47-StAR BAC transgenic mice. B, The StAR BAC transgenes direct wild-type levels of StAR mRNA in the adrenal cortex and testis. Adrenal or testicular RNA was extracted from newborn pups and analyzed by real-time RT-PCR. None of these pups received corticosteroid treatment before being killed. Transcript levels are normalized to wild-type mice (1.0), and data are shown as mean ± sem (n = 5). C, Analysis of StAR protein expression in primary steroidogenic tissues. Recombinant StAR proteins were produced in HEK 293 cells as described in Materials and Methods. Tissue lysates were prepared from adrenal glands, testes, and ovaries of adult mice (8 wk of age) with or without trophic hormone stimulation, and levels of StAR protein were quantitated by immunoblot analyses using anti-StAR and anti-actin antisera. The WTTg(a) and N47Tg(a) BAC transgenic lines were used for this protein expression study. The WT-StAR protein was detected as a preprotein of 37 kDa and a mature protein of 30 kDa, whereas the N47-StAR protein has a molecular mass of 31 kDa. All steroidogenic tissues from Star^−/−WTTg mice and adrenal cortex and ovary from Star^−/−N47Tg mice expressed levels of StAR protein comparable to those from Star^+/+ mice despite the lower amount of StAR protein in testis from Star^−/−N47Tg mice. Adrenal expression of endogenous and BAC-directed StAR proteins was increased 2- to 3-fold at 3 h after three intermittent injections of 25 μg ACTH. Testicular expression of the various StAR proteins was significantly enhanced by hCG treatment, but the levels in Star^−/−N47Tg mice were lower than those in Star^−/−WTTg mice. Ovarian StAR expression was about 5-fold increased by hCG plus PMSG treatment in females.

Deletion of the mitochondrial targeting signal is predicted to dramatically alter the localization of StAR protein within steroidogenic cells. We therefore performed cell fractionation studies with adrenal cells to examine where StAR localized. As expected, Cyp11a1 was found exclusively in the mitochondrial fraction (Fig. 4A), validating the success of the fractionation. However, the localization of the WT- and N47-StAR proteins differed considerably: the WT-StAR protein colocalized with Cyp11a1 in the mitochondria, whereas the N47-StAR protein was found in the supernatant (Fig. 4A). Although technical difficulties precluded complete separation, a similar change in the subcellular localization was seen with extracts from testes and ovaries (supplemental Fig. Sl, published as supplemental data on The Endocrine Society's Journals Online web site at http://mend.endojournals.org).

Figure 4.

Intracellular Localization of the Various StAR Proteins in the Adrenal Glands of Star^+/+ (WT-STAR) or Star^−/−N47Tg (N47-StAR) Mice

A, In fractionated adrenal extracts, the WT-StAR protein colocalized with Cyp11a1 protein in the mitochondria (Mt), whereas the N47-StAR protein was exclusively detected in the supernatant (Sn). B, High-power magnification of immunogold labeling in electron microscopy of a zona fasciculata cell from Star^+/+ mouse (top) or Star^−/−N47Tg mouse (bottom) immunodetected with anti-StAR antiserum. Whereas the majority of WT-StAR protein labeling (101 particles) was located inside the mitochondria (m) relative to the cytosolic content (28 particles), N47-StAR protein is dominantly immunolocalized to the cytoplasm (107 particles) vs. inside the mitochondria (23 particles). The zona fasciculata cells are filled with typical large lipid droplets (L), and the mitochondria exhibit ample tubulovesicular cristae. The lipid droplets of zona fasciculata cells in Star^−/−N47Tg mouse were smaller and the mitochondrial cristae showed more tubular and less vesicular appearance than those of the Star^+/+ mouse. No labeling was observed in the nucleus (N) or the lipid droplets. Scale bars, 500 nm.

To confirm this discrepant localization, we performed localization studies using gold-conjugated anti-StAR antibodies. As shown in Fig. 4B, the majority of WT-StAR protein located inside the mitochondria (101 immunogold particles in mitochondria vs. 28 particles in the cytosol). In contrast, N47-StAR protein dominantly localized to the cytoplasm (107 particles) relative to the mitochondria (23 particles). These data collectively support our ability to express comparable levels of StAR protein but also document markedly different localizations within steroidogenic cells.

Differing Functions of the WT- and N47-StAR Transgenes in Vivo

One important indicator of function of the StAR transgenes is their ability to rescue neonatal lethality (8,9). Although Star^+/+ and Star^+/− mice almost invariably survived the postnatal period, Star^−/− mice invariably died within the first week of life (11,12,13). To assess function, we examined the survival of Star^−/− mice carrying either the WT- or N47-StAR transgene (Star^−/−WTTg or Star^−/−N47Tg, respectively). As shown in Fig. 5A, we again observed a marked divergence in survival of Star^+/+ and Star^−/− mice, with the latter always dying in the postnatal period. However, analysis of Star^−/−WTTg and Star^−/−N47Tg mice revealed an interesting dichotomy. Consistent with its expression at levels equal to endogenous StAR, the WT-StAR transgene uniformly rescued the Star^−/− mice, with survival comparable to that seen with Star^+/+ or Star^+/− mice. These findings validate the utility of BAC transgenesis to rescue StAR function. In contrast, Star^−/−N47Tg mice typically (∼60%) died in the immediate postnatal period. Given that the essential role of StAR in vivo is to facilitate the biosynthesis of adrenal corticosteroids, these findings suggest that comparable expression of StAR protein lacking the mitochondrial targeting signal cannot fully reconstitute StAR function in vivo. However, a minority (∼40%) of Star^−/−N47Tg mice survived the postnatal period, a finding never seen in Star^−/− mice, suggesting that N47-StAR protein retains some capacity to facilitate cholesterol transfer to the Cyp11a1 complex inside the mitochondria.

Figure 5.

Phenotypic Characterization of Star^−/−WTTg and Star^−/−N47Tg Mice

A, Impaired postnatal survival of Star^−/−N47Tg and Star^−/− mice. Newborn pups were genotyped on d 5–9 and observed until 8 wk of age. Percent survivals were determined at various ages. Star^+/+ mice almost invariably survived, whereas the Star^−/− mice universally died within 1 wk after birth. Although the BAC-derived WT-StAR protein restored survival to levels comparable to those seen in the Star^+/+ mice, the Star^−/−N47Tg provided only a partial rescue to about 40% survival. *, P < 0.05. B, Defective virilization in Star^−/−N47Tg and Star^−/− mice. Star^−/− mice were kept alive to adulthood with corticosteroid replacement. Shown are the external genitalia of adult males (8 wk of age); note the normal genitalia in Star^−/−WTTg mice, hypoplastic penis and scrotum with a shortened anogenital distance in Star^−/−N47Tg mice, and female genitalia in Star^−/− mice.

Because StAR also regulates testosterone synthesis in fetal and adult testes (3,4), virilization provides another indirect indicator of StAR function in male mice. Consistent with our previous findings (11,12,13), Star^−/− males had completely feminized external genitalia and bilateral inguinal testes, supporting the indispensable role of StAR in male sexual development. As shown in Fig. 5B, Star^−/−WTTg males were phenotypically normal, with fully descended testes and normal male external genitalia; they also had well-developed seminal vesicles and prostate (data not shown). In contrast, Star^−/−N47Tg males had undermasculinized external genitalia, with hypoplastic penis and scrotum and a shortened anogenital distance. Moreover, about 50% of Star^−/−N47Tg males had unilateral or bilateral cryptorchidism, which may cause secondary testicular dysfunction (26). Interestingly, the external genitalia of Star^−/−N47Tg males were grossly normal until 4–5 wk of age, when postnatal sexual maturation normally occurs, but showed no further development postnatally even at 24 wk, suggesting that the N47-StAR protein retains sufficient activity in utero to produce enough testosterone for sex differentiation. Nevertheless, postnatal sexual maturation is apparently impaired, probably reflecting both decreased StAR function in testosterone biosynthesis by Leydig cells and adverse effects due to the cryptorchidism.

Circulating Levels of Steroid Hormones in StAR KO Mice Carrying the WT- or N47-StAR Transgene

To confirm more directly the function of the WT- and N47-StAR transgenes in primary steroidogenic tissues, we determined circulating levels of steroid hormones in 8-wk-old mice of different genotypes. To evaluate StAR function in the adrenal cortex, we measured serum corticosterone (the predominant glucocorticoid in mice) and plasma ACTH; we also measured the acute steroidogenic response to ACTH. As shown in Table 1, Star^−/− mice that were kept alive to adulthood with corticosteroid replacement had markedly decreased serum corticosterone and significantly elevated plasma ACTH, consistent with severe primary adrenal insufficiency. Whereas Star^−/−WTTg mice had normal corticosterone and ACTH levels, indicating full rescue of glucocorticoid production, Star^−/−N47Tg mice that survived to 8 wk of age had basal ACTH levels that did not differ significantly from normal; they did, however, have significantly reduced basal and ACTH-stimulated levels of corticosterone. These data argue that the protein encoded by the N47-StAR transgene is not fully active in facilitating steroidogenesis in the adrenal cortex, even in the minority of mice that can survive to 8 wk without corticosteroid replacement.

Table 1.

Plasma ACTH and Serum Steroid Hormone Levels at Eight Weeks of Age

	ACTH (pg/ml), basal (n = 10)	Corticosterone (ng/ml)		Testosterone in Male (ng/ml)		Progesterone in Female (ng/ml)		Estradiol in Female (pg/ml)
		Basal (n = 20)	Stimulated (n = 10)	Basal (n = 10)	Stimulated (n = 5)	Basal (n = 10)	Stimulated (n = 5)	Basal (n = 10)	Stimulated (n = 5)
Star+/+	66 ± 9.9	197 ± 18	418 ± 28	2.2 ± 0.7	28 ± 2.5	4.9 ± 0.3	26 ± 4.0	19 ± 1.3	24 ± 2.2
Star−/−WTTg	46 ± 4.3	186 ± 16	393 ± 30	2.3 ± 1.1	30 ± 3.6	5.0 ± 0.8	24 ± 2.7	18 ± 1.0	21 ± 2.7
Star−/−N47Tg	80 ± 20	123 ± 21a	200 ± 34a	1.7 ± 0.2	3.4 ± 1.1a	3.0 ± 0.3a	5.8 ± 0.5a	18 ± 2.0	18 ± 1.6
Star−/−	339 ± 82a	4.6 ± 1.1a	27 ± 6.2a	0.5 ± 0.1a	0.7 ± 0.1a	0.6 ± 0.4a	1.7 ± 0.1a	19 ± 7.5	19 ± 0.2

At basal state, blood samples were collected between 1200 and 1500 h. At stimulated state, blood samples were analyzed after injection(s) of ACTH or gonadotropins as described in Mateials and Methods. Star^−/− mice were kept alive with sc injections of hydrocortisone and fludrocortisone acetate until 24 h before they were killed. The values were expressed as the mean ± sem. 

^aP < 0.05. 

We also analyzed gonadal steroids in 8-wk-old mice, both basal and stimulated by treatment with gonadotropins. As shown in Table 1, male Star^+/+ mice had a marked (approximately 10-fold) increase in serum testosterone levels in response to human chorionic gonadotropin (hCG), whereas male Star^−/− mice had low basal testosterone levels and showed no response to gonadotropin treatment. In this regard, the two StAR BAC transgenes again behaved differently. Whereas Star^−/−WTTg mice had normal basal testosterone levels and responded normally to gonadotropin stimulation, the Star^−/−N47Tg mice had somewhat lower basal levels of testosterone, which did not differ significantly from those in either Star^+/+ or Star^−/−WTTg mice, but showed minimal response to hCG. Similarly, basal progesterone levels in female Star^−/−N47Tg mice were modestly reduced, and the mice showed a markedly blunted response to gonadotropins relative to Star^+/+ and Star^−/−WTTg females. These data again argue strongly that the N47-StAR protein cannot fully replace the WT-StAR protein in facilitating the acute induction of steroidogenesis in the gonads. In contrast, estradiol levels did not differ between Star^−/−WTTg and Star^−/−N47Tg females (or even in Star^−/− females); this may reflect the very high potency of estradiol in vivo and a sufficient reserve in the steroidogenic capacity to maintain estradiol levels.

Differential Effects of the WT- and N47-StAR Transgenes on Histology of the Primary Steroidogenic Tissues

We next sought to explore further the function of the StAR transgenes in the different steroidogenic tissues. One hallmark of StAR deficiency in steroidogenic organs is the accumulation of cholesterol inside the cytosol, driven by persistent trophic hormone stimulation in the setting of a block in cholesterol translocation into the mitochondria (8,9). We therefore examined steroidogenic tissues from mice of different genotypes, looking both at histology and at lipid deposits. In the adrenal glands, Star^−/−N47Tg mice showed marked vacuolization (Fig. 6), consistent with florid lipid deposition, whereas Star^−/−WTTg mice appeared normal. Testes of Star^−/−N47Tg mice exhibited apparent Leydig cell hyperplasia in the interstitial region but had normal-appearing sperm in the seminiferous tubules, suggesting residual capacity for spermatogenesis. Star^−/−N47Tg ovaries also had vacuolated regions in the interstitial and theca cells, suggestive of impaired cholesterol translocation. However, some corpora lutea were noted, consistent with some capacity for ovulation in these mice.

Figure 6.

Hematoxylin-Eosin Staining of Adrenal Glands, Testes, and Ovaries from Adult Mice (Eight Weeks of Age)

Organs were collected from Star^−/−WTTg and Star^−/−N47Tg mice and processed for histology as described in Materials and Methods. Shown are sections of adrenal glands (×100 magnification), testes (×100 magnification), and ovaries (×40 magnification). The arrow indicates Leydig cell hyperplasia in the interstitial region of the testis section from the Star^−/−N47Tg mouse. CL, Corpus luteum.

We next directly examined neutral lipid deposits by staining with oil red O. As shown in Fig. 7, the adrenal cortex and testes of Star^+/+ mice have minimal lipid deposits, whereas the ovaries showed moderate oil red O staining consistent with some cholesterol accumulation. In contrast, all of the primary steroidogenic cells in the Star^−/− mice exhibited florid lipid deposits. Consistent with the phenotype described above, the WT-StAR BAC transgene rescued cholesterol trafficking in the steroidogenic cells, such that lipid deposits were comparable to those seen in Star^+/+ mice. Strikingly, the N47-StAR BAC transgene had minimal if any effect on the oil red O staining in the adrenal cortex and ovaries but was associated with some decreased intensity of oil red O staining in the interstitial region of the testes. These results again argue that the N47-StAR protein is significantly impaired in its ability to mediate cholesterol translocation but also suggest that there may be tissue-specific differences in the function of the N47-StAR protein.

Figure 7.

Oil Red O Staining of Adrenal Glands, Testes, and Ovaries from Star^+/+, Star^−/−WTTg, Star^−/−N47Tg, and Star^−/− Mice

Star^−/− mice were kept alive to adulthood with corticosteroid replacement. Organs were harvested from adult mice (8 wk of age) and processed as described in Materials and Methods. Shown are lipid deposits in adrenal gland (×100 magnification), testis (×40 magnification), and ovary (×40 magnification) from mice of the indicated genotypes.

In addition to StAR, a number of other proteins play important roles in providing cholesterol substrate for steroidogenesis, including the scavenger receptor-B1 (SRB1), low-density lipoprotein receptor (LDLR), and hormone-sensitive lipase (HSL). We therefore used quantitative PCR analysis to examine the levels of transcripts encoding these proteins in the steroidogenic tissues from mice of the different genotypes. As shown in Fig. 8, significantly increased expression of SRB1 and HSL and decreased expression of LDLR were seen in the adrenal gland of StAR^−/− mice relative to levels in StAR^+/+ mice. Interestingly, although the Star^−/−WTTg mice restored expression of SRB1, LDLR, and HSL to WT levels in the adrenal gland, SRB1 expression in Star^−/−N47Tg mice remained elevated. These findings again highlight the inability of N47-StAR to fully restore normal function in adrenocortical cells.

Figure 8.

Expression of Genes Related to Cholesterol Delivery in the Primary Steroidogenic Tissues

Organs were removed from mice of the indicated genotypes, and total RNA was prepared and subjected to real-time RT-PCR analysis as described in Materials and Methods. Shown are the relative expression levels of SRB1, LDLR, and HSL in the adrenal glands, testes, and ovaries of mice with the indicated genotypes, all normalized to expression levels in WT mice (1.0). *, P < 0.05 relative to levels in wild-type tissues. Data are shown as mean ± sem (n = 5).

DISCUSSION

Although genetic evidence in human patients with lipoid CAH and in KO mice unequivocally established the essential role of StAR in steroidogenesis, the precise mechanism of StAR action has not been fully defined despite considerable effort by several groups (5,6,7). The crystal structures of proteins that share the StAR-related lipid transfer domain suggest that StAR forms a hydrophobic pocket that can accommodate one molecule of cholesterol (19,22), whereas biochemical analyses indicate that StAR facilitates cholesterol transfer by actions at the outer membrane (23). Perhaps most surprisingly, a truncated StAR protein lacking the mitochondrial targeting signal reportedly increased steroidogenesis in transfected COS-1 and in mitochondria isolated from MA-10 Leydig cells (14,15,27).

To explore the relevance of these studies in authentic steroidogenic cells at expression levels comparable to those of WT-StAR, we developed a mouse model for expression of different StAR proteins in the primary steroidogenic cells, all in the absence of endogenous StAR activity due to homozygosity for the StAR KO allele. Our results reveal a complexity of StAR function that both supports and extends beyond previous systems of analysis. Whereas the WT-StAR transgene completely rescued all aspects of StAR function, the N47-StAR transgene functioned at a level permitting male sex differentiation in utero but did not confer full virilization of the external genitalia at puberty. These data again support important differences in the function of the WT- and N47-StAR proteins in steroidogenesis. However, despite their markedly blunted testosterone responses to hCG and their hypoplastic external genitalia, the seminiferous tubules of Star^−/−N47Tg transgenic mice contained apparently normal sperm (Fig. 6), and the males examined were fertile (data not shown). In addition, most of the mice died in the immediate postnatal period; the frequency of survivors was increased by treatment with corticosteroids (data not shown), indicating that this death likely was due to adrenal insufficiency. Finally, the Star^−/−N47Tg mice had lipid deposits in the adrenal cortex and ovary comparable to those seen in Star^−/− mice, again supporting impaired function. Despite this, with the exception of estradiol levels, basal and stimulated circulating levels of steroid hormones of Star^−/−N47Tg mice are significantly greater than those of Star^−/− mice. Thus, although the function of N47-StAR in vivo clearly is not equal to that of WT-StAR, our data are inconsistent with the model that StAR must reach the intermembranous space in the mitochondria to act.

In general, transgenic structure-function experiments may be confounded by differences in copy number or insertion site of the transgene or by its failure to accurately match the endogenous expression of the gene studied. In our studies, we observed highly comparable results with three independently derived transgenic lines for both the WT-StAR and the N47-StAR transgenes, both with respect to rescue of the StAR KO phenotype and with respect to basal and trophic hormone-induced levels of StAR expression. Thus, we believe that our model system provides a novel strategy to assess key aspects of StAR structure and function within the context of authentic steroidogenic cells. It should be noted that we did not achieve comparable levels of expression in several transgenic lines whose copy number was only 1 or 2 (data not shown); thus, the transgenes studied here may not completely parallel the efficiency of the Star locus. In addition, we did not observe any phenotype associated with the various StAR BAC transgenes that might reflect effects of transgene insertion or overexpression of Star or other gene(s) on the BAC clone.

Although our transgenic mouse rescue system was developed to provide insights into StAR function in the steroidogenic tissues in vivo in a manner not possible in previously studied systems, there are limitations to the analyses that can be performed. Thus, although we can document the levels of basal and trophic hormone-induced StAR expression and the effects on phenotype and steroid hormone levels, we have not yet defined the basis for the apparent tissue-specific degree of rescue that we observed. First, testicular StAR expression in Star^−/−N47Tg mice is lower than that in Star^+/+ and Star^−/−WTTg males, even after hCG stimulation, whereas ovarian expression is comparable among the three genotypes (Fig. 3C). The Star^−/−N47Tg male mice frequently have cryptorchidism and diminished basal levels of testosterone, whereas Star^−/−N47Tg females have normal estradiol levels (Table 1). Cryptorchidism, which may reflect decreased androgen action (26), could induce secondary testicular dysfunction that impairs StAR expression, thereby decreasing testosterone production and possibly delaying maturation of the hypothalamus-pituitary-gonadal axis (28). Second, cholesterol accumulation due to impaired StAR action may also have adverse effects on steroidogenesis. The phenotype of human patients with lipoid CAH prompted a two-hit model: the first hit is loss of StAR-dependent steroidogenesis, whereas the second hit is loss of StAR-independent steroidogenesis due to trophic hormone-driven cholesterol accumulation and cellular damage (9). Studies in StAR KO mice have confirmed the important role of trophic hormones in cholesterol accumulation (13), an important tenet of this model. The proposed secondary damage may progress relatively independently in the different steroidogenic tissues, allowing for tissue-specific effects of our transgenic rescue. Finally, the mechanism of cholesterol homeostasis may not be fully equivalent in the various steroidogenic tissues. Cultured steroidogenic tissues and cells show a variety of responses to extracellular cholesterol-rich lipoproteins, various types of intracellular donor particles of cholesterol ester, and different levels of SRB1 expression (29); the latter finding is consistent with the apparent tissue-specific effects of the WT- and N47-StAR transgenes on expression of SRB1, LDLR, and HSL in this study (Fig. 8). Nonetheless, our data clearly highlight a complexity of StAR function in vivo that has not been revealed by previous studies in cell transfection and mitochondrial reconstitution experiments (14,15,27).

The truncated N47-StAR protein lacks the mitochondrial targeting signal and thus does not localize to the mitochondria (Fig. 4). In contrast, the WT-StAR protein is translated as a 37-kDa preprotein that contains the N-terminal mitochondrial targeting signal, which directs StAR import into the mitochondria (3). Coincident with import, the N-terminal signal is sequentially cleaved by two processing peptidases to yield a 32-kDa intermediate form and 30-kDa mature form, respectively (4,17). The cleavage sites are located between amino acids 39/40 and 55/56 for bovine StAR (30), and the corresponding putative cleavage sites are found in the mouse Star sequence; thus, the predicted size is compatible with our finding that mouse N47-StAR protein has a migration on SDS-PAGE intermediate to these two forms. Although the N47-StAR protein localizes to the cytosol and was not detected inside the mitochondria, we cannot exclude the possibility that there are transient interactions of the N47-StAR protein with components of the outer mitochondrial membrane; the partial rescue of viability and steroidogenesis in StAR KO mice carrying the N47-StAR transgene would suggest that this indeed may occur (31). In fact, one possibility for the lack of significant differences in basal circulating levels of steroid hormones and the restoration of viability in about 40% of the N47-StAR transgenic mice is that a small subset of StAR that does reach the outer mitochondrial membrane is not susceptible to mitochondrial import (14,15) and thus is able to facilitate cholesterol transfer for a prolonged period of time (32).

Although primarily developed to facilitate analysis of targeted gene expression by the StAR BAC, the StAR/eGFP transgene is specifically expressed in the steroidogenic cells of the adrenal cortex, testis, and ovary. Thus, it provides a useful reagent for selectively isolating the steroidogenic cells that normally express StAR. For example, we previously used transgenesis with the regulatory sequences of steroidogenic factor 1 (SF-1) and the eGFP reporter to study sex-specific differences in gene expression in the somatic cells of the developing gonads (33,34). A similar strategy could be used to purify the steroidogenic cells, particularly in embryonic testes where StAR is expressed at high levels and facilitates the production of testosterone to mediate male sex differentiation. Despite reports of StAR expression in potentially steroidogenic cells such as neuronal and glial cells (35), we have not detected eGFP expression in the brains of StAR/eGFP mice (data not shown). We do not know whether this reflects a level of expression below the sensitivity of our assay or the lack of important regulatory elements for central nervous system expression in the Star sequences used here. Nonetheless, the StAR/eGFP BAC transgene should provide a facile approach to analyze steroidogenic cell lineages of adrenal glands and gonads.

In summary, our transgenic expression studies document that N47-StAR, when expressed at wild-type levels in the adrenal cortex and gonad of transgenic mice, restores some but not all aspects of steroidogenesis. The partial deficiency of N47-StAR differs from conclusions drawn from transfection analyses in nonsteroidogenic COS-1 cells, highlighting the potential pitfalls in trying to rely on expression or reconstitution studies outside of the context of bona fide steroidogenic cells. One goal for future studies is to develop a more physiological system for studying StAR function by oncogene-driven immortalization of steroidogenic cells from the StAR KO mice (36); these cells should provide a system for studying StAR function in the context of normal levels of expression of the other components of steroidogenesis normally found in cells of the primary steroidogenic tissues.

MATERIALS AND METHODS

Experimental Animals

All mouse studies were approved by the Institutional Animal Care Research Advisory Committee at University of Texas Southwestern. A WT-StAR BAC clone (Fig. 1A) was isolated from a C57BL/6J mouse genomic DNA library and provided by the Genome Center at Washington University (St. Louis, MO). StAR/eGFP and N47-StAR BAC clones were generated by homologous recombination to modify the WT-StAR BAC (37). In brief, two 1-kb recombination cassettes were cloned into pBluescript KS (−) (Stratagene, La Jolla, CA). One recombination cassette for the StAR/eGFP BAC contained two 500-bp fragments flanking the StAR translation initiation site, eGFP cDNA (Clontech, Mountain View, CA), and a polyadenylation signal from bGH (Fig. 1C). The other recombination cassette consisted of the 3′ region of exon 1, the entire region of intron 1, and the 5′ region of exon 2 of Star gene; four nucleotide substitutions were introduced into the StAR coding sequence by site-directed mutagenesis that changed three adenine residues from potential initiator ATGs to thymidine residues and a thymidine residue three nucleotides 5′ of the new initiator methionine to an adenine residue to better match the Kozak sequence (Fig. 1D). The recombination cassettes were subcloned into the shuttle vector pSV1.RecA and then transformed into recombination-deficient E. coli containing the WT-StAR BAC. After homologous recombination and resolution, BAC modification was verified by Southern blot analysis of BAC DNA resolved by pulsed-field gel electrophoresis (Fig. 1B) and then confirmed by direct sequencing.

The BAC DNAs for pronuclear injection were prepared with the Large-construct kit (QIAGEN, Valencia, CA) followed by purification by CsCl gradient ultracentrifugation; the DNAs were then resuspended in microinjection buffer [10 mm Tris-HCl (pH 7.4), 0.1 mm EDTA, and 100 mm NaCl]. Each BAC was injected as supercoiled DNA into the pronucleus of C57BL/6J fertilized eggs to produce transgenic pups, which were identified by genomic PCR using a forward primer for the vector sequence (5′-AGCTGAAGCCATATTGGGGAACAAG-3′) and a reverse primer for the Star 5′-flanking region (5′-AGTGAATTGTAATACGACTCACTATAGGGC-3′). Transgenic mice were always maintained as heterozygotes. Either WT- or N47-StAR transgenic mice were crossed with Star^+/− mice maintained as a congenic line on the C57BL/6J background (Jackson Laboratory, Bar Harbor, ME) to produce Star^−/− mice carrying either the WT-StAR BAC transgene (Star^−/−WTTg) or the N47-StAR BAC transgene (Star^−/−N47Tg), as genotyped by quantitative Southern blotting (Fig. 3A). Star^−/− mice were also produced by matings of Star^+/− mice, which were rescued by daily postnatal sc injections with hydrocortisone (Sigma Chemical Co., St. Louis, MO) and fludrocortisone acetate (Sigma) until 8 wk of age as described (12,13). All mice were housed in temperature-controlled rooms with a 12-h light, 12-h dark cycle and were given food and water ad libitum.

To evaluate the capacity for steroidogenesis, some mice at 8 wk of age were treated with ACTH or gonadotropins. The ACTH 1–24 fragment, hCG, and pregnant mare serum gonadotropin (PMSG) were purchased from Sigma and dissolved in sterile saline. To stimulate the adrenal cortex, mice received a single ip injection of 75 μg ACTH at 1200 h and were killed at 1230 h (Table 1) or three ip injections of 25 μg ACTH at 1200, 1300, and 1400 h followed by euthanasia at 1500 h (Fig. 3C). To stimulate testosterone production, males were injected ip with 10 U hCG at 1200 h once daily for three consecutive days (d 1–3) and then killed at 1200 h on d 4. To induce ovulation, females were injected ip with 10 U PMSG at 1200 h on d 1, followed by 10 U hCG at 1200 h on d 3; mice were then killed at 1200 h on d 4.

Southern Blot Analysis

Southern blotting was used to analyze resolution of the StAR BAC clones and to genotype mice carrying the StAR BAC transgenes. For BAC resolution, 0.1 μg purified BAC was digested with restriction enzymes and resolved by pulsed-field gel electrophoresis at 4 V/cm, 1–15 sec linear ramping for 16 h at 14 C (Bio-Rad Laboratories, Hercules, CA). For genotyping, 10 μg of DNA purified from tail snips was digested overnight with NcoI and NdeI at 37 C and separated on a 1% agarose gel at 100 V for 2 h. The digested DNA was transferred to a Nytran membrane (Whatman, Florham Park, NJ) using the TurboBlotter Downward Transfer system (Whatman). A 0.8-kb probe corresponding to the exons 1 and 2 and intron 1 of Star gene was generated by PCR and labeled with ³²P using the RediprimeII random-prime labeling system (Amersham Biosciences, Piscataway, NJ). The probe was hybridized with the membrane at 68 C for 16 h in ExpressHyb solution (Clontech). Radioactivity on the membrane was detected using a Storm 820 scanner (Molecular Dynamics, Sunnyvale, CA) and quantified using ImageQuant software (Molecular Dynamics).

RNA Extraction and Real-Time RT-PCR

Total RNA was extracted from adrenal glands, testes, or ovaries using TRIzol reagent (Invitrogen, Carlsbad, CA). The isolated RNA was treated with DNase I (Roche, Indianapolis, IN) at 37 C for 30 min to eliminate contaminating genomic DNA. Reverse transcription of 2 μg of the RNA was carried out using random hexamers (Roche) and SuperScript III reverse transcriptase (Invitrogen). For quantitative analysis, the 6FAM-dye-labeled TaqMan MGB probes for mouse StAR (Mm00441558_m1), mouse SRB1 (Mm00450236_m1), mouse LDLR (Mm00440169_m1), and mouse HSL (Mm00495359_m1) and for eukaryotic 18S rRNA (Hs99999901_s1) were purchased (Applied Biosystems, Foster City, CA); TaqMan gene expression assay was performed using the ABI Prism 7700 Sequence Detection system (Applied Biosystems) according to the manufacturer's protocol. Expression values were analyzed by the standard curve method, normalized for 18S rRNA using the Applied Biosystems software. All reactions were performed in triplicate to assess well-to-well variability, and only curves having a high correlation coefficient (r² > 0.99) were used.

Protein Preparation and Immunoblotting

Recombinant StAR proteins were produced using the mouse WT-StAR cDNA subcloned into pACCMV.pLpA vector. The N47-StAR cDNA encoding a mutated StAR protein with deletion of the first 47 amino acids was created from the WT-StAR cDNA as described (15). Using FuGENE 6 according to the manufacturer's instructions (Roche), 1 μg of each plasmid DNA was transiently transfected into HEK293 cells maintained in DMEM supplemented with 10% fetal bovine serum. Twenty-four hours after transfection, cells were harvested in 1 ml cold 10% 6.1 n trichloroacetic acid (Sigma) by incubating on ice for 1 h. Cell lysates were centrifuged for 15 min at 19,000 × g, and the supernatant was removed. Protein pellets were resuspended in sample buffer (7.2 m urea, 1.6% Triton X-100, 0.8% dithiothreitol, 2% lauryl sulfate; all purchased from Sigma). For analyses of StAR protein levels, whole adrenal glands, testes, or ovaries were homogenized with a sterile pestle grinder (Kontes, Vineland, NJ) and prepared as described above. To collect mitochondria-enriched fractions, fresh adrenals, testes, and ovaries were homogenized using a Con-Torque homogenizer with a loosely fitting Teflon pestle in 0.25 m sucrose, 10 mm Tris-HCl, 1 mm EDTA (pH 7.4). A portion of the homogenate was centrifuged at 400 × g for 10 min, and the supernatant was then centrifuged at 14,000 × g for 20 min. The final supernatants and the mitochondrial pellets were resuspended in the homogenization buffer.

Lysates (10 μg protein) were mixed with Tris-glycine SDS sample buffer (Invitrogen), separated on a 12% Tris-glycine gel (Invitrogen) along with prestained molecular weight markers (Bio-Rad) at 120 V for 2 h, and electrophoretically transferred to polyvinylidene difluoride membrane (Bio-Rad) at 40 V for 4 h. Immunoblot analyses used a rabbit polyclonal anti-StAR antiserum (1:10,000) kindly provided by Dr. Dale Hales, rabbit anti-β-actin antiserum (1:5000) (Novus, Littleton, CO), or a rabbit anti-Cyp11a1 antiserum (1:10,000) (Chemicon, Temecula, CA). Proteins were visualized by exposure to x-ray film after treatment of the membrane with chemiluminescence Luminol reagent (Santa Cruz Biotechnology, Santa Cruz, CA). Band intensities were quantified with YabGelImageX1.0 (http://homepage.mac.com/yabyab/rb/gelimage.html).

Hormone Assays

Mice at 8 wk of age were anesthetized, and blood was collected by cardiac puncture between 1200 and 1500 h. RIA analyses for steroid hormones were performed by Drs. David Hess and Richard Yeoman, Oregon Regional Primate Center, including corticosterone [limit of detection, 5 pg/tube; intraassay and interassay coefficients of variation (CV), 4.5 and 8.0%, respectively; recovery, 99.2%], testosterone (limit of detection, 5 pg/tube; intraassay and interassay CV, 4.1 and 3.9%, respectively; recovery, 94.3%), progesterone (limit of detection, 5 pg/tube; intraassay and interassay CV, 9.7 and 11%, respectively; recovery, 93%), and estradiol (limit of detection, 10.0 pg/ml; intraassay and interassay CV, 4.5 and 5.7%, respectively). Plasma concentrations of ACTH were measured with a commercially available ELISA kit (MD Biosciences, St. Paul, MN) with range of 8–475 pg/ml, limit of detection of 0.46 pg/ml, and intraassay and interassay CV of 4.2 and 6.2%, respectively.

Histological Analysis

Morphological analysis was performed with frozen or Bouin's-fixed tissues harvested from anesthetized mice. The frozen tissues were embedded in optical cutting temperature compound (Sakura Finetex USA Inc., Torrance, CA) and sectioned at 10 μm with a Cryostat CM1900 (Leica Corp. Instruments GmbH, Nussloch, Germany). The sections were analyzed using an Optiphot microscope (Nikon, Melville, NY) equipped with a UV light source and filters to visualize eGFP expression or by staining with oil red O (Sigma) and hematoxylin (Richard-Allan Scientific, Kalamazoo, MI). The Bouin's-fixed tissues were embedded in a mix of purified paraffin and plastic polymers of regulated molecular weights (Oxford Labware, St. Louis, MO), sectioned at 7 μm with a Rotary Microtome HM330 (Microm GmbH, Heidelberg, Germany), and then stained with hematoxylin and eosin (Richard-Allan Scientific).

Decapsulated adrenal tissues were cut in small fragments (1 mm³) and fixed in freshly prepared 0.1 m Na-cacodylate buffer containing 0.05% electron microscopy-grade glutaraldehyde and 3% paraformaldehyde. After overnight incubation at 4 C, the tissue fragments were washed in PBS and dehydrated in a graded series of alcohols (10, 25, 50, and 70%). The fragments were then infiltrated with LR White resin (London Resin Co., Basingstoke, UK) and placed in gelatin capsules (EMS, Fort Washington, PA) for polymerization at 50 C for 24 h. The 70-nm sections were cut with an Ultratome 3 and collected on coated nickel grids (300 square mesh; Agar Scientific, Stansted, UK) coated with 1% parlodion in amyl acetate (EMS). Before incubation with antiserum, nonspecific antigenic sites were blocked by incubation for 5 min at room temperature with normal goat serum (1:100 dilution) in antiserum incubation buffer [0.9% NaCl, 10 mm Tris-HCl (pH 8.2), 0.1% Tween 20]. The sections were incubated overnight (4 C) with a 1:20 dilution of rabbit anti-StAR antiserum, followed by incubation with a 1:10 dilution of 10-nm gold-labeled goat antirabbit IgG. The sections were then examined on a transmission Philips Technai 12 electron microscope (Eindhoven, The Netherlands) equipped with a MegaView camera (Soft Imaging System GmbH, Münster, Germany).

Statistical Analysis

The mean values between groups were analyzed by Mann-Whitney U test; differences with P < 0.05 were considered significant.