Cloning of Factors Related to HIV-inducible LBP Proteins That Regulate Steroidogenic Factor-1-independent Human Placental Transcription of the Cholesterol Side-chain Cleavage Enzyme, P450scc*

The cholesterol side-chain cleavage enzyme, cytochrome P450scc, initiates the biosynthesis of all steroid hormones. Adrenal and gonadal strategies for P450scc gene transcription are essentially identical and depend on the orphan nuclear receptor steroidogenic factor-1, but the placental strategy for transcription of P450scc employs cis-acting elements different from those used in the adrenal strategy and is independent of steroidogenic factor-1. Because placental expression of P450scc is required for human pregnancy, we sought factors that bind to the −155/−131 region of the human P450scc promoter, which participates in its placental but not adrenal or gonadal transcription. A yeast one-hybrid screen of 2.4 × 106 cDNA clones from human placental JEG-3 cells yielded two unique clones; one is the previously described transcription factor LBP-1b, which is induced by HIV, type I infection of lymphocytes, and the other is a new factor, termed LBP-9, that shares 83% amino acid sequence identity with LBP-1b. When expressed in transfected yeast, both factors bound specifically to the −155/−131 DNA; antisera to LBP proteins supershifted the LBP-9·DNA complex and inhibited formation of the LBP-1b·DNA complex. Reverse transcriptase-polymerase chain reaction detected LBP-1b in human placental JEG-3, adrenal NCI-H295A, liver HepG2, cervical HeLa, and monkey kidney COS-1 cells, but LBP-9 was detected only in JEG-3 cells. When the −155/−131 fragment was linked to a minimal promoter, co-expression of LBP-1b increased transcription 21-fold in a dose-dependent fashion, but addition of LBP-9 suppressed the stimulatory effect of LBP-1b. The roles of LBP transcription factors in normal human physiology have been unclear. Their modulation of placental but not adrenal P450scc transcription underscores the distinctiveness of placental strategies for steroidogenic enzyme gene transcription.

Steroid hormones regulate a wide variety of physiologic functions. Mineralocorticoids, produced by the adrenal cortex, are needed to retain sodium and maintain blood pressure (1); glucocorticoids, also produced by the adrenal cortex, raise blood sugar but also play roles in numerous physiologic processes (2); and androgens and estrogens, produced by the gonads, are required for reproduction (3,4). In human beings, absence of mineralocorticoids leads to death in infancy, absence of glucocorticoids may lead to death during times of severe physiologic stress, and absence of sex steroids would, eventually, lead to death of the species. Thus, the regulation of steroid hormone biosynthesis is of fundamental interest. The role of steroid hormones in the fetus is less obvious, and there are important species differences among mammals. Human fetuses can develop normally, reach term gestation, undergo normal parturition, and make initial adaptations to extrauterine life in the absence of mineralocorticoids, glucocorticoids, or sex steroids (5). By contrast, normal human gestation is totally dependent on the action of progesterone to suppress uterine contractility and thus to maintain pregnancy (6). Progesterone is initially provided by the mother's ovarian corpus luteum, but after about 8 -10 weeks virtually all progesterone is produced by the placental syncytiotrophoblast cells, which are fetal tissue (7,8). Therefore, placental synthesis of progesterone is essential for the initiation of human life (5).
The synthesis of placental progesterone, and of all other steroid hormones, begins with the conversion of cholesterol to pregnenolone by mitochondrial cytochrome P450scc, which is the rate-limiting and hormonally regulated enzymatic step in steroidogenesis (9). Human P450scc is encoded by a single gene (10), formally termed CYP11A (11), that is located on chromosome 15q23-q24 (12) and is expressed in the adrenals (13), gonads (13), placenta (14), and brain (15). Because of its key role in the production of all steroid hormones, the transcription of the P450scc gene has been the subject of intensive study (reviewed in Refs. 16 -18). Such studies led to the discovery of steroidogenic factor-1 (SF-1), 1 also known as Ad4-BP, an orphan nuclear receptor that is essential for fetal adrenal and gonadal development as well as for expression of all the steroidogenic genes in these tissues (19). However, although SF-1 is expressed in many tissues (20), very little SF-1 is expressed in other tissues that contain P450scc including brain, embryonic gut, and placenta (20 -22), and studies of P450scc regulation in the brain (23) and placenta (24,25) indicate that the SF-1 sites in the P450scc promoter are not involved in P450scc transcription in these tissues. As placental P450scc expression and progesterone synthesis are mandatory for successful pregnancy and as little is known about SF-1-independent expression of genes for steroidogenic enzymes, we have studied the transcription of P450scc in human placental JEG-3 cells (24,25). We now report the cloning and characterization of two transcription factors that modulate the human placental expression of P450scc; both factors are related to the LBP-1 family of transcription factors induced by HIV, type I infection of lymphocytes (26).
JEG-3 cDNA Library-Both randomly primed and oligo(dT)-primed cDNA made from JEG-3 cell RNA was ligated with phosphorylated EcoRI-NotI-SalI adapters and passed through Sephacryl S-300 to exclude DNA smaller than 300 -400 bp. The cDNA was ligated to the GAL4 activation domain (32,33) in the LEU2 plasmid pGAD10 and propagated in E. coli. The plasmid DNA library was introduced by LiAc/polyethylene glycol transformation into the yeast YM4271 (P450scc Ϫ155/Ϫ131 ϫ4 HIS3/LacZ) reporter strain and plated onto 16 243 ϫ 243-mm minimal media (SD) Ura Ϫ -Leu Ϫ -His Ϫ plates containing 45 mM 3-amino-1,2,4-trizole. Growth of a small aliquot on SD/Leu Ϫ selective media was used to estimate the number of transformants. After 7 days of growth on minimal medium, survivors were assayed for ␤-galactosidase activity (34). These plasmid DNAs were isolated from their host yeast and individually transformed into E. coli DH5␣ for amplification, restriction endonuclease mapping, and sequencing.
5Ј-RACE-Total JEG-3 cell RNA (1 g) was reverse transcribed into cDNA using 50 ng of primer 9-GSP1 (see Table I), isolated with Glass-MAX DNA spin cartridges (Life Technologies, Inc.), then oligo(dC)tailed with terminal deoxynucleotidyl transferase. The dC-tailed cDNA was amplified by a nested PCR procedure, first using the Abridge Anchor primer (Life Technologies, Inc.) and 9-GSP2, followed by reamplification with primers AUAP (Life Technologies, Inc.) and 9-GSP2, yielding a predominant PCR product of ϳ600 bp. This double-stranded cDNA was blunt-ended by T4 DNA polymerase, ligated into SmaIdigested pBluescript II SKϩ, and sequenced. The full-length cDNA was assembled by ligating the 5Ј-RACE clone into construct 9pSG5 containing the 3Ј-end of the cDNA, using the XhoI and Bst107I sites. 5Ј-RACE was also performed for clone 32 using primers 32-GSP-1 and 32-GSP-2, and the full-length cDNA was assembled by ligating the 5Ј-RACE clone into pGAD10 -32 containing the 3Ј-end of the cDNA using the XhoI and BclI sites.
RT-PCR-Total RNA was isolated from JEG-3 cells, NCI-H295A cells, HeLa cells, HepG2 cells, COS-1 cells, and human adrenal tissue. Random primers were used for the first strand cDNA syntheses using reverse transcriptase Superscript II (Life Technologies, Inc.). PCR was done using 5 l of the first strand cDNA product and oligonucleotides specific for clones 1, 9, 32, human P450c17, and GAPDH (see Table I).
Full-length cDNAs for clones 1, 9, and 32 were re-cloned into pSG5 (39) for transfection of mammalian cells. The human P450scc constructs 1xwtϪ155/Ϫ131/TK32LUC and 1xmtϪ155/Ϫ131/TK32LUC were described previously (25). As an internal control, 100 ng of pRL-CMV vector (Promega) was co-transfected for each well. Luciferase assays were performed with Dual-Luciferase™ assay system (Promega)  Preparation of Protein Extracts-Nuclear extracts were prepared from JEG-3 cells and NCI-H295A cells as described (24). Protein concentrations were determined by Bradford assay. Yeast transformed with HIS multicopy vector YEp90 expressing the cDNA for clones 1, 9, or 32 were grown in 50 ml of SD selective media overnight at 30°C, transferred into 300 ml of YPD media, and incubated for 3 h at 30°C (40). Cell pellets were collected by centrifugation, washed once with H 2 O, and then resuspended in 1.5 ml of 20 mM Tris-HCl, pH 7.5, 20% glycerol, 0.4 M KCl, 2 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride with protease inhibitors (Roche Molecular Biochemicals). The yeast were vortexed vigorously for 2 min with an equal volume of 450 -600-m glass beads (Sigma), and the debris was pelleted at 4°C for 15 min at 13,000 rpm. The protein concentration of the supernatant was determined, and aliquots of these yeast extracts were stored at Ϫ70°C.
Electrophoretic Mobility Shift Assays-Double-stranded wild-type and mutant probes corresponding to bases Ϫ155/Ϫ131 of the human P450scc promoter (see Table I) were end-labeled by [␥-32 P]ATP (Amersham Pharmacia Biotech) and T4 polynucleotide kinase (New England BioLabs) and purified through a G-50 column. About 10 fmol of labeled probe (20,000 cpm) were incubated at room temperature for 15 min with 5-8 g of nuclear extracts or yeast cell extracts and various competitor oligonucleotides in a final volume of 15 l in 10 mM Tris-HCl, pH 7.5, 100 mM KCl, 1 M EDTA, 4% glycerol, 5 mM dithiothreitol, 0.1 mg/ml bovine serum albumin with 1 g of poly(dI-dC) added as a nonspecific competitor. For supershift experiments, 1 l of undiluted rabbit antihuman LBP-1 antiserum was added following formation of the DNA⅐protein complex and was incubated for an additional 15 min at room temperature. DNA⅐protein complexes were analyzed by electrophoresis in 4% native polyacrylamide gel in 50 mM Tris base, 38 mM glycine, 2 mM EDTA, pH 8.0, and 0.35 l of ␤-mercaptoethanol at 20°C for 90 min at 230 volts.

Construction and Screening of a Yeast One-hybrid Library-
Our previous work demonstrated that the DNA segment between Ϫ155 and Ϫ131 was required for placental, but not for  (24,25,30,36). Furthermore, the corresponding Ϫ155/Ϫ131 oligonucleotide formed complexes with nuclear extracts from human placental JEG-3 cells that were not seen with nuclear extracts from human adrenal NCI-H295 cells (25). Therefore we sought to clone JEG-3 cell factors that bound to this DNA. Construction of JEG-3 cell cDNA expression libraries and screening 6 ϫ 10 6 unique gt11 clones with radiolabeled Ϫ155/Ϫ131 dimers failed to identify specific clones. 2 Therefore we used the yeast one-hybrid system (41) to search for proteins that bind to this DNA.
Yeast strain YM4271 was genetically modified to contain selectable HIS3 and LacZ reporter genes under the control of four tandem copies of the Ϫ155/Ϫ131 element of P450scc. A human placental cDNA library containing 2.4 ϫ 10 6 unique clones was constructed and propagated in E. coli, and the plasmid DNA was transformed into the YM4271 reporter strain. Of approximately 9 ϫ 10 7 transformants plated onto selective media, 108 robust, apparently Leu ϩ -His ϩ -Ura ϩ colonies were identified. After further propagation, 52 plasmids showed ␤-galactosidase activity, and 36 plasmids contained identifiable cDNA inserts. Most clones appeared to be unique on the basis of insert size and endonuclease digestion pattern; however, clone NH1 was present in 7 copies, clone NH9 in 6 copies, and clone NH32 in 2 copies. All clones were subjected to at least one sequencing run from a plasmid primer. BLAST searches showed that none of the unique copy clones bore identifiable relationships to any known transcription factors, but the multicopy clones NH1, 9, and 32 were all related to LBP-1; these clones were then sequenced in their entirety (Fig. 1).
Sequences of LBP-related Clones-The insert in clone NH1 was 3843 bp long, had a long 3Ј-untranslated region, and encoded a protein of 540 amino acids with a predicted molecular weight of 62 kDa. The encoded amino acid sequence of NH1 was 98% identical to human LBP-lb. Because the differences between the published LBP-1b amino acid sequence and the sequence encoded by clone NH1 were mainly confined to one region, we considered whether these were alternately spliced products of the same gene. Therefore, we used RT-PCR to amplify the cDNA regions that were different between NH1 and LBP-1b, using total RNA from JEG-3 and NCI-H295A cells and primers flanking the variant region. Sequencing of two JEG-3 and two NCI-H295A clones showed the sequence found in the NH1 cDNA, suggesting that the NH1 sequence is the correct LBP-1b sequence and that the reported sequence of LBP-1b (26) has a short, frameshifted, incorrect amino acid sequence (Fig. 2). Hence we refer to the NH1 cDNA sequence as LBP-1b.
The insert in clone NH9 was 4777 bp long, had a long 3Јuntranslated region, and had an open reading frame of 1395 bp without an initiating methionine in a Kozak sequence. Therefore we performed 5Ј-RACE (42) using JEG-3 cell RNA reversetranscribed into cDNA using the specific primer 9-GSP1 (Table  I) followed by nested PCR. The 591-bp double-stranded RACE/ PCR product was cloned and sequenced, showing complete sequence overlap with the 5Ј-end of NH9, indicating that it was an accurate RACE product. After appropriate ligation of this RACE product, the full-length clone NH9 was 4909 bp long and encoded a protein of 479 amino acids with a predicted molecular weight of 55 kDa that had 83% amino acid sequence identity to our corrected sequence of LBP-1b and hence is hereafter termed LBP-9.
The insert in clone NH32 also appeared to be less than full-length, having an open reading frame of only 1230 bp and lacking an initiating methionine codon in an appropriate Kozak sequence. This clone was also completed by 5Ј-RACE yielding a sequence of 2250 bp that lacked a complete 3Ј-untranslated region but contained an open reading frame for 618 amino acids whose sequence was distantly related to human LBP-1b and to the Drosophila transcription factor E1f-1/NTF-1 (43,44). This protein has a predicted molecular mass of 68 kDa that is hereafter designated LBP-32. No previously defined DNA binding motifs, such as a zinc finger, leucine zipper, helix-loophelix, or homeobox was identified in any of our three clones.
Tissue Distribution of Expression of LBP-1b, -9, and -32-Our previous work indicated that the placental expression of human P450scc involved some transcription factors found in both the adrenals and placenta and some transcription factors found in the placenta but not in the adrenals (24,25). Therefore we sought to determine whether the expression of clones LBP-1b, -9, and -32 was unique to the placenta or occurred in a broader array of human cell types. The highly sensitive procedure of RT-PCR followed by Southern blotting detected LBP-1b expression in steroidogenic human placental JEG-3 cells and adrenal NCI-H295A cells, in human adrenal tissue, in nonsteroidogenic human liver HepG2 cells, human cervical carcinoma HeLa cells, and monkey kidney COS-1 cells (Fig. 3). By contrast, LBP-9 and -32 were expressed abundantly in placen- tal JEG-3 cells, at very low levels in non-steroidogenic cells, and were not detected in human adrenal NCI-H295A cells or in human adrenal tissue. P450c17, a steroidogenic enzyme expressed in the adrenals, gonads, and brain but not the placenta (13, 15) was found in NCI-H295A cells and human adrenal tissue as predicted but not in JEG-3 cells or the non-steroido-genic cell lines, demonstrating the specificity of this RT-PCR experiment.
Binding of LBP-1b, -9, and -32 to the P450scc Promoter-The clones expressing LBP-1b, -9, and -32 had been identified through trans-activation of selectable markers, presumably through specific binding to the incorporated tetramer of the Ϫ155/Ϫ131 sequence of the human P450scc promoter. Therefore we sought to determine whether the proteins encoded by these three clones would bind this DNA in vitro. The inserts of each clone were sub-cloned into the multicopy vector YEp90, expressed in yeast, and yeast protein extracts were prepared; these yeast extracts and JEG-3 cell nuclear extracts were then used in electrophoretic mobility shift assays. As shown in Fig.  4, JEG-3 nuclear extracts created two protein⅐DNA complexes with end-labeled Ϫ155/Ϫ131 double-stranded DNA. Complex B appeared to be nonspecific, as it was not competed by a 100-fold molar excess of unlabeled oligonucleotide; by contrast, complex A was readily competed by a 100-fold excess of unlabeled probe.  3. Tissue distribution of expression of LBP-1b, -9, and -32. Random primers were used to prepare cDNA from 1 g of total RNA from each cell line or tissue as indicated, and 5 l of the cDNA were then amplified for 30 cycles to yield fragments of 690, 862, 856, and 820 bp, respectively, for LBP-1b, -9, and -32 and human P450c17; a similar procedure was used to amplify a 502-bp GAPDH cDNA fragment as control. The PCR products were resolved on 1% agarose gel, blotted, and probed with the corresponding LBP-1b, -9, and -32, human P450c17, and GAPDH cDNA as indicated. Only 1 ⁄25 of the cDNA product was loaded. Negative control: 5 l of cDNA synthesis reaction without reverse transcriptase was used as template for PCR amplification. Positive control: 5 ng of plasmid DNA for LBP-1b, -9, and -32, P450c17, or GADPH, respectively, was used as template for PCR amplification.

FIG. 4. Bandshift and antibody supershift experiments.
JEG-3 cell nuclear extract or yeast-expressed LBP-9 or LBP-1b were incubated with radiolabeled Ϫ155/Ϫ131 double-stranded probe in the absence (-) or presence (ϩ) of a 100-fold molar excess of the unlabeled probe as a competitor. LBP-1 antibody was added to the protein⅐DNA complex and incubated for an additional 15 min at 25°C. LBP-9 appears to be responsible for JEG-3 cell complex A, but complex C formed by yeastexpressed LBP-1b was not detected in JEG-3 cells in this experiment. Bkg, background.
When the protein⅐DNA complexes were incubated with rabbit antiserum to human LBP-1a/b (generously provided by Dr. R. Roeder), the anti LBP-1 inhibited the formation of complex A but not complex B. Thus JEG-3 nuclear extract contains a protein that binds to Ϫ155/Ϫ131 and is immunologically related to LBP-1.
Yeast extracts containing the LBP-1b and LBP-9 proteins, but not LBP-32, also formed complexes with the doublestranded Ϫ155/Ϫ131 DNA (Fig. 4). Yeast-expressed LBP-1b forms complex C, which is competed by cold probe and inhibited by the anti LBP-1 antiserum, similarly to complex A formed by the JEG-3 nuclear extract. However, the mobility of complex C was clearly different from that of complex A, suggesting that LBP-1b does not form JEG-3 cell complex A. The absence of a band corresponding to complex C in JEG-3 nuclear extracts could mean that LBP-1b is of low abundance in JEG-3 cells or that LBP-1b has weak affinity for Ϫ155/Ϫ131. The complex formed by LBP-9 had the same apparent mobility as JEG-3 complex A, was inhibited by excess cold probe, and unlike the other complexes, was supershifted by the antiserum to LBP-1. Thus LBP-9 may be the protein generating complex A, but the difference between inhibition of complex A formation and supershifting of the LBP-9 complex may mean that complex A is not formed by the LBP-9 protein. Yeast extracts containing LBP-32 did not yield detectable complex formation with the Ϫ155/Ϫ131 oligonucleotide (not shown), hence LBP-32 was not considered further.
Regulation of the P450scc Promoter by LBP-1b and LBP-9-To assess the potential roles of LBP-1b and LBP-9 in the regulation of human P450scc gene transcription, we assessed the capacity of mammalian expression vectors for these two proteins to transactivate a single copy of the Ϫ151/Ϫ131 sequence linked to the minimal 32-base promoter of the thymidine kinase gene (TK32) fused to the luciferase reporter transfected into JEG-3 cells. The Ϫ155/Ϫ131 sequence increased TK32LUC activity 3.4-fold, but mutation of 10 bases in the Ϫ155/Ϫ131 sequence reduced this to a 2-fold increase (Fig. 5A).
To determine whether LBP-1b or LBP-9 influenced this basal activity, we co-transfected JEG-3 cells with the Ϫ155/Ϫ131/ TK32LUC reporter and vectors expressing either LBP-1b or LBP-9. LBP-1b increased the activity of the wild-type Ϫ155/ Ϫ131 sequence 21-fold but had no effect on the 10-base mutant of Ϫ155/Ϫ131, whereas LBP-9 had no apparent effect on either the wild-type or mutant Ϫ155/131 sequence (Fig. 5A). When the Ϫ155/Ϫ131 sequence fused to TK32LUC was co-transfected with increasing amounts of the vector expressing LBP-1b, transcription was increased in a dose-dependent manner (Fig. 5B). Thus LBP-1b had a clear stimulatory effect on transcription fostered by the Ϫ155/Ϫ131 sequence of human P450scc. Because the basal transcription from the Ϫ155/Ϫ131/TK32LUC construction was low, it was not clear whether LBP-9 had no effect or exerted a suppressive effect. Therefore, we examined the effect of LBP-9 on LBP-1b-induced transcription from Ϫ155/Ϫ131/TK32LUC in JEG-3 cells. Co-transfection of the Ϫ155/Ϫ131/TK32LUC reporter construct with 500 ng of the vector for LBP-1b and with increasing amounts of the vector for LBP-9 showed that LBP-9 suppressed the LBP-1b-induced activation of LUC expression in a dose-dependent fashion (Fig.  5C). Thus, LBP-9 appears to be a transcriptional suppressor, and, in the amounts of protein expressed by our pSG5-based vectors, the suppressive action of LBP-9 appears to override the activating action of LBP-1b. DISCUSSION The orphan nuclear receptor SF-1 is required for the production of steroid hormones in the adrenals and gonads but not in the placenta, brain, or other "extra-glandular" tissues. SF-1independent transcription of P450scc has been demonstrated in the human placenta (24) and rat brain (23), prompting a search for factors that can substitute for the essential role of SF-1. Some candidate factors have been identified. SF-1-independent transcription of the rat gene for steroid 17␣-hydroxylase (P450c17) can be regulated by two factors operationally termed StF-IT-1 and StF-IT-2 (45) and by ku autoantigen (46). StF-IT-1 has recently been identified as the oncoprotein SET (47), a factor not previously known to be a transcriptional regulator. Thus it appears that an unexpectedly broad array of proteins can regulate the transcription of the genes for the steroidogenic enzymes. We have now added factors related to the LBP group of transcription factors to this growing family.
The LBP family of transcription factors was initially characterized as a single cellular factor that bound to two different sites in the HIV, type I promoter (48 -51). Later work showed that there are two related LBP genes, each of which encodes Luciferase activity is expressed relative to the TK32LUC construct without additional upstream sequences. Values represent the average of two independent experiments, each performed in triplicate after normalization for the internal control (renilla luciferase activity of pRL-CMV). Left column, activity of the wild-type or mutant Ϫ155/Ϫ131/TK32LUC co-transfected with 250 ng of empty pSG5 vector; middle column, co-transfection with 250 ng of pSG5-based vector expressing LBP-1b; right column, co-transfection with 250 ng of pSG5 vector expressing LBP-9. 100 ng of pRL-CMV (Promega) was included in each transfection as internal control. B, LBP-1b stimulates P450scc Ϫ155/Ϫ131 transcriptional activity in a dose-dependent manner. 1 g of Ϫ155/Ϫ131/TK32 reporter was co-transfected with increasing amounts of pSG5-based vector expressing LBP-1b as shown. C, LBP-9 suppresses the LBP-1b-stimulated transcriptional activity of P450scc Ϫ155/Ϫ131. 500 ng of vector expressing LBP-1b and increasing amounts of vector expressing LBP-9 (5, 50, and 500 ng as indicated) were co-transfected together with 1 g of Ϫ155/Ϫ131/TK32LUC reporter. LUC activity is expressed relative to the TK32LUC construct without additional upstream sequences. Values represent the average of two experiments, each performed in triplicate, after normalization for the internal control (renilla luciferase activity of pRL-CMV). two alternately spliced transcripts, so that LBP-1a and LBP-1b arise from one gene, and LBP-1c and LBP-1d arise from a second gene (26). LBP-1c is identical to the ␣-globin transcription factor CP2 (52), and proteins in the LBP family are all related to Elf-1/NTF-1, which is essential for Drosophila embryogenesis (43,44). Thus the LBP proteins represent an evolutionary ancient family of transcription factors that participate in development. Sequences related to retroviruses are found throughout the human genome (53,54) and frequently regulate expression of adjacent cellular genes, especially those expressed in the placenta (55,56). We find no evidence for the insertion of a retroviral regulatory sequence as has been described for placental expression of human pleiotropin (57), although sequences similar to Ϫ155/Ϫ131 are not found in the bovine or rat P450scc promoters. Hence, whereas it may not be surprising to find LBP-related proteins participating in transcriptional regulation in the placenta, to our knowledge this is the first report of an action of LBP-1b on an endogenous, non-viral promoter or of any LBP-related factor participating in the transcriptional regulation of a gene involved in steroid hormone biosynthesis.
Our mobility shift data show that LBP-1b and LBP-9 bind to the Ϫ151/Ϫ131 segment of the human P450scc promoter. In comparison with our previous studies of the P450scc promoter in JEG-3 cells, it appears that LBP-9 forms what was previously called Complex IV, LBP-1b forms a previously undetected complex, and neither LBP-1b or LBP-9 forms what was previously called Complex VII (25). Our previous data suggested that a 55-kDa protein forming Complex VII was required for basal, placental-specific expression of P450scc, whereas complex IV appeared to be involved in modulating P450scc expression (25). Our present data are consistent with those earlier observations, suggesting that LBP-1b and LBP-9, respectively, amplify or diminish the level of expression initiated by other factors. This action of LBP-related proteins to function as quantitative "volume controls" rather than as basal "on/off switches" is consistent with their similar action to modulate the level of HIV, type I transcription and is similar to the reciprocal action of Sp1 and Sp3 on numerous genes (58,59) including adrenal expression of bovine P450scc (60). Thus LBP proteins are important newly described modulators of placental P450scc expression, but other, as yet uncharacterized proteins are probably required for basal placental-specific expression of this gene.