| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Obstetrics, Gynecology, and Reproductive Sciences (S.H.M., S.R.B., C.D., J.-L.V., P.B.B.), The Center for Reproductive Sciences, University of California, San Francisco, California 94143; and Department of Obstetrics and Gynecology, and Center for Research on Reproduction and Women’s Health (N.B.H.), University of Pennsylvania, Philadelphia, Pennsylvania 19104
Address all correspondence and requests for reprints to: Synthia H. Mellon, Ph.D., Department of Obstetrics, Gynecology, and Reproductive Sciences, University of California, San Francisco, 513 Parnassus Avenue, Box 0556, San Francisco, California 94143-0556. E-mail: mellon{at}cgl.ucsf.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
-hydroxylase, producing 17-hydroxy C21 steroids and C17,20 lyase, producing C19 steroids. These two activities are regulated by posttranslational modification (1, 2) and by allosteric factors (3, 4); hence, steroid production is regulated both transcriptionally and posttranslationally (reviewed in Ref. 5). Hydroxylase and lyase activities are necessary for the development and maintenance of reproductive function (5), and appear to be important for production of neurosteroids that regulate neural differentiation and function (6) and for early embryonic development (7). Transcriptional regulation of P450c17 is both species-specific and tissue-specific, and involves transcription factors that vary among mammalian species and among steroidogenic tissues (8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20). P450c17 is expressed in the gonads of all vertebrates, in the adrenals of primates (21) and ungulates (22), but not rodents (23, 24), and it is expressed in placenta of rodents (25, 26), but not primates (21). P450c17 expression is also regulated developmentally. The mammalian fetus expresses P450c17 in the testis but not the ovary (10, 21, 27), and it is required for male sexual differentiation (21, 28). It is also expressed transiently in the neonatal rat liver (29) in the developing rodent nervous system (30). In the mouse embryo, its expression appears to be essential for development beyond embryonic day 7 (7), although human beings lacking P450c17 activity develop to term.
P450c17 of all species is regulated by steroidogenic factor-1 (SF-1) in those steroidogenic tissues that express SF-1 (12, 13, 14, 18, 31, 32, 33, 34). Rat P450c17 transcription is also regulated by developmentally regulated factors (9, 10), by tissue-specific factors (11, 12), and by multiple orphan nuclear DNA-binding receptors (12, 13, 14).
We have identified and characterized several transcriptionally regulated regions in the rat P450c17 gene in detail (9, 12, 13, 14). We found that the region of the rat P450c17 gene at –447/–419 is regulated by SF-1 together with two novel factors we termed steroidogenic factor-inducer of transcription-1 and -2 (StF-IT-1 and StF-IT-2) (12, 35). We subsequently identified StF-IT-1 as a phosphoprotein called SET, which inhibits cell cycle (36, 37) and is an inhibitor of protein phosphatase 2A (2, 38) in addition to regulating transcription (9, 10). We have now identified StF-IT-2 as the protein known as testis brain RNA-binding protein (TB-RBP) (39, 40) also known as translin (41), and we now show that translin activates SF-1-mediated transcription in a DNA-specific manner.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Our established protocol for purification of StF-IT-2 began with fast protein liquid chromatography of MA-10 whole-cell extracts; fractions containing StF-IT-2 were identified by EMSAs (gel shift assays), combined, and purified further by HPLC on a MonoQ column (Fig. 1
). SDS-PAGE showed that the fractions containing StF-IT-2 DNA binding activity were enriched for a protein of approximately 25 kDa. The approximately 25,000 molecular mass protein from fraction 29, which contained the majority of the StF-IT-2 binding activity, was excised, digested with trypsin, purified by HPLC, and analyzed by tandem mass spectrometry (MS/MS) (Table 1). The sample contained three different proteins, each of which could be identified by three or more peptides. Two were proteins of approximately 25 kDa: translin (26.2 kDa) and 14-3-3
(27.7 kDa); the other was nexin 2 (58 kDa), which was not considered further, because its molecular mass did not match that identified by SDS-PAGE. The peptide masses identified by mass spectrometry encompassed 37% coverage of the sequence of translin and 24% coverage of the sequence of 14-3-3 but only 7% of nexin 2.
|
|
The family of 14-3-3 proteins are approximately 30-kDa acidic proteins, so-named because of their purification profile (55, 56, 57). They are a highly conserved, ubiquitously expressed protein family composed of at least seven isoforms. They bind to phosphoserine and phosphothreonine (58) and hence can interact with a wide variety of proteins yielding many different functions (59, 60, 61, 62, 63). These functions include modulating enzymatic activity, modulating protein localization, preventing dephosphorylation, promoting protein stability, and inhibiting and mediating protein interactions.
Translin Antibodies Supershift StF-IT-2:DNA Complex
To determine whether translin or 14-3-3 were StF-IT-2, we performed gel shift assays using antibodies against each of these proteins (Fig. 2). Using the –447/–419 rP450c17 oligonucleotide as probe, MA-10 cell extracts produced two protein:DNA complexes. The more rapidly migrating complex was an SF-1:DNA complex, identified previously by antibody supershift assays (35) and competition assays (12), and the more slowly migrating complex was StF-IT-2:DNA. Addition of an antibody against mouse translin/TB-RBP caused a specific supershift of the St-F-IT-2:DNA complex obtained from MA-10 cell protein (Fig. 2, lane 4) and from fraction 28 of the MonoQ column (Fig. 2, lane 7). Addition of preimmune serum (Fig. 2, lane 5), or antibodies against either 14-3-3 or 14-3-3ß (Fig. 2, lanes 11 and 12) did not alter the migration of this complex. Immunodepletion of MA-10 extracts with the antibody against translin, but not with preimmune serum, decreased the intensity of the StF-IT-2:DNA complex (data not shown). Thus, the StF-IT-2:DNA complex contains translin and does not contain 14-3-3 or 14-3-3ß.
|
, lane –447/–419). Cotransfection with a translin expression construct did not increase basal activity, whereas cotransfection with an SF-1-expression construct (65) increased luciferase activity 4-fold. However, when the –447/–419-luciferase reporter construct was transfected with both the SF-1 and the translin expression vectors, luciferase activity increased 10-fold. Thus, translin has no effect by itself, but increases SF-1-mediated transcription about 2- to 3-fold.
|
). In comparison with the wild-type –447/–419 luciferase reporter construct that could be stimulated approximately 4-fold by SF-1, Mut 1 and Mut 2 had low basal activity that could be stimulated about 1.5- to 2-fold by SF-1 and Mut 1/2 could not be stimulated by SF-1 (Fig. 3B). Translin increased SF-1-mediated transcription from either the Mut 1 or Mut 2 constructs about 2-fold, similar to the stimulation that translin elicited from the wild-type construct. However, when Mut 1/2 was used as reporter, SF-1 could not stimulate transcription with or without translin. Thus, the action of translin requires only a single SF-1 site and SF-1 binding and stimulation are required for translin action. Hence, translin appears to act as an SF-1 coactivator, rather than as a transcription factor.
Translin Binds to Single-Stranded DNA
Translin can bind to single-stranded RNA (39, 47, 48, 49) and to single-stranded DNA sequences found at breakpoints of chromosomal translocations (41). To determine whether translin bound to the –447/–419 region of the rat P450c17 gene as single- and/or double-stranded DNA, we assessed the binding of translin prepared by an in vitro transcription/translation system to labeled sense and antisense oligonucleotides and to the annealed double-stranded –447/–419 oligonucleotide. Gel shift assays showed that translin could bind to the gel-purified, double-stranded oligonucleotide (Fig. 4A, DS Kinase), but the degree of binding was much greater with either single-stranded oligonucleotide probe (Fig. 4A; only single-stranded sense strand probe, SS-S, is shown). These data suggest that translin can bind to either double- or single-stranded DNA. However, these experiments were done with probes labeled with [
-32P]ATP and polynucleotide kinase, and such kinased probes may contain a small amount of single-stranded DNA. Therefore, we also synthesized a double-stranded probe that was labeled using [-32P]dCTP and the Klenow fragment of DNA polymerase, which cannot label a single-stranded probe. Incubation of this probe with translin did not yield any protein:DNA complex (Fig. 4A, DS Klenow). By contrast, when the double-stranded, Klenow end-labeled probe was incubated with MA-10 extracts that contain SF-1 and translin, only SF-1 yielded its protein:DNA complex, indicating that SF-1 could bind to this double-stranded probe, as expected, whereas translin could not (Fig. 4B). These data suggest that translin only binds to single-stranded DNA.
|
). Our data show that 250-fold excess of both the sense and antisense strands of –447/–419 oligonucleotide can compete with the labeled –447/–419 sense strand for translin binding.
Because this –447/–419 oligonucleotide contains SF-1 or estrogen response element half-site-like binding sites, we used additional oligonucleotides that contained previously identified SF-1 or similar binding elements as competitors. These oligonucleotides were the –418/–399 and –84/–55 regions of the rat P450c17 gene, as well as the –95/–58 region of the mouse P450c21 gene. Our gel shift data indicate that both the sense and antisense strands of each of these oligonucleotides could compete with the –447/–419 sense strand for translin binding (Fig. 4C).
As competitors for translin binding, we also used –447/–419 oligonucleotides in which the SF-1 sites were mutated (Fig. 4D). In the oligonucleotides, listed in Table 2, each SF-1 site was mutated independently (Mut 1, Mut 2), or together (Mut 1/2). Competition gel shifts demonstrate that, when of either of the SF-1 sites within the –447/–419 oligonucleotide were mutated, the resulting sense and antisense oligonucleotides were able to compete with wild-type –447/–419 sense oligonucleotide for translin binding. In addition, oligonucleotides that contained mutations of both SF-1 sites also competed for translin binding.
|
and named Prm-2, GS-1, and Bcl-Cl1, to assess translin binding. We confirmed that translin binds to these single-stranded DNA sequences, as others had shown previously (data not shown). Thus, whereas other laboratories have described a consensus translin binding site that may resemble an SF-1 site (41, 68), other DNA sequences that do not resemble SF-1 sites are also bound by translin. Hence we did not identify specific single-stranded DNA sequences to which translin bound.
Translin’s Activity Does Not Require DNA Binding
To determine whether translin needed to bind to DNA to elicit its effects, we created a previously described variant of translin containing three mutations, R86T, H88N, and H90N; this form of translin cannot bind to nucleic acid (54). We expressed recombinant wild-type and mutant translin in bacteria, and used these proteins in gel shift assays to determine whether mutant translin could bind to the –447/–419 DNA. Gel shift assays confirmed that the mutant translin could not bind to DNA (Fig. 5A). To determine whether the lack of DNA binding affected translin’s ability to stimulate transcription, we subcloned the wild-type and mutant translins into a mammalian expression vector. We cotransfected HeLa cells with either of these translin-expression vectors together with the –447/–419 luciferase reporter and the SF-1 expression vector. Luciferase assays demonstrated that, whereas neither the wild-type nor mutant translin could stimulate P450c17 transcription alone, both the wild-type and mutant translins increased SF-1-stimulated transcription about 2- to 3-fold (Fig. 5B). Taken together, the data suggest that translin does not need to bind to DNA to increase SF-1-stimulated transcription.
|
) demonstrate that increasing the amount of translin has no effect on SF-1 binding to DNA. These data suggest that translin does not alter SF-1 binding to DNA.
|
). When [35S]translin, synthesized in vitro, was incubated with GST-translin bound to a glutathione-Sepharose 4B matrix, the [35S]translin was retained on the matrix, indicating that translin could bind to itself (Fig. 7A), consistent with previous data showing that translin forms multimers (44). The signal for labeled translin was relatively weak because translin contains a single methionine residue. We then prepared [35S]SF-1 in vitro and incubated it with GST-translin bound to the glutathione-Sepharose 4B matrix (Fig. 7B). Whereas a small amount of [35S]SF-1 was retained by the matrix that contains only GST, there was a substantially greater amount of [35S]SF-1 retained by the matrix containing GST-translin. Thus, SF-1 binds directly to translin. Although these pull-down assays show an interaction between SF-1 and translin, this interaction was not detected by gel shift analysis (Fig. 6); such receptor:coactivator:DNA interactions are typically undetected by gel shift analyses.
|
). These SF-1 mutants are shown in cartoon form on the left panel of Fig. 7B. As shown in Fig. 7A, full-length SF-1 binds to GST-translin. When we deleted the carboxyl-terminal activation function-2 (AF2) domain (
AF2, amino acids 451–462 deleted from full-length SF-1) or the ligand binding domain (LBD, amino acids 225–462 deleted from full-length SF-1), the remaining SF-1 fragments could still bind to GST-translin. Thus, a region of SF-1 between amino acids 1 and 225 binds to translin. The ligand binding/AF2 domains alone (LBD, amino acids 225–462) do not bind to GST-translin. However, when amino acids 79–225 comprising the hinge region are added to the LBD construct (DBD), binding to translin is restored. Thus, the GST-pull-down experiments suggest that the hinge region of SF-1 encompassing amino acids 79–225 is necessary for translin binding. We expressed amino acids 79–225, the hinge region of SF-1 alone, in the absence of any additional regions of SF-1. This peptide, however, did not bind GST-translin. Because this is only a small SF-1 peptide, the lack of binding to GST-translin could have been due to a lack of proper three-dimensional structure. Hence, we created additional full-length SF-1 mutants that lacked only specific segments of the hinge region, to test directly their ability to be activated by translin in cell culture.
To determine whether the hinge region of SF-1 is important for translin activation, we created three SF-1 mutants: two SF-1 mutants deleted the proline-rich region and lacked either amino acids 101–169 (SF-1 101–169) or lacked amino acids 101–225 (SF-1 101–225), and one mutant contained this proline-rich region but lacked amino acids 170–225 (SF-1 170–225). Each of these constructs was transfected into HeLa cells with the –447/–419 luciferase reporter, plus or minus the translin expression vector (Fig. 7C). Transfection data demonstrate that both the SF-1 101–169 and SF-1 101–225 could not activate transcription of the –447/–419 luciferase reporter, suggesting that the proline-rich region of SF-1 is required for SF-1’s transcriptional activity. However SF-1 170–225 could stimulate transcription from the –447/–419 luciferase reporter, although this stimulation was about half of that seen with the full-length SF-1. Translin cotransfected with SF-1 170–225 could not activate transcription of the –447/–419 luciferase reporter. Translin also could not activate the SF-1 mutants lacking the proline-rich region. These data suggest that the hinge region of SF-1 between amino acids 170 and 225 are important for translin’s ability to augment SF-1 transcriptional activation.
The 170–225 region of the SF-1 hinge contains serine 203 that has been shown to be important for SF-1’s transcriptional activity (69). We created a seine 203 to alanine mutant of SF-1 (Ser 203 Mut) in our SF-1-expression vector, and transfected this mutant construct together with the –447/–419 luciferase reporter construct into HeLa cells (Fig. 7D). Ser 203 Mut SF-1 could stimulate luciferase activity from the P450c17 reporter construct about 5-fold (Fig. 7D). Translin could also increase this SF-1-stimulated transcription 2-fold further. These data demonstrate that serine 203 does not play a role in mediating the effects of translin.
Translin Activation of SF-1-Stimulated Transcription Is Dependent upon DNA Sequence
To determine whether translin augments the transcription of other genes that are regulated by SF-1, we cloned several different SF-1-responsive elements into the p36Luc reporter construct. These included another region of the rat P450c17 gene, –84/–55 (13, 14), the rat P450aro gene, –87/–68 (70, 71, 72), and the mouse P450c21 gene, –95/–58 (65, 73). These three elements are proximal SF-1 regulatory elements, and each is located within 100 bp of the TATA box of each gene. When these luciferase reporter constructs were transfected into HeLa cells, they had very low basal activity (Fig. 8A). Cotransfection of these reporter constructs with an SF-1-expression construct increased luciferase activity 4-, 13-, and 4-fold, respectively. When we cotransfected the reporter constructs containing the SF-1 binding domains from the rat P450c17 –84/–55, rat P450aro, or mouse P450c21 with SF-1 and translin expression vectors, luciferase activity was not significantly different from the activity obtained with the SF-1 expression vector alone (Fig. 8A). Transfection with the luciferase reporter constructs plus translin gave luciferase activities no different from basal activity. Thus, translin had little effect on SF-1-stimulated transcription of these three SF-1-responsive elements.
|
). However, cotransfection with SF-1 and translin expression vectors did not significantly increase luciferase activity compared with that obtained with the SF-1 expression vector alone. Thus, the presence of two SF-1 elements does not necessarily create a translin-responsive site. We also used a concatamer of the P450c21 element containing five repeated SF-1-responsive elements, which had been used originally to demonstrate SF-1-dependent transcription (65, 73). This DNA was cloned into p36Luc, and transfected into HeLa cells. SF-1 stimulated luciferase activity 200-fold, substantially greater than the 4-fold stimulation obtained when a single SF-1 site was used. Furthermore, whereas translin alone had no effect, cotransfection of translin and SF-1 expression vectors further augmented SF-1-stimulated transcription 2.5-fold. Therefore, translin appears to discriminate among SF-1-response elements, and only activates transcription from a subset of these elements.
Translin-Associated Factor X (TRAX) Also Augments SF-1-Dependent Transcription
TRAX is a 33-kDa protein that is 28% identical with translin and has been shown to interact with translin (66, 74, 75, 76, 77, 78). Both translin and TRAX have putative leucine zipper regions, found in the carboxyl terminus of translin and in the mid-region of TRAX (41, 46). Although these two proteins are known to interact, TRAX, unlike translin (79), does not form homomultimers. TRAX alone does not bind directly to nucleic acids, although heterodimers of translin and TRAX can bind to DNA (76). Although we did not identify TRAX in our MS/MS analysis of StF-IT-2, we determined whether TRAX was found in MA-10 cells, whether TRAX could also be part of the StF-IT-2:DNA complex, and whether TRAX could augment transcription.
To determine whether TRAX was part of the StF-IT-2:DNA complex, the –447/–419 oligonucleotide probe was incubated with MA-10 cell extracts in the presence of anti-TRAX antibodies (Fig. 9A). TRAX antibodies, but not preimmune serum, supershifted the protein:DNA complex that was identified as StF-IT-2, in a manner similar to that seen when the translin antibody was used (Fig. 2B). However, TRAX does not bind directly to the –447/–419 DNA. Unlike translin, which binds to the –447/–419 DNA, recombinant TRAX prepared in bacteria does not bind to this DNA (Fig. 9B). Thus, because the TRAX antibody could supershift the StF-IT-2:DNA complex, the data suggest that translin and TRAX probably form heteromers in MA-10 cells, and these heteromers bind to DNA.
|
). As shown previously for translin, using a translin expression vector (Fig. 3A), TRAX (0.15 µg) had no effect on transcription from the wild-type –447/–419 luciferase reporter construct. When SF-1 was cotransfected, TRAX (0.15 µg) had only a slight effect on SF-1 stimulated transcription. HeLa cells were transfected with SF-1 and an equal mixture of TRAX (0.075 µg) and translin (0.075 µg). In this transfection (Fig. 9C, eighth bar on graph, +SF-1/(Translin/Trax)/2), the total amount of translin and TRAX together equaled the amount of translin or TRAX added individually (shown in the sixth and seventh bars, +SF-1/Translin; +SF-1/Trax). Transfection of this equal mixture of translin and TRAX resulted in a 2-fold activation of SF-1-stimulated transcription, similar to that seen with translin alone plus SF-1. When 0.15 µg of TRAX plasmid was transfected with 0.15 µg of translin plasmid, SF-1-stimlated transcription was now stimulated 4-fold (ninth bar in graph, +SF-1/Translin/Trax). Thus, TRAX appears to work together with translin to activate SF-1-stimulated transcription. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Translin is a 228-amino acid protein that has many cellular roles. It is involved in DNA rearrangements through binding to single-stranded DNA sequences at breakpoints (44), cell proliferation (50), mRNA transport and/or stabilization and translational regulation (51, 52), and as a linker protein that binds specific mRNAs to microtubules (39, 53). Translin’s interactions with mRNAs have been found in the nucleus, in cytoplasm, and between male germ cells, implicating translin as a protein that is involved in RNA transport between cells (52). Translin binds - and -tubulins, consistent with a role in microtubule organization, stabilization, and/or chromosome segregation. In neurons, inhibition of translin results in a lack of mRNA sorting for those mRNAs bound by translin, including ligatin and calcium- or calmodulin-dependent protein kinase II (100). The role of nuclear receptor coactivators in RNA processing, as well as in a variety of other activities has recently been reviewed (101). Several splicing-related coactivators such as p68, p72, CAPER, and PGC-1, have RNA binding motifs. One unique coactivator, SRA, is an RNA that may be important structurally for coactivation (102). Because translin is an RNA binding protein, it may likewise act as a coactivator of SF-1 through RNA binding.
The domains of SF-1 that interact with DNA and with other proteins have been identified (69, 80, 81, 84, 91, 103, 104). As with other members of the nuclear hormone receptor family of transcription factors, SF-1 contains a zinc finger motif that is responsible for DNA binding. Immediately downstream from the zinc-finger motif lies a basic region (termed the Ftz-F1 box) that facilitates binding of SF-1 to specific DNA sequences, and which is similar to the A box seen in other nuclear hormone receptors. SF-1 also contains an LBD at the C terminus. Recently, potential phospholipid ligands for SF-1 have been identified through crystallization studies (105, 106, 107). The LBD is usually the dimerization site for those nuclear hormone receptors that act as homodimers or heterodimers, but SF-1 acts as a monomer (104). Transactivation by SF-1 appears to depend on a region in the LBD, termed the AF2 region, and a region consisting of the Ftz-F1 box and an adjacent proline cluster (103). However, our previous data suggest that transactivation by SF-1 within the –447/–419 region of the P450c17 gene does not require the AF2 region (35). In addition to the AF2 region, an N-terminal AF1 region crucial for ligand-independent transactivation by other nuclear hormone receptors is needed for maximal SF-1-mediated transcription and interaction with nuclear receptor cofactors (103). This interaction requires phosphorylation of serine Ser 203 located in the hinge region (69). Our GST-pulldown assays and SF-1 deletional mutants suggest that translin interacts with the hinge region of SF-1. This region is important for stabilizing the ligand binding domain (108) and interacts with other proteins (86). Phosphorylation of Ser 203 in the hinge region enhances the interaction of GRIP1 and SMRT with the AF1 and AF2 regions of SF-1 (69), whereas sumoylation of lysines within the hinge region increase interactions with DEAD-box proteins, and result in transcriptional repression (85, 86). Mutation of serine 203 in our studies did not affect translin’s ability to stimulate SF-1-mediated transcription, indicating that the mechanism by which translin augments SF-1’s activity is different from that of other coactivators like GRIP1 and SMRT.
The mechanism by which translin augments SF-1-stimulated transcription is unclear, but it appears to depend on the DNA sequence to which SF-1 binds. Our previous data suggested that there are sequence-specific differences in SF-1 binding to DNA, and we could assess these differences by the degree to which SF-1 could bend DNA (35). Furthermore, the different putative ligands recently identified for SF-1 may play distinct roles in SF-1 regulation of gene expression, and may also determine the nature of the coactivators/corepressors that interact with SF-1. Recent studies identifying novel ligands for SF-1 have indicated that exogenously added C12–C16 fatty acids increased interaction between SF-1 and TIF2 coactivators (106). Such ligand-dependent differences in coactivator/corepressor recruitment have also been demonstrated for other nuclear hormone receptors (109, 110, 111, 112, 113). Thus, translin may discriminate between the SF-1:DNA interactions that may be mediated by different ligands.
The role of translin in the SF-1-mediated transcription of the rat P450c17 gene is not known. Mice in which the translin gene has been ablated are viable, small at birth, and have behavioral and reproductive abnormalities (114). Translin knockout mice have slightly elevated concentrations of FSH and slightly reduced concentrations of LH, progressively reduced sperm counts in male mice, and reduced fertility and litter size in female mice that are not due to obvious abnormalities in reproductive organs. Plasma or gonadal steroid concentrations were not reported, but are likely to be within normal ranges, because alterations may have been reflected in abnormal gonadotropin concentrations. Gene array analysis of testis mRNA from wild-type and translin knockout mice did not detect differences in expression of steroidogenic enzyme mRNAs, suggesting that changes, if any, would be less than 2-fold (N. Hecht, unpublished data). Recent behavioral analysis of translin knockout mice indicate that these mice exhibit sex-specific differences in learning and memory, locomotor activity, anxiety-related behavior, as well as tonic-clonic seizures induced by handling (115). Some of these behavioral abnormalities may be related to alterations in monoamine neurotransmitter levels in several forebrain regions that are seen in the translin knockout mice, and may be modified by gonadal hormones.
Ablation of translin also causes reduction in TRAX protein but not mRNA, suggesting that these two proteins may stabilize one another (114, 116). By use of small interfering RNA, we were unable to reduce translin protein expression in MA-10 cells to any great extent, suggesting that translin is both abundantly expressed and relatively stable in mouse Leydig MA-10 cells (our unpublished data). Our previous results suggested that SF-1 and translin were coexpressed in the developing gonad and in mouse Leydig MA-10 cells suggesting that the –447/–419 region of the rat P45017 gene could be coactivated by both of these factors during development as well as in the adult gonad (35).
Thus, there are many coactivators and corepressors regulating SF-1 transcriptional activity in a variety of tissues, contributing to the pleiotropic effects of SF-1. Translin and TRAX are new additions to this family.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Purification of StF-IT-2
The S100 fraction was adjusted to 0.5 M ammonium sulfate by adding solid ammonium sulfate under constant stirring. The high salt S100 fraction was then loaded onto a 30-ml phenyl Sepharose 6 fast flow column (Amersham Biosciences, Uppsala, Sweden), which had been previously equilibrated with 50 mM phosphate buffer containing 0.5 M ammonium sulfate (pH 7.0). The protein elution was monitored at 280 nm.
To determine which fractions contained the StF-IT-2 protein, gel shift assays were performed as described previously (13). Fractions containing StF-IT-2 were then pooled and further purified on a 1 ml MonoQ column (Amersham Biosciences). Before loading, the pooled fractions were dialyzed against 10 mM Tris buffer (pH 8.0), and the column was equilibrated in the same buffer. Increasing NaCl concentrations linearly to 0.4 M, in 20 bed volumes of 10 mM Tris (pH 8.0), eluted retained proteins. Eluted proteins were detected at 280 nm. Fractions were analyzed for StF-IT-2 by gel shift assays. Eluted proteins were also characterized by separation on a 10% polyacrylamide gel by electrophoresis followed by a Coomassie blue staining. The approximately 30-kDa proteins from fraction 29 were excised from the polyacrylamide gel, and sent to the Proteomic Core Facility at The Rockefeller University (New York, NY) for analysis. Protein from this band was treated with trypsin, purified by HPLC, and analyzed by MS/MS.
Gel Shift Assays
Gel shift assays were performed as described previously (13). Oligonucleotides corresponding to the –447/–419 region of the rat P450c17 gene promoter and mutants of this region (12) were used in these studies, and are presented in Table 2. Sense and antisense oligonucleotides were mixed in a 1:1 molar ratio, boiled, and reannealed slowly at room temperature before end-labeling using [-32P]ATP and T4 polynucleotide kinase, or labeling with [-32P]dCTP and Klenow DNA polymerase, and mixed with the cellular extracts (150,000 cpm/reaction) in the presence of 100 µg/ml polydeoxyinosinic deoxycytidylic acid, 50 µg/ml salmon sperm DNA, 5 mM dithiothreitol, and 1 mg/ml BSA, and incubated at room temperature for 45 min. For antibody supershift assays, 1–3 µl of antiserum was added to the gel shift reaction mix (75, 118, 119). Antisera were affinity-purified polyclonal rabbit antibodies directed against recombinant mouse TB-RBP (75) or were polyclonal rabbit antibodies directed against the C-terminal amino acids of translin or TRAX (118, 119), or commercially available antibodies against 14-3-3 proteins (Santa Cruz Biotechnology, Santa Cruz, CA). One third of the total reaction was loaded onto a 6% nondenaturing polyacrylamide gel and run in 0.5x Tris-borate-EDTA as running buffer, to separate the free labeled probe from probe bound by protein. The gel was then dried and exposed to a phosphorimager screen or x-ray film at –70 C using an intensifying screen.
Transient Transfections and Reporter Gene Assay
Luciferase reporter constructs were prepared using the minimal mouse prolactin promoter linked to the firefly luciferase gene in the vector p36Luc (64). Double-stranded oligonucleotides were cloned into this vector using BamHI restriction sites. All clones were sequenced to ensure proper orientation and to ensure that only a single copy of each oligonucleotide was cloned into the vector. Cells were transfected with the reporter gene constructs containing wild-type or mutant –447/–419 regions using a lipofection reagent (Fugene 6; Roche Diagnostics, Indianapolis, IN) and cells were assayed for luciferase activity 48 h later. All transfections used pRLTK-Renilla luciferase as internal standard for transfection efficiency. Firefly and Renilla luciferases were assayed using a dual-luciferase reagent and protocol according to manufacturer’s instructions (Promega, Madison, WI).
Preparation of Translin and TRAX
Mouse translin (40) was cloned into the pTarget mammalian expression vector (Promega). Mouse TRAX cDNA was obtained by RT-PCR amplification of MA-10 cell RNA, and was cloned into the pCDNA3.1 mammalian expression vector for use in transfections (Invitrogen, Carlsbad, CA). These cDNAs were also used to create recombinant protein using an in vitro transcription/translation system (Promega). These cDNAs were also cloned in the EcoRI/BamHI sites of pET42A (Novagen, Madison, WI; emdbiosciences.com) to obtain recombinant protein in bacteria, and to obtain GST-fusion proteins. Bacterially expressed GST-fusion proteins were purified on a glutathione-Sepharose 4B matrix. The GST was cleaved from the fusion protein with Factor Xa (Promega), and protein was eluted from the matrix.
Glutathione-Sepharose 4B Column Chromatography
GST-translin and GST-TRAX were created in pET42A, an N-terminal GST-fusion vector. Recombinant protein was prepared from BL21DE3pLysS bacteria that were transfected with the fusion proteins; proteins were induced by isopropyl ß-D-thiogalactoside treatment of bacteria for 3 h. After induction, bacteria were precipitated, resuspended in 300 µl PBS, and sonicated to lyse cells, and the lysate was cleared by centrifugation at 10,000 x g. The supernatant was then mixed with glutathione-Sepharose 4B matrix. Protein was bound in PBS containing 0.5% Triton X-100, for 30 min at room temperature, and the matrix was washed five times in binding buffer containing 10% glycerol.
For pull-down assays, proteins were labeled in vitro with [35S]methionine using a transcription/translation system, and then added to the GST-protein matrix, incubated 1 h, and washed five times with binding buffer to remove nonspecific binding. Specifically bound proteins were resuspended in SDS gel loading buffer, boiled 5 min, separated on 12% SDS polyacrylamide gels, and exposed to x-ray film and intensifying screens at –70 C overnight.
Preparation of Full-Length and Fragments of Mouse SF-1
Full-length mouse SF-1 cDNA (65) was cloned into the EcoRI site of pBlueBacHis2B, and were used to infect Sf9 insect cells at a multiplicity of infection of 10:1. Cells were grown at 27.5 C for 36–48 h, sonicated, and recombinant protein containing a 6-His-tag at the amino terminus was purified over a cobalt affinity resin (TALON column; Clontech, Palo Alto, CA).
Full-length mouse SF-1 cDNA was also cloned into the EcoRI/BamH1 site of pCDNA3.1 (Invitrogen). Plasmid DNA was used in a coupled transcription/translation reaction (Promega) to generate recombinant protein, or was used directly as a eukaryotic expression plasmid in cell transfections. Fragments of SF-1 protein were created by PCR amplification of full-length mouse SF-1 cDNA. The oligonucleotides used to create these peptides and internal SF-1 mutants are listed in Table 2. SF-1AF2 was created by site-directed mutagenesis (QuikChange; Promega), mutating amino acid 451 from Asn to a stop codon. Serine 203 mutant was also created by changing the serine to an alanine by site-directed mutagenesis
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Present address for P.B.B.: IBM Healthcare and Life Sciences, Research Triangle Park, North Carolina 27709.
S.H.M., S.R.B., C.D., J.-L.V., N.B.H., and P.B.B. have nothing to declare.
First Published Online October 19, 2006
Abbreviations: AF2, Activation function-2; GRIP1, glucocorticoid receptor-interacting protein 1; GST, glutathione-S-transferase; LBD, ligand binding domain; MS/MS, tandem mass spectrometry; SF-1, steroidogenic factor-1; SMRT, silencing mediator of retinoic acid and thyroid hormone receptor; StF-IT-1, steroidogenic factor-inducer of transcription-1; TB-RBP, testis brain RNA-binding protein; TRAX, translin-associated factor X.
Received for publication August 31, 2005. Accepted for publication October 12, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
-hydroxylase/c17,20 lyase). DNA Cell Biol 20:613–624[CrossRef][Medline]-hydroxylase and steroidogenic acute regulatory protein gene promoters in normal and polycystic ovary syndrome theca cells. J Clin Endocrinol Metab 85:2304–2311-hydroxylase/17,20 lyase) in human adrenal NCI-H295A cells. Mol Endocrinol 15:1277–1293-hydroxylase promoter function in theca cells isolated from patients with polycystic ovary syndrome involves nuclear factor-1. Mol Endocrinol 18:588–605-hydroxylase/17,20-lyase), and P450c21 (21-hydroxylase) in the human fetus. J Clin Endocrinol Metab 63:1145–1150[Abstract]-hydroxylase mRNA in senescent bovine adrenal gland. Gerontology 37:262–271[Medline]-hydroxylase/17,20 lyase) in cultured human granulosa cells. J Clin Endocrinol Metab 63:202–207[Abstract] (17-hydroxylase/C17,20 lyase) activity of the junctional zone of the rat placenta. J Endocrinol 125:217–224[Abstract] and P450scc gene expression and regulation in the rat placenta. Endocrinology 130:1309–1317[Abstract]-hydroxylase/C17–20 lyase, and P450 aromatase in fetal mouse gonads. Endocrinology 135:262–268[Abstract]-hydroxylase/C-17,20-lyase cytochrome P450 in rat liver. Endocrinology 138:3166–3174
