| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Urologic Surgery (X.Y., K.S., Y.W., A.G., R.J., R.M.), and Center for Reproductive Biology Research (K.S., M.-C.O.-C.), Vanderbilt University School of Medicine, Nashville, Tennessee 37232
Address all correspondence and requests for reprints to Robert J. Matusik, Department of Urologic Surgery, Vanderbilt University School of Medicine, Nashville, Tennessee 37232. E-mail: robert.matusik{at}vanderbilt.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Lipocalins are a family of proteins characterized by a highly conserved three-dimensional structure, which forms a pocket for small hydrophobic ligands binding, such as fatty acid, steroids, and odorants. Lipocalins bind and deliver these ligands to specific cells (5). Members of lipocalin family exhibit diverse biological functions. Some lipocalins have protective antiinflammatory activity (6) such as Neutrophil gelatinase-associated lipocalin (Lcn2), which is secreted from activated neutrophils and induced in epithelial cells during inflammation (7). Lipocalins also have been implicated in reproductive processes, such as prostaglandin D2 synthase (8).
The Lcn5 gene was mapped to the proximal region of mouse chromosome 2 A3-B, which is homologous to the human chromosome 9q (9, 10). In this region, a gene cluster composed of 14 lipocalins exists. Within this cluster, six genes [Lcn12, Lcn8 (mEP17), Lcn5 (mE-RABP), Lcn10, Lcn13, and Lcn9] are expressed specifically in the epididymis. The Lcn8, Lcn5, Lcn10, and Lcn9 genes display a highly regionalized and complementary expression pattern within the mouse caput epididymidis, suggesting a possible functional redundancy. Lcn12 and Lcn5 are androgen regulated (9, 10). This epididymis-specific lipocalin cluster is flanked by other members of the lipocalin family, prostaglandin D2 synthase and Lcn2, which are expressed in the epididymis as well as other tissues. This organization suggests that a locus control region enhances expression of a large gene cluster. However, epididymis specificity with the correct region-specific patterning is achieved in transgenic mice by using 5'-flanking fragments of the Lcn5 and Lcn8 promoters. This indicates that relatively small fragments of DNA contain sufficient information to control epididymis-specific gene expression.
The expression of Lcn5 is restricted to the middle/distal caput epididymidis (1). The region-specific expression is a common feature of epididymal genes, and it constitutes the basis for regionalization of epididymal epithelium, which has been shown to be essential for spermatozoa maturation (11). However, little is known about how region-specific gene expression is achieved. The highly region-specific expression pattern of Lcn5 provides a good model to study the molecular mechanism involved in tissue- and region-specific transcription regulation of epididymal genes. To define the regulatory DNA region involved in the tissue-specific expression of Lcn5, transgenic mice were established harboring 5 kb genomic DNA fragment of the 5'-flanking region linked to chloramphenicol acetyltransferase (CAT) reporter gene (12). The 5-kb fragment was capable of driving the expression of CAT reporter gene to the principal cells of the mid-distal caput epididymidis, exactly in the same region where endogenous Lcn5 gene is expressed. Also, this fragment of DNA maintained androgen regulation of the reporter gene (12). These results demonstrated that the 5-kb DNA fragment contains all the information required for the temporal, spatial, and hormonal regulation of the Lcn5 gene.
In this study, the minimal regulatory region required for androgen regulation and region-specific patterning in the epididymis of transgenic mice was located between –1.8 to –0.6 kb of the promoter. This 1.8-kb fragment is the intergenic region between the 5'-flanking region of Lcn5 and the final 3'-exon of Lcn8. Because Lcn8 is not androgen regulated and is expressed in the initial segment, a region where Lcn5 is not expressed, the regulatory mechanism for these two closely related lipocalins must be distinct. Deletion analysis identified a Forkhead box subclass A (Foxa) binding site and two new AR binding sites (ARBSs). Foxa1, Foxa2, and Foxa3 proteins were formally known as hepatocyte nuclear factor 3 (HNF3
, HNF3ß, and HNF3
, respectively) (13, 14). They were discovered as controlling liver-specific gene expression. Although the three genes are on different chromosomes (15), they share about 85% sequence identity in the DNA-binding domain (13), which exhibits 82% similarity with the Drosophila forkhead protein (16). The Foxa proteins bind to the same consensus binding site, but each exhibit different affinity. The forkhead domain displays a novel protein folding called "winged helix," which contains three -helices flanked by two loops, or "wings" (17). Helix 3 makes primary contacts with the DNA major groove, and wing 2 makes contacts with the minor groove (17). The probasin (PB) gene is another androgen-regulated lipocalin that is prostate specific (18). This promoter has been extensively used to target genes to the prostate in transgenic animals and a series of ARBS coordinately function to androgen-regulate PB (19, 20). In the PB promoter, adjacent and overlapping to ARBS-1 and ARBS-2 are Foxa1 binding sites that are required for gene expression (21). We have reported that the DNA-binding domain of Foxa1 directly interacts with the DNA-binding domain of AR, suggesting that the two transcription factors interact as they are bound to their overlapping sites in the PB promoter. Further, Foxa1 binding sites were found adjacent to ARBSs in all the prostate-specific promoters examined (21). Not only is Foxa1 required for AR action, but also it is required for prostate ductal morphogenesis and epithelial cell maturation (22). Foxa1 null prostates show severely altered ductal pattern that resembles primitive epithelial cords similar to those found in the embryonic prostate (22). Recently, Foxa1 binding sites have been identified as being frequently adjacent to estrogen receptor-binding sites (23) and in breast cancer cells, Foxa1 regulates gene expression (24). Further data suggest that the function of Foxa1 extends to glucocorticoid-regulated genes (25). Taken together, these data suggest that Foxa transcription factors play an important role in steroid hormone-responsive promoters.
We previously reported that Foxa2 is expressed in the epididymis but not Foxa1 (26). Like Foxa1, Foxa2 physically interacts with the DNA-binding domain of AR. However, Foxa2 down-regulates the epididymal Lcn5, opposite to the role of Foxa1 to enhance androgen-regulated genes in the prostate. The increased expression of Foxa2 in epididymal segments that do not express Lcn5 suggests that Foxa2 functions to restrict region-specific expression of Lcn5.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
), confers both androgen-dependent and epididymis-specific gene expression in transgenic mice whereas a 0.6-kb proximal promoter fragment of Lcn5 does not (12). To test whether sequences of the Lcn8 gene are required for Lcn5 gene regulation or whether the 1.8-kb intergenic region between the two genes is sufficient for dictating hormone-regulated, tissue- and region-specific gene expression, two new transgenic lines were made. A fragment of DNA that was 3.8 kb 5'-flanking to the start of transcription for Lcn5 (also containing Lcn8 exons 2–7) and the 1.8-kb 5'-flanking region that stops at the last exon of Lcn8 were isolated and linked to CAT (Fig. 1A). Transgenic founders established three lines for the 3.8-kb and seven lines for the 1.8-kb 5'-flanking regions.
|
and 2). Transgene copy number was calculated by Southern blot (Fig. 1B), and CAT assay (Fig. 1C) was performed on tissues taken from wild-type and transgenic mice. Tissues analyzed included the epididymis, testis, vas deferens, prostate, seminal vesicle, spleen, kidney, heart, lung, brain, stomach, small intestine, liver, muscle, bulbourethral gland, bladder, ovary, uterus, and oviduct. Transgenic CAT activity was not detectable in any tissue other than the epididymis (Fig. 1C; CAT activity was shown only for the epididymis). Two of the three 3.8-kb Lcn5CAT transgenic lines and five of seven 1.8-kb Lcn5CAT transgenic lines expressed CAT activity in the epididymis (Fig. 1C). The remaining lines showed no CAT expression in any tissue. Expression of CAT activity was copy number dependent (Fig. 1, B and C).
|
) or CAT gene probe (Fig. 2B), respectively. Consistent with our previous report, endogenous Lcn5 gene was specifically expressed in middle/distal caput epididymidis (Fig. 2A). In situ hybridization performed on transgenic mouse epididymis demonstrated that the expression of CAT transgene was restricted to the middle/distal caput epididymidis, in exactly the same region where endogenous Lcn5 gene is expressed (Fig. 2B). These results indicate that the 1.8-kb Lcn5 promoter is sufficient to drive gene expression in a region-specific fashion.
5'-Sequential Deletion Analysis of the Lcn5 Promoter
Because the 1.8-kb fragment contains all the cis-DNA elements required for Lcn5 region-specific expression in the epididymis, we focus our effort to analyze this region to identify the important cis-acting elements and associated protein factors involved in the gene regulation. To look for the minimum cis-regulatory region, 5'-sequential deletion was performed. Fourteen sequentially smaller fragments were generated by PCR, which resulted in deletion of 100 bp each time starting 1.8 kb and finishing 0.5 kb upstream of the Lcn5 promoter. The 14 promoter fragments were inserted into pGL3 at MluI/HindIII site so that the Lcn5 promoter could drive luciferase reporter gene (Fig. 3A). PCR primer sequences were listed in Table 1. The resulting 14 constructs were termed as 1.8 k-Luc to 0.5 k-Luc, according to their position from the start site of transcription. The 1.8 k-Luc construct encompasses the same 5'-flanking region as used in transgenic mice. All the luciferase constructs were cotransfected with AR into DC2 cells, which is a Simian virus 40 T antigen-immortalized cell line derived from epididymal distal caput where endogenous Lcn5 is expressed (27). Nonepididymal HeLa cells were also used to analyze the sequential deletion constructs. As shown in Fig. 3B, luciferase activity reflected different transactivation ability of the 14 promoter fragments. Consistent with our previous report (28), two ARBSs identified in proximal promoter region functioned to show androgen induction of the 0.6 k-Luc construct. The 1.8 k-Luc construct showed a 38-fold increase in activity in the presence of androgen. However, the strongest trans activity and androgen response were achieved with the 1.3-kb fragment, where androgen induction reached 42-fold. Upstream of 1.3 kb, there was a slight decrease of promoter activity and androgen induction, suggesting the existence of inhibitory elements in this region.
|
|
), MCF7, and Chinese hamster ovary cells (data not shown).
Identification of New ARBSs
An androgen response region has been identified within the first 600 bp of the Lcn5 promoter, and two functional ARBSs have been located at position –459/–445 and –102/–88, ARBS-1 and ARBS-0, respectively (28). However, the 600-bp promoter fragment is not sufficient to dictate epididymis-specific gene expression in transgenic mice (12). Because there is a large drop in androgen induction between –1.3 to –1.2 kb, we carefully examined this area for additional ARBSs. Transcription factor search program (www.cbil.upenn.edu/tess) yields a number of potential transcription factor binding sites in this region that might be involved in the Lcn5 gene regulation, but ARBSs were not predicted. This is not surprising because transcription factor search programs use the same consensus sequence for a glucocorticoid response element as for an androgen response element. Therefore, probe1 and probe2 (Fig. 4A) were designed to cover the –1.3-kb to –1.2-kb Lcn5 region in an EMSA. As shown in Fig. 4B, both labeled probe1 and probe2 were capable of binding to purified AR-glutathione-S-transferase (GST) protein, and the binding was competed by 100-fold or 300-fold excess cold ARBS oligomers, which correspond to ARBS-2 in the PB promoter (29, 30).
|
) were designed spanning –1.3 kb to –1.2 kb of the Lcn5 promoter region and used as cold competitors in EMSA (Fig. 4C). Among these competitors, 300-fold molar excess of unlabeled competitor 11 (comp. 11) and competitor 6 (comp. 6) were capable of eliminating AR binding to the P32 labeled probe 1 (Fig. 4C), indicating that both comp. 6 (lane 6) and comp. 11 (lane 10) contain ARBSs. These new ARBSs were denoted as ARBS-2 and ARBS-3, respectively, following the previous nomenclature reporting ARBS-0 and ARBS-1 in the endogenous promoter (28). Competitor 9 (comp. 9) (lane 5) sequence partially overlapped with ARBS-2, thus enabling its slight ability for interfering the AR protein binding to the probe 1.
To compare the AR binding affinity of the newly identified ARBSs, 20 fmol of ARBS-2 or ARBS-3 probe was incubated with different amounts of AR-GST protein in gel-shift assays (Fig. 4D). Shifted bands were cut from the gel and placed into scintillation solution to determine the amount of radioactivity of each band. Results, shown in Fig. 4E, identified that the AR-GST protein binding affinity with ARBS-3 is slightly higher than with ARBS-2.
Mutagenesis Study of ARBSs on Lcn5 Promoter
Two AR binding motifs (ARBS-0 and ARBS-1) have been identified in the proximal promoter region (28). In the context of the 0.6-kb Lcn5 promoter, point mutation of ARBS-0 resulted in a slight decrease of androgen response. In contrast, mutation of ARBS-1 led to a total loss of the androgen responsiveness, suggesting that it was a major cis-acting element (28). To determine whether the two newly identified ARBSs are functional, mutagenesis was carried out on ARBS-2 and ARBS-3, along with the previously reported major AR response element ARBS-1 (28), in the context of the –1.8-kb Lcn5 promoter (Fig. 5A) by replacing the ARBSs with a 10-bp oligonucleotide which has been analyzed by TESS transcription factors searching program to ensure that no additional transcription factor binding sites could be introduced. M1, M2, and M3 were created for mutations of ARBS-1, ARBS-2, and ARBS-3, respectively. Sequences for wild-type and mutant ARBS were listed in Table 1. The resulting mutant constructs were transfected into DC2 cells. ARBS-1, ARBS-2, and ARBS-3 mutation decreased Lcn5 promoter activity to 20%, 56%, and 43%, respectively, as shown in Fig. 5B. Considering the fact that the 0.6-kb construct (containing ARBS-1) gave only 2-fold androgen induction in a 5'-promoter deletion study, it suggests that ARBS-1 functions cooperatively in the context of the 1.8-kb fragment with multiple ARBSs. These results confirmed that ARBS-1 is required for androgen regulation on Lcn5 promoter, although the proximal 600 bp is not sufficient to achieve either strong or epididymis-specific promoter activity.
|
). To determine whether the ARBS-2 and ARBS-3 compensate for each other, the 100 bp between –1.3 kb and –1.2 kb of the Lcn5 promoter was modified by PCR to generate the mutation construct M4. With deletion of the 100 bp containing both ARBS-2 and ARBS-3, promoter activity was decreased to 34%. The M5 construct, with both a 100-bp deletion and ARBS-1 mutation, reduced androgen response to 16% of the intact promoter. These results suggest that the two newly identified ARBSs are functional and that additional weak AR binding motifs may exist in the 1.8-kb promoter that may coordinate with the known ARBSs to synergistically regulate the Lcn5 gene.
Identification of Foxa2 Binding Site in the Vicinity of ARBSs
Functional ARBSs have been identified in both the proximal (within 600 bp) and distal (–1.3 to –1.2 kb) Lcn5 promoter, which contribute to androgen responsiveness of this gene. However, androgen response elements are not sufficient to dictate epididymal specific gene expression because many nonepididymal promoters are regulated by the AR. Our previous work on androgen-regulated prostate-specific gene expression demonstrated that the adjacent regions of the AR binding motif are always involved in the tissue-specific regulation (29, 30). Thus, to identify other transcription factors involved in epididymal gene regulation, we focused on factors bound to an adjacent region of androgen response elements.
Transfection data from the 5'-flanking sequential deletion study have indicated the existence of inhibitory cis-elements upstream of the 1.3-kb promoter fragment, which is located near the two newly identified ARBSs. The inhibitory region may bind to some tissue-specific transcription factors or ubiquitous transcription factors that interact with AR to dictate tissue-specific regulation. Transcription factor search program (www.cbil.upenn.edu/tess) revealed several transcription factors binding to this area, including Foxa, Hunchback (Hb), and Sox9/SRY.
Sequence alignment for the mouse, rat, and human Lcn5 promoters was also performed to identify the DNA regions where transcription factor conservation existed among the three species. As shown in Fig. 6A, no significant homology exists between –1.8 to –1.5 kb of the mouse and rat promoter, but between –1.5 to –1.35, –1.35 to –0.35, and –0.35 to 0 kb, there is 87%, 84%, and 89% homology, respectively. No significant sequence homology exists between the rodent and human promoter, but a conserved pattern of transcription factor binding site (predicted by transcription factor search program) was observed (Fig. 6B). In the high homology region (–1.5 to –1.35 kb) of rat and mouse Lcn5 promoter, Foxa, Sox9/SRY, and Hb binding sites are found upstream of ARBSs. Binding sites for these transcription factors were also found on the human Lcn5 promoter.
|
A gel shift fragment (probe3 in Fig. 7A), covering the potential Foxa binding motif located in the –1.4-kb to –1.3-kb Lcn5 promoter region (Fig. 6B), was synthesized and used to study the Foxa protein’s binding. This region corresponds to the inhibitory region in the sequential deletion study and is also highly conserved between rat and mouse Lnc5 promoters. Because Foxa2, but not Foxa1 and Foxa3, is expressed in epididymis (26), in vitro synthesized Foxa2 protein was used for gel-shift assay. Figure 7A showed the sequence of the oligomers used in this study. In vitro translated Foxa2 bound to the probe3 in gel-shift assay (Fig. 7B). The binding of Foxa2 transcription factor to Lcn5 promoter was further confirmed by antibody supershift experiments using Foxa2 antibody and nuclear extract from DC2 cells [Fig. 7C; repeated experiment previously published (26)].A shifted band was observed with the addition of Foxa2 antibody (Fig. 7C). If Foxa is critical for Lcn5 gene regulation, similar Foxa binding sites would be conserved in the Lcn5 promoter of human and rat genes. Transcription factor binding site search found three Foxa binding sites on the corresponding area for each gene: two on human, and one on rat Lcn5 promoter (Fig. 6B). Three more oligonucleotides were synthesized based on the predicted Foxa binding motifs on human and rat E-RABP promoter and used in gel-shift assay. As shown in Fig. 7D, when used as competitors in gel-shift assay, all three oligomers were capable of eliminating the Foxa2/DNA complex marked by arrowhead. Further, Foxa2 binding on the Lcn5 promoter in vivo was determined by chromatin immunoprecipitation assay, and Foxa2 physically interacts with AR as previously reported (26). This suggests that Foxa motif is conserved among the human, rat, and mouse Lcn5 promoter.
|
). Foxa2 was detected mainly in nuclei of epididymis principal cells and showed a region-specific expression pattern along the epididymis: no staining was observed in the initial segment, light staining was seen in proximal caput, gradiently heavier staining was seen in the middle and distal caput, and the strongest staining was observed in corpus and cauda, regions with little or no expression of Lcn5.
|
, cotransfection of Foxa2 inhibited 1.8 kb Lcn5 promoter activity [repeated experiment previously published (26)]. The inhibitory effect of Foxa2 on the Lcn5 promoter is consistent with the data from gel shift assays that a Foxa2 binding site exists immediately upstream of the 1.3-kb Lcn5 promoter and the 5'-flanking deletion experiments that show the existence of inhibitory element(s) upstream of the –1.3-kb Lcn5 fragment.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Development of a complex eukaryote requires differential transcription of more than 37,000 genes in precise spatial and temporal patterns (31). The regulatory proteins expressed start a cascade that can account for cell determination where the cell is assigned a developmental fate and, subsequently, cell differentiation where the assigned cell now emerges with its own unique character. It is estimated that about 2000–3000 transcription factors exist in mammals. They can be placed in three groups: 1) general transcription factors, such as NF1, that are ubiquitously distributed and constitutively important for most genes; 2) conditional transcription factors, such as steroid receptors, that are signal-dependent; and 3) cell-specific factors, such as forkhead proteins, that are restrictively distributed and expressed in certain cell types during development (32). In the regulatory regions of some well-characterized genes, multiple proteins are required to act on the same enhancersome for optimal and specific gene expression. However, we still have a limited understanding of the underlying mechanism that controls cell-specific gene expression. Several transcription factors that are highly specific for pituitary (33), muscle (34), testis (35), mammary gland (36), and prostate (37, 38) have been described, but single factors alone cannot explain tissue-specific expression. Combinatory use of transcription factors would lead to an exponentially large number of regulatory decisions, ensuring that each gene is expressed in the right place and at the right time under a unique regulation (31). By defining in a number of genes the minimal DNA promoter elements necessary to control tissue specificity, in species ranging from Drosophila to human, the complete cascade is estimated to require as little as four to eight regulatory proteins (39).
Cell determination for the reproductive organs occurs in the embryo, but the final stages of cell differentiation take place in adulthood. For example, the AR is required in the embryonic Wolffian ducts to develop and stabilize into the epididymis, but differentiation is completed at sexual maturation. When differentiated function of the epididymis is completed, epididymal-specific gene expression is not uniform throughout the organ; rather it is patterned differently in specific segments (40). The function of this patterning is to expose the spermatozoa to a continuously changing microenvironment as it passes through this long convoluted duct (41). This changing regionalized microenvironment results in spermatozoa acquisition of forward motility, the ability to fertilize, and the maintenance of viability during storage in the cauda epididymidis (42).
The epididymal duct is lined by the same cell type, a polarized columnar principal cell. Gene regulation in the principal cell is achieved by two different mechanisms: extracellular and intracellular regulation. For extracellular regulation, luminal testicular fluid has been reported to have a larger impact on gene expression in the initial segment of epididymis (5). Growth factors and other regulatory factors, such as retinoid carried in testicular fluid, display a gradient distribution along epididymis tube and mutually affect epididymal gene expression (43). AR is expressed throughout the duct such that androgens act intracellularly to regulate genes in segment-specific patterns. Lcn8 expression is controlled by luminal testicular fluid in the initial segment. Lcn5, located 1.8 kb downstream from the Lcn8 gene, is regulated by androgens in the middle/distal caput epididymidis (1). Both genes are epididymal specific, yet regulation and segment expression are distinct. The mechanism(s) that controls this regionalization of gene expression is poorly understood.
The AR is a critical trigger (44); however, AR alone cannot explain the differences in organ-specific gene expression in the male reproductive tract. To identify AR coregulators that would be involved in organ-specific and region-specific gene expression, we have examined androgen-regulated prostate and epididymal genes. Recently we reported that Foxa1 is expressed in the prostate (45) and that it interacts with the AR to control multiple prostate-specific genes (21). Further, we found that Foxa1 is required for prostate ductal morphogenesis (22). Interestingly, Foxa2 is only expressed in the embryonic prostate buds and prostatic neuroendocrine cells (45, 46). Because Foxa1 and Foxa2 can bind to the same DNA motif, we tested whether they would have a redundant function on prostate-specific promoters. Surprisingly, Foxa2 will activate prostate-specific genes in the absence of androgens and even in the absence of the AR (46). In the epididymis, we found that Foxa2 rather than Foxa1 was expressed (26). Further, the Foxa2 expression pattern increased in a gradient such that the strongest staining was in the corpus and cauda, regions with little or no Lcn5 expression (40, 47). Like Foxa1, Foxa2 is a DNA binding protein that is involved in opening chromatin structure, but our data show that it can also be tethered to DNA through binding to the AR. This suggests that Foxa2 can also function as coactivator/repressor of gene expression.
We were surprised to see Foxa2 having an inhibitory role on androgen action on Lcn5, an epididymal gene, because with prostate-specific genes Foxa2 will activate gene expression in the absence of androgen (46). This suggests distinct regulatory differences between how Foxa1 and Foxa2 interact with the AR and other cofactors in the context of epididymal- and prostate-specific promoters. Our data provide further support for a fundamental role of the Foxa proteins to serve as cell-specific factors that interact with signaling-dependent transcription factors such as AR and estrogen receptors (21, 22, 23) to control tissue-specific gene expression.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
In Situ Hybridization
Epididymis taken from wild-type or transgenic mice was fixed in 4% paraformaldehyde-PBS overnight and embedded in paraffin. Probes were labeled using [35S]UTP and T7/Sp6 MAXIscript (Ambion, Inc., Austin, TX). Hybridization and washing were the same as previously described (1).
5'-Deletion Constructs and ARBS Mutation Constructs
Fourteen sequentially deleted promoter fragments were generated by removing 100 bp each time from –1.8 kb to –0.5 kb of Lcn5 promoter using the PCR approach. High-fidelity DNA polymerase (Pfx, Invitrogen, Carlsbad CA) was used for PCR. Primers were designed so that PCR can amplify sequences from –1.8 kb to +26 bp Lcn5 gene. Fourteen forward primers recognize sequences of –1.8 kb to –0.5 kb (with 100-bp interval) of Lcn5 promoter, respectively. The resulting PCR fragment was inserted into pGL3-Basic vector (Promega Corp., Madison, WI) at MluI/HindIII sites to generate Lcn5 luciferase constructs, which were termed as 1.8 k-Luc to 0.5 k-Luc. For promoter mutagenesis studies, ARBS mutation constructs were generated using a PCR-based site-directed mutagenesis method as previously described (30). ARBS-1 mutant sequence was the same as previously reported (28). ARBS-2 and -3 were replaced by 10-bp oligomers as listed in Table 1.
Cell Culture and Transfection Assays
Immortalized mouse epididymal epithelial cell line DC2 was cultured as previously described (27). Human cervical adenocarcinoma cell line HeLa was obtained from American Type Culture Collection and cultured as recommended by ATCC. Transient transfection assays were carried out using lipofectamine 2000 (Invitrogen) on 24-well plates. For every well, 0.6 µg test plasmid DNA, 0.2 µg AR expression vector, and 12.5 ng pRL-SV40 plasmid were cotransfected into host cells. Medium was replaced 6 h later by DMEM (Life Technologies, Inc., Gaithersburg, MD) with 5% charcoal/dextran-treated fetal bovine serum (HyClone Laboratories, Logan, UT) in the presence or absence of 10–8 M dihydrotestosterone and cultured for 24 additional hours. The luciferase activity was determined using dual luciferase reporter assay system (E1960; Promega Corp.) and LUMIstar (BMG Laboratory Technologies, Inc., Durham, NC). Renilla luciferase activity served to normalize the transfection efficiency. Results were presented as luciferase/min/mg protein or as percentage using 1.8 k-Luc luciferase as 100.
EMSAs
Nuclear extract from DC2 cells was prepared as previously described (30). In vitro translated Foxa2 protein was synthesized using pcDNA-Foxa2 plasmid as template and TNT T7 Quick Coupled Transcription/Translation System (Promega Corp.). AR-GST fusion protein was purified as described previously (48). All oligonucleotides for EMSA were synthesized by Integrated DNA Technologies, Inc. (Coralville, IA). Probes were end labeled using [-32P]dATP and DNA polymerase I Klenow fragment (New England Biolabs, Beverly, MA) and purified using a G50 column. In vitro translated Foxa2 protein (1 µl) or 4 µg of nuclear extract from DC2 cells was preincubated with Foxa2 antibody (P19; Santa Cruz Biotechnology, Inc., Santa Cruz, CA) or cold competitors (nonradiolabeled oligonucleotides) at room temperature for 20 min, followed by an additional 20-min incubation with 32P-labeled probe. The DNA/protein complex was resolved on 5% nondenatured polyacrylamide gel and processed for radioautography.
Immunohistochemical Staining of Foxa2
Epididymis was fixed in Bouin fixation fluid and embedded in paraffin. Sections were cut (5 µm) for immunohistochemistry study. Foxa2 antibody was purchased from Santa Cruz Biotechnology (P19). The staining was performed using Vectastain ABC kit (Vector Laboratories, Inc., Burlingame, CA), and sections were counterstained with Harris hematoxylin (Surgipath, Richmond, VA).
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Disclosure statement: The authors have nothing to disclose.
First Published Online June 1, 2006
1 X.Y. and K.S. contributed equally to this study. 
Abbreviations: AR, Androgen receptor; ARBS, AR binding site; CAT, chloramphenicol acetyltransferase; comp.6, competitor 6; E-RABP, epididymal retinoic acid-binding protein; Foxa, Forkhead box subclass A; GST, glutathione-S-transferase; HNF, hepatocyte nuclear factor; Lcn5, lipocalin 5; PB, probasin.
Received for publication January 6, 2006. Accepted for publication May 24, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
belongs to a gene family in mammals that is homologous to the Drosophila homeotic gene fork head. Genes Dev 5:416–427 (Forkhead Box A1) and androgen receptor in transcriptional regulation of prostatic genes. Mol Endocrinol 17:1484–1507 target promoters reveals that FOXA1 defines a domain of the estrogen response. Proc Natl Acad Sci USA 102:11651–11656NURSA Molecule Pages Link:
This article has been cited by other articles:
 |
|
 |
|
  U. Simanainen, K. McNamara, R. A. Davey, J. D. Zajac, and D. J. Handelsman Severe Subfertility in Mice with Androgen Receptor Inactivation in Sex Accessory Organs But Not in Testis Endocrinology, July 1, 2008; 149(7): 3330 - 3338. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. V. Heemers and D. J. Tindall Androgen Receptor (AR) Coregulators: A Diversity of Functions Converging on and Regulating the AR Transcriptional Complex Endocr. Rev., December 1, 2007; 28(7): 778 - 808. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||
