| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Testis |
Department of Chemistry and Biochemistry and The School of Medicine, University of South Carolina, Columbia, South Carolina 29208
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
gene regulation, meiosis, spermatogenesis, testis, transcription factor
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Immunohistochemical and in situ hybridization studies have shown that H1t expression is upregulated in mid-late pachytene spermatocytes in the rat [6–8] and mouse [9– 11], although a sensitive reverse transcription-polymerase chain reaction investigation indicated that trace levels of H1t mRNA could be detected before meiosis in mice [12]. The tightness of its transcriptional regulation was demonstrated by a transgenic study in which an H1t promoter-diphtheria toxin fusion gene led to complete elimination of germ cells in males, but had no other apparent effect [13]. This tight regulation is made more intriguing because the H1t promoter region is strikingly similar to those of the 5 standard H1 variants expressed ubiquitously in somatic cells [14–17]. Accordingly, the potential of the H1t promoter is somehow constrained to tissue and developmental dependency by additional regulatory elements. In fact, a number of studies have pointed to various upstream regions of the gene that have an inhibitory effect in somatic cells [18–21].
Study of the H1t promoter has shown that it shares five sequence motifs with most other standard H1 genes: a TATA box, a CAAT consensus for NFY, a GC box for Sp1, and a pair of H1-specific upstream elements related to AACACA. Between the CAAT and Sp1 sites lies an H1t-specific sequence first identified by Wolfe and Grimes, which includes the palindrome CCTAGG [22]. They also detected a testis-specific nuclear binding protein for this region, which they designated Te1 [22, 23]. Clare et al. [24] confirmed their observation but found that an additional set of complexes of similar mobilities was also formed from all somatic nuclear extracts. The identities of these bands have remained obscure, although Grimes recently suggested that the Te1 palindrome motif was related to the so-called X box in the promoters of the major histocompatibility complex II (MHCII) genes [25].
Analysis of nuclear proteins that bind the MHCII X box led to identification of the RFX family of DNA binding proteins, which share a variant form of the winged helix domain [26–28, reviewed in 29, 30]. This family is encoded by 5 genes in mammals (RFX1–5), which in some cases have splice variants [30–32]. The MHCII genes themselves are uniquely controlled by RFX5, due to its unique ability to associate with a set of additional proteins [30]. This is shown decisively by the fact that one variant of the human bare lymphocyte syndrome results solely from the absence of RFX5 [33]. The regulatory targets of the other family members are only partly understood. RFX proteins bind to DNA as dimers [28], and RFX1-4 can form hetero- as well as homodimers [26, 31, 34], leading to a diverse mix of nuclear factors that have very similar or identical binding site specificities.
Here we report a study demonstrating that both the testis-specific and the ubiquitous somatic nuclear binding proteins for the H1t palindrome motif are indeed RFX regulatory factors. The testis-specific complex is due to RFX2, while the ubiquitous complexes are due to RFX1 and its dimer partners. By immunohistochemistry, we have shown that RFX2 is upregulated at the beginning of the pachytene stage of spermatogenesis. We have also shown that RFX binding sites occur in the promoter regions for two other genes upregulated during male meiosis, lamin C2 and Sgy. While this manuscript was in preparation, Wolfe et al. reported that the Te1 binding protein for the palindrome motif is RFX2 [35]. Our work confirms and extends their contribution.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Sprague Dawley rats and CD1 mice were obtained from Harlan (Indianapolis, IN) and killed by CO2 inhalation. Use of mice and rats in this investigation was approved by The University of South Carolina Animal Use and Care Committee.
Band Shift Assays
Nuclear extracts (NE) from testis and liver (2–3 mg protein/ml) were prepared from six 40-day-old rats according to the procedure of Lichsteiner et al. [36] as described in detail by Sierra [37]. Double-stranded oligonucleotides were labeled at 3' recessed ends using Klenow DNA polymerase [38] and 
-[32P]-dATP (Amersham Biosciences, Piscataway, NY). NE from Raji and HeLa S3 cells were purchased from Active Motif (Carlsbad, CA).
Oligonucleotides used for probes were designed to have 5' single-strand extensions of GATC at either end. The upper strand on each oligo is shown, with the H1t palindrome motif underlined where applicable:
GATGC (H1tpal); AGACAGA (RFX, IL-5R [34]); CGACCGT (Sgy [39]); ATACCT (Lamin C2 [40]); 
AGGTACG (c-Mos [41]); 
CAACGC (Pdha-2 [42]); The mouse Sgy (Ensemble mouse transcript 33057) sequence is located at –100 relative to the 5' end of the cDNA [39]. The mouse lamin C2 sequence is located at –227 relative to the initiating ATG [40].
Binding reactions were done in a final volume of 12.5 µl in a mixture containing 50 mM NaCl, 10 mM HEPES (pH 7.9), 5 mM MgCl2, 0.5 mM EDTA, 5% glycerol, and 0.5–1 µg of poly[d(I-C)] (Roche Molecular Biochemicals, Indianapolis, IN), using 50 000 CPM of probe and an incubation of 20 min on ice. For antibody supershift reactions, the protein extract was incubated in binding buffer in the presence of antibody for 1 h on ice, the labeled probe fragment was then added, and incubation continued a further 20 min on ice. Primary antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, CA): RFX1 (I-19, sc-10652), RFX2 (C-15, sc-10657 or E-20, sc-10660), RFX3 (T-18, sc-10663). Complexes were resolved in a 5% polyacrylamide gel containing 45 mM Tris, 45 mM boric acid, 1 mM EDTA. The gel was dried under heat and vacuum and exposed to Kodak BioMax film in the presence of an intensifying screen.
In Vitro Synthesis of RFX2
A full-length cDNA clone for mouse RFX2 (ATTC MGC-6105) was obtained from American Type Culture Collection (Manassas, VA), and the coding region was transferred to pCI (Promega, Madison, WI) to permit expression from a phage T7 promoter. This plasmid was used to program a TNT-coupled transcription/translation system derived from reticulocytes (Promega) following the protocol supplied.
Western Blotting
Tissues from adult mice were homogenized in four volumes of ice-cold 50 mM sodium phosphate, pH 7.8, in the presence of a cocktail of protease inhibitors (Cat no. 1-836-153; Roche). The homogenate was then rapidly mixed with a 2x modified SDS sample mix (1x = 2% SDS, 50 mM sodium phosphate, pH 7.8, 50 mM beta mercaptoethanol, 10% glycerol) already in a boiling water bath and incubated 3 min. Viscosity was decreased by forceful passage through a 22-gauge needle. Purified populations of late spermatocytes and round spermatids were prepared from four adult mice according to the protocol described by Heyting and Dietrich [43] with elutriation conditions modified for mouse cells [44]. These cells were collected by low-speed centrifugation and further purified by equilibrium centrifugation in Percol gradients as described [43]. Purity of fractions was estimated to be approximately 95% based on nuclear morphology after fixation with cold acetic acid:methanol (1:3) and staining with Giemsa. Protein was estimated as described [45], using the Pierce Chemical (Rockport, IL) BCI assay with serum albumin as standard.
SDS samples (25 µg of protein for whole cell/tissue extracts, 2–3 µg of NEs) were resolved in a 10% gel [46], using a 100:1 ratio of acrylamide to bisacrylamide [47]. Proteins were transferred by electroblotting to Hybond ECL nitrocellulose membranes (Amersham) using an alkaline transfer buffer [48]. Membranes were blocked in 1% Carnation skim milk, 0.05% Tween (Bio-Rad, Richmond, CA) 10 mM Tris HCl (pH 8.0), 150 mM NaCl (TBST) for 1 h at room temperature or at 4°C overnight. Primary antibodies were used at 0.5 µg/ml in the blocking solution. After washing in TBST, detection was done with Santa Cruz peroxidase-conjugated bovine anti-goat secondary antibody at a dilution of 1:250, in the blocking mixture, for 45 min at room temperature. After washing in TBST, detection employed Amersham Enhanced ECL Reagent and Kodak X-Omat LS film (Rochester, NY).
Immunohistology
Mouse testis was bisected and fixed overnight at 4°C in 4% paraformaldehyde. Paraffin-embedded tissue was sectioned (8 µm), mounted (Superfrost plus slides; Fisher, Atlanta, GA), and rehydrated by passage through xylene and graded ethanol solutions.
Antigen retrieval [49] was performed as described [50]. Sections were submersed in 250 ml 10 mM sodium citrate (pH 6.0) in a slide holder and placed in a microwave pressure cooker (Nordicware, Minneapolis, MN) containing 650 ml distilled water. Heating was carried out in a microwave oven at 800 W for 20 min followed by 15 min at 300 W. When cool, slides were washed in PBS, and endogenous peroxidase activity was blocked by a 5-min incubation in 1% hydrogen peroxide, 70% ethanol. After a wash in PBS, slides were incubated in 300 mM glycine for 5 min and washed in PBS. Sections were blocked in 10% rabbit serum in PBS for 1 h at room temperature. RFX proteins were localized by incubation with a 1:5000 dilution of RFX2 antibody (C-15, Santa Cruz) or 1:1000 dilution of RFX1 antibody (I-19, Santa Cruz) in 5% normal rabbit serum, PBS, 0.1%Triton X100 overnight at 4°C. Sections were washed with PBS containing 0.1% Triton, incubated with the second antibody (1:300 dilution of horseradish peroxidase-conjugated rabbit anti-goat) for 1 h at room temperature, and washed in PBS. Bound antibodies were localized by staining with the DAP-Plus Substrate kit from Zymed Laboratories (South San Francisco, CA) according to manufacturer's protocol. Successive sections were stained with Gill No. 2 hematoxylin (Fisher Scientific) and the different antisera. Images were recorded on an Olympus DP10 digital camera mounted on an Olympus microscope and were manipulated using Adobe Photoshop software.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The H1t promoter region contains a conserved palindromic motif CCTAGG at –60 from the Cap site that is thought to be a cis-acting regulatory element [5, 22, 24]. An oligo centered on this motif was previously shown to form a group of retarded complexes in band-shift analysis of rat testis NE [24]. The upper set was observed in extracts from several somatic organs and cell lines, while the lowermost band was apparently unique to testis (see Fig. 5 of [24]). Mutation of either half of the palindrome eliminated formation of both the ubiquitous and testis-specific complexes [24].
|
As a first step to test if the binding proteins detected by Clare et al. [24] are related to the RFX family [25], we compared testis and liver NE by band-shift analysis, using labeled oligos containing either the H1t palindrome or the strong, interleukin-5 receptor (IL-5R) RFX site [34] (Fig. 1, lanes 1–4). Consensus oligos for the common DNA binding proteins Sp1 and NFY were run in parallel to provide mobility references (Fig. 1, lanes 5 and 6). The pattern of retarded complexes was quite similar for both RFX and H1t probes with both testis and liver NE, although less extract was required to generate prominent bands with the RFX oligo (Fig. 1, lanes 1–4). In the previous report, Clare et al. [24] showed that only the lower band formed with H1tpal is testis-specific. The upper bands (Fig. 1, "somatic") were found to varying degrees in all other organs examined [24].
|
If the H1 probe is binding to RFX proteins, then cross-competition should occur between the H1t pal and RFX oligos. When RFX was the labeled probe, a 50-fold excess of unlabeled RFX oligo eliminated both somatic and testis-specific complexes (Fig. 2A, lanes 1 and 2). In cross-competition, the unlabeled H1t palindrome reduced formation of all complexes on the RFX probe but required a higher concentration (Fig. 2A, lanes 3–6). When the experimental situation was reversed, both unlabeled H1tpal and RFX fragments competed away the retarded bands formed on the labeled H1tpal oligo. The RFX oligo was again more effective at a 50-fold excess (Fig. 2A, lanes 7–9). As a control, a 100-fold excess of an unrelated oligo containing the Sp1 consensus did not compete (Fig. 2A, lane 10). These results suggest that the same proteins bind to both probes, with the stipulation that binding to H1tpal is weaker than to RFX.
|
The sizes of RFX proteins differ significantly [27]: 104, 76.5, and 83.5 kDa for mouse RFX1,2,3 respectively (SWISS-PROT P48377, P48379, P48381). Because both homo and heterodimers form [26, 34], complexes of six different masses are possible from RFX1–3. RFX1 is expressed ubiquitously with little tissue-to-tissue variability [26]. In contrast, RFX 2 was the most prominent RFX mRNA detected in mouse testis, but was barely detected elsewhere, except for intestine and stomach [26]. RFX3 mRNA was prevalent in brain, intestine, and testis [26]. RFX4 is most prevalent in testis, followed by brain [31, 32]. However, in mouse testis, the only splice variants detected lacked the DNA binding domain and are of uncertain function [32].
To help identify the complexes obtained by band-shift assay, we used extracts from cell lines for which the RFX band-shift pattern is known. Raji cells, a human B-cell line, generate primarily RFX1 homo- and RFX 1,3 heterodimers [26] while HeLa cells produce primarily RFX1 homodimers [26, 51]. These extracts were compared with rat testis NE using the RFX probe (Fig. 2B). The mobility references indicate that the major somatic complex is comparable with an RFX1 heterodimer, while the mobility of the lower, testis-specific band implicates one of the smaller RFX proteins, such as RFX2.
To make positive identification of the RFX proteins involved in these complexes, we next used specific antisera. Using the RFX probe with rat testis NE, a specific antiserum to RFX1 eliminated the entire set of somatic binding complexes and produced a supershifted band near the top of the gel, indicating that all these complexes involve RFX1 (Fig. 3, lane 2). The testis-specific band was not affected by anti-RFX1. In contrast, antisera to two different epitopes of RFX2 eliminated the lower, testis-specific band completely, while generating new supershifted complexes. Some of the low-mobility, somatic complexes were also reduced in intensity, perhaps indicating that they involve RFX2 heterodimers (Fig. 3, lanes 3 and 4). When H1t palindrome was used as labeled probe, identical results were obtained. Antiserum to RFX3 had no effect on either the somatic or testis-specific complex, nor did nonimmune serum (Fig. 3, lanes 9 and 10). From these results, we conclude that the testis-specific binding complex for the H1t palindrome is likely to be an RFX2 homodimer. The somatic band is due to RFX1, presumably a heterodimer, because the mobility suggests something smaller than an RFX1 homodimer (compare Fig. 2B). The possibility of an RFX1–3 heterodimer is not ruled out because it is not certain that the antiRFX3 antibody cross-reacts with the rat protein.
|
Localization of Elevated RFX2 to Germ Cells from Pachytene Spermatocytes to Early Spermatids by Western Blotting and Immunohistochemistry
The early studies of Reith et al. [26] showed that RFX2 mRNA was most prominent in testis but did not localize expression to specific cell types. Therefore, we examined the expression pattern of RFX2 within the testis by Western blotting of extracts from enriched cell populations and by immunohistochemistry.
Late spermatocytes and round spermatids were purified from adult mouse testis, and extracts were resolved by SDS PAGE, transferred to nitrocellulose membranes, and probed with antiserum to RFX2 (Fig. 4). A major band of 76.5 kDa was readily identified in whole mouse testis, but was faint or missing from mouse brain and kidney, or Raji cells, which have little RFX2-binding activity [26]. This same band was enhanced in both late spermatocyte and round spermatid extracts. The 76.5-kDa band was also prominent in rat testis NE, but not detected in rat liver NE. Similar results were obtained with a second anti-RFX antiserum (SC E-20, results not shown). This Western blot demonstrates that enrichment of RFX2 extends to both spermatocytes and round spermatids.
|
More precise localization of RFX2 was achieved by immunostaining of adult mouse testis sections. Serial sections were stained with hematoxylin, antiRFX2, or antiRFX1 (Fig. 5). Examination of the sections stained with anti-RFX2 indicates that the ring of germ cells (spermatogonia and early spermatocytes) next to the basement membrane in tubules display a background level of staining, while the large late spermatocytes and early round spermatids are intensely stained. Specifically, the early pachytene spermatocytes of stage I–III tubules are only weakly stained compared with the very prominent round spermatids present in the same tubules (Fig. 5b). However, by stages IV–VI, which are difficult for us to differentiate, immunopositive staining is definite in pachytene cells (Fig. 5e). This leads to the highly immunopositive late pachytene spermatocytes of stages VIII and IX (Fig. 5, h and k). As these cells divide to generate haploid cells, staining remains high (Fig. 5, n and b). Intense staining remains over early spermatid nuclei (Fig. 5e) but diminishes somewhat as spermatids mature. By step 8, it is noticeably weaker than in the spermatocytes of the same tubule (Fig. 5h). Immunostaining is then rapidly lost from haploid cells as they undergo nuclear elongation from steps 9 to 12 (Fig. 5, k and n). Spermatogonia and prepachytene spermatocytes, as well as Sertoli cells, do not stain above background.
The situation for anti RFX1 was quite different. Sertoli cell nuclei stained positively at all stages of the cycle (Fig. 5, c, f, i, l, and o). However, the only germ cells to stain above background were late round spermatids of stage IV– VI (Fig. 5f). The immunostaining for RFX2 is summarized diagrammatically and compared with the occurrence of H1t mRNA in Figure 6.
|
RFX2 Binding Sites Occur in the Promoter Regions of the Mouse Lamin C2 and Sgy Genes
Because of its striking upregulation in spermatocytes and early round spermatids, RFX2 may participate in the regulation of multiple genes that function during meiosis or beyond. In fact, the CCTAGG motif of the H1t palindrome has been identified as similar to presumptive DNA control motifs for other testis-specific promoters such as c-Mos and Odf-1 [41]. RFX1–3 share a conserved DNA binding domain and recognize similar or identical binding sites. The consensus site (Fig. 7A) is an inverted 6-base pair (bp) repeat with a variable spacer (0–3 nt) [52]. The H1t palindrome is the center of this larger region (Fig. 7A), and strong RFX binding sites depend critically on several of the outlying nucleotides [28]. In a preliminary examination of promoters that are activated in meiosis, we identified several containing similarities to the RFX consensus. In particular, the testis-specific promoter for mouse lamin C2 [40] and Sgy [39] have potential matches, while the promoter for the pyruvate dehydrogenase 2a gene has a perfect match to an RFX half site [42]. As pointed out earlier [41], the cMos promoter has a good homology to the central CCTAGG palindrome. We used oligos for these regions and tested them by band-shift assay with rat testis NE (Fig. 7B). Both lamin C2 and Sgy gave band patterns indistinguishable from the H1t palindrome while no comparable bands were generated by the cMos or Pdha2 (results not shown) probes. The affinity of the lamin C2 site for RFX was comparable with that of H1t, while the affinity of Sgy was significantly higher. We employed the specific antisera used above to show that the complexes formed with lamin C2 and Sgy were indeed due to RFX1 and RFX2 (Fig. 7B, lanes 3–8). To study binding of DNA fragments in a simplified system, we generated RFX2 in vitro and showed that it binds to RFX, H1tpal, lamin C2, and Sgy (Fig. 7B, lanes 10–13). The in vitro protein generates bands of slightly greater mobility compared with that from rat testis NE (compare Fig. 7B, lanes 9 and 10). The reason for this is unknown, but could relate to additional proteins that are present on the in vivo complex or to possible splice variations. As equal amounts of probe and protein were added to the these binding reactions, it is apparent that H1tpal and lamin C2 are relatively weak binding sites for the plasmid-generated RFX2, but that Sgy is considerably stronger and more comparable with the IL-5R probe (Fig. 7B, lanes 10–13).
|
Data from Hybridization Array Experiments Support the Appearance of RFX2 in Pachytene Spermatocytes
Recently, results of cDNA array hybridization for genes covering the initial wave of spermatogenesis were reported [53]. When plotted graphically (Fig. 8), expression of RFX2 and Sgy mRNAs are strikingly similar during the development of spermatogenesis, appearing above basal level at Day 14, corresponding to the first appearance of pachytene cells, and rising sharply thereafter. These results for cDNA measurements agree with our immunohistochemical location of RFX2.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
Clare et al. [24] detected a set of protein complexes formed from testis NE and the H1t palindrome region. One appeared to be testis-specific, while the remainder were widely distributed in other organs. We have now explained those earlier results by assigning the larger, ubiquitous binding complexes to various RFX1 hetero- or homodimers, and the smaller, testis-specific complex to RFX2. Several lines of evidences support these conclusions. 1) Oligonucleotides containing either the H1tpal site or a strong RFX site from the IL-5R promoter gave similar binding complexes after band-shift analysis of rat testis nuclear proteins. 2) These two oligos showed cross competition with one another in the band-shift assays, indicating that they form specific complexes with proteins having similar or identical binding sites. 3) Specific antisera were used to assign the more slowly migrating complexes to RFX1 and the more rapidly migrating complex to RFX2. By comparison with known RFX1 homo- and heterodimer complexes formed from HeLa and Raji cell extracts, we surmised that the RFX2 complex is a homodimer, while the major RFXI complex is some manner of heterodimer. 4) In vitro-synthesized RFX2 formed bound complexes of identical mobility with either oligo. The only difference observed between the two DNA sequences is that the IL-5R site is of higher affinity and therefore reveals less abundant RFX binding complexes more readily.
The properties of the RFX family explains the multiple complexes of varying abundance detected by band-shift assay with either the H1t or IL-5R probes. Of this interesting 5-gene family, RFX1–3 are the most likely to contribute to the pool of DNA-binding proteins in rat testis extracts. RFX5 dimerizes only with itself and appears to function primarily in certain cells of the immune system or elsewhere in the event of interferon gamma stimulation [30, 33]. RFX4 is expressed strongly in testis, but all testis splice variants are defective in some manner, lacking either the presumptive transcriptional activation domain [32] or the DNA-binding domain [31]. Furthermore, the only genes known to depend on RFX4 are found in neural cells because an accidental disruption of the RFX4 gene in mice led to brain defects [32]. The remaining proteins, RFX1–3, are the most similar of the family, with identical or nearly identical binding site specificities and the ability to cross dimerize [26]. While mRNA for each was detectable in all organs examined, testis showed variable enrichment for all three [26]. Enhanced expression was most dramatic for RFX2, where the elevation relative to kidney was estimated at 600-fold [26]. This enormous difference in expression makes it easy to see how RFX2 could appear testis-specific, particularly when using the suboptimal binding site from H1t in band-shift assays. The testis enhancement of RFX3 was less marked, while RFX1 has the most uniform tissue expression of the family [26]. RFX3 was also relatively elevated in brain and intestine [26]. Quite recently, a targeted disruption of the mouse RFX3 gene was reported to affect nodal cilium development and left-right asymmetry specification. Rare male homozygotes survived at least a month, but effects, if any, on spermatogenesis were not mentioned [54].
What is the functional significance of RFX binding to the H1t promoter? Evidence that the H1t palindrome site is important for promoter regulation is largely circumstantial. The fact that a testis-specific binding factor appeared in spermatocytes, associated with the major expression of the gene, certainly suggested that it is an activating protein. However, it was not possible to demonstrate a role for this DNA region using in vitro transcription [24]. The close association of the palindrome region with the adjacent upstream, Sp1 consensus has made the functional importance of the CCTAGG element difficult to dissect. A transgenic mouse with a DNA substitution that eliminated the palindrome motif, the Sp1 site, and adjacent 5' sequences failed to express the transgene [55]. However, given the importance of the Sp1 site in several activity assays [18, 19, 24, 56], the transgene silence could be attributed as readily to the missing Sp1 site as to the absence of the RFX site. In their recent report, Wolfe et al. reported that cotransfection of an RFX2 expression plasmid in rat testis GC-2spd cells stimulated expression of an H1t promoter-luciferase construct [35]. This result is encouraging, but the stimulation was not very responsive to mutations in the RFX binding site. Perhaps, as suggested, there is redundancy between the palindrome site and a related site upstream of the Sp1 element.
The regulatory effect of RFX proteins is extremely context dependent. In the genes that have been studied in some detail, RFX activation depends strikingly on collaboration with proteins at adjacent promoter binding sites, on coactivators expressed in a particular cell lineage, or on both. The action of RFX5 on MHCII promoters may set the paradigm for the family. MHCII genes are expressed constitutively only in professional antigen-presenting cells of the immune system, such as B lymphocytes and dendritic cells [30]. RFX5 is transcriptionally effective only when a) complexed to two additional subunits (RFXANK/B and RFXAP), b) adjacent promoter sites are filled with X2BP and NFY, and c) a tissue-specific coactivator protein, CIITA, is present and bound to the platform generated by this multiprotein array [29, 30].
The requirement for specialized promoter environments and cell lineages extends to other RFX proteins, although few details are available for proteins other than RFX1. Because of the overall similarities of RFX1–3 [26], RFX1 may serve as a model for the other two. RFX1 serves as a positive activator for enhancer 1 of the hepatitis B virus and other viral enhancers [57, and references therein], the transforming promoter of the Epstein-Barr virus [58], the ribosomal protein rpl30 promoter [59], the IL-5R gene in eosinophils [34], and others [26, 29]. In the case of Epstein-Barr virus, the RFX1 site is only functional as long as adjacent sites for CREB and BSAP are also filled [58]. In the IL-5R promoter, the RFX site lies next to an AP1 site. When RFX sites are simply multimerized and placed upstream of a reporter gene, their effect on gene expression in transfection assays is markedly dependent on the cell line used. In many cell lines, multimerized RFX1 sites have no effect on transcription [34]. Dissection of RFX1 has shown that it contains both transcriptional activation and repressive domains and that these tend to cancel one another out for the intact protein [34, 57]. Presumably, binding in the context of surrounding factors leads to a net expression of one or the other of the activities. In some contexts, the effect of RFX1 is inhibitory; for example, a site within intron I of the c-Myc gene [60], repression of chorionic somatomammotropin transcription in pituitary cells [61], suppression of microtubule-associated protein 1A in nonneuronal cells [62], or suppression of collagen [63]. Of particular interest, it remains to be determined if RFX1, and other RFX proteins, are dependent on tissue-specific coactivators, similar to CIITA. At present, very little is known about the contextual requirements of RFX2 for gene activation.
The binding site for RFX proteins in the H1t palindrome region is relatively weak compared with the strong IL-5R promoter site, particularly with in vitro-synthesized RFX2 (Fig. 7B). It may be that binding is cooperative with adjacent Sp1 and NFY sites (see Fig. 9). Indeed, for the MHCII promoters, it is well established that binding of the RFX5 complex is cooperative with the adjacent NFY site [33, 64]. However, in that case, the CCAAT element is oriented in the opposite direction, and cooperativity depends critically on the spacing between the two binding sites [64]. In the case of MHCII promoters, band-shift assays readily demonstrated formation of complexes that contain both RFX5 and NFY [33]. With the H1t promoter, we have not yet observed such complexes when using longer promoter fragments in testis NE (unpublished results). Detection of H1t mRNA by in situ hybridization [7, 11] correlates well with the point of maximum expression of RFX2 (Fig. 6), and it may be that high expression is required to fill the H1t promoter site.
The striking upregulation of RFX2 during development of spermatocytes prompted us to search for additional promoter sites in which RFX2 might play a role. We identified RFX binding sites in the 5' flanking regions of both the spermatocyte-specific lamin C2 transcript [40] and the mouse Sgy gene. While the lamin C2 site is comparable with the H1t site in binding, the Sgy site is substantially stronger. Lamin C2 results from a testis-specific promoter that lies within intron 1 of the lamin A gene [40, 65]. Sgy is a member of a family of secreted inhibitors of the Wnt signaling pathway [66]. At the same time, similarities to the CCTAGG central element of the H1t RFX site present in the c-mos [41] and Pdha2 [42] genes did not form RFX-like complexes in band-shift assays. This emphasizes the importance of the outlying nucleotides in the typical 14-bp RFX1 site.
The report of Wolfe et al. [35] and the results presented here provide the first potential targets for the enormously upregulated expression of RFX2 in the adult testis. Immunohistochemistry has shown that this expression coincides with the development of pachytene spermatocytes. It now remains to demonstrate clear functional connections between these promoter binding sites and gene activation and to determine whether specialized promoter contexts and novel coactivator proteins are required for gene activation during meiosis in the male.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Correspondence: M. Kistler, Department of Chemistry and Biochemistry, University of South Carolina, 631 Sumter Street, Columbia, SC 29208. FAX: 803-777-9521; mkistler{at}mail.chem.sc.edu
Received: 19 May 2004.
First decision: 14 June 2004.
Accepted: 22 June 2004.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Kim, D. Li, Y. Cui, K. Mueller, W. C. Chears, and J. DeJong Regulatory Factor Interactions and Somatic Silencing of the Germ Cell-specific ALF Gene J. Biol. Chem., November 10, 2006; 281(45): 34288 - 34298. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. I. Niesen, A. R. Osborne, H. Yang, S. Rastogi, S. Chellappan, J. Q. Cheng, J. M. Boss, and G. Blanck Activation of a Methylated Promoter Mediated by a Sequence-specific DNA-binding Protein, RFX J. Biol. Chem., November 25, 2005; 280(47): 38914 - 38922. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
