Association of Purα with RNAs Homologous to 7 SL Determines Its Binding Ability to the Myelin Basic Protein Promoter DNA Sequence*

Cell type and developmental stage expression of the myelin basic protein (MBP) gene in mouse brain is regulated at the transcriptional level. Earlier studies from our laboratory have led to the identification of a DNA binding protein from mouse brain, named Purα, which interacts with the MB1 regulatory motif of the MBP and stimulates its transcription in glial cells. In this report, we demonstrate that a cellular RNA, with significant homology to 7 SL RNA is associated with Purα. Results from band shift competition studies indicate that Purα-associated RNA (PU-RNA), inhibits the interaction of immunopurified Purα with the MB1 DNA sequence. Results from Northern blot studies indicated that PU-RNA is expressed during various stages of brain development. Of interest, this RNA was found in association with Purα that was produced in the mouse brain at the early stage of brain development. Results from Northwestern analysis using a PU-RNA probe identified the regions within Purα that are important for Purα/PU-RNA association. Production of Purα at the early stage of brain development and its association with PU-RNA at this stage, when Purα exhibits poor binding ability to the MB1 DNA sequence, suggests that PU-RNA may function as a co-factor that negatively regulates Purα interaction with the MBP promoter sequence.

Cell type-specific gene activation is controlled by tissue-specific transcription factors that usually recognize and interact with DNA sequences located within the promoter or enhancer. In some instances, binding of these regulatory proteins to the DNA sequence is facilitated upon self-association of the proteins or their interaction with other members of transcription factors. Such a protein-protein interaction may, in turn, alter the protein conformation and facilitate its binding to the DNA molecules and/or its communication with other transcription factors.
The focus of our investigations during the past 10 years has been to decipher the regulatory pathways that participate in cell type-specific gene transcription in the brain. We have focused our attention on the myelin basic protein (MBP) 1 gene, whose product is a major component of the myelin sheath in the central nervous system (for a review see Ref. 1), and its expression is controlled in manners both cell-and stage-specific during brain development (2)(3)(4). In mouse brain, expression of the MBP occurs postnatally such that it is first detected at the end of the first postnatal week, increases dramatically to peak at 18 -21 days, and decreases to about 20% of peak levels in mature animals (5). Earlier studies have indicated that programmed expression of MBP is regulated at the level of transcription (6 -8). Functional dissection of the MBP regulatory sequence led to the identification of a regulatory motif named MB1, which spans from nucleotide Ϫ14 to Ϫ50 with respect to the transcription start site. Analysis of nuclear proteins derived from mouse brain identified a 39-kDa protein that forms a nucleoprotein complex with a DNA fragment containing the MB1 sequence (9). Subsequent studies indicated that the 39-kDa protein, which is named Pur␣, recognized the GGC/GGArich sequences within MB1 DNA in a single-stranded configuration and has the ability to increase transcription of the MBP promoter both in vitro and in vivo (10,11). Of interest, Pur␣ binding activity to the MB1 sequence occurs in a developmental stage-specific manner that coincides with the pattern of MBP transcription (10). Evidently, at the early stage of brain development (days 3-7, postnatally), the level of Pur␣ association with MB1 is extremely low, whereas during the phase of myelination (18 -20 days) and in adults (day 30), its binding activity to MB1 drastically increases. Preliminary examination of Pur␣ production indicated that although its levels increase between the first and second week of brain development, a significant amount of Pur␣ is detected at the early stage of mouse brain development, which may not account for its extremely low level of association with the MB1 DNA sequence at this stage. These observations imply the participation of a co-factor(s) that may determine Pur␣ binding activity to the DNA molecule. The ability of Pur␣ to interact with a GGN repeat in a single-stranded form prompted us and others (12) to examine its binding ability to RNA molecules. Here, we describe our results demonstrating that Pur␣ obtained from brain extract is in association with RNA molecules that have the ability to inhibit its binding activity to the MB1 DNA. Of interest, this RNA, which we have named PU-RNA, has significant homology to the 7 SL RNA, is expressed during various stages of brain development, and is found in association with Pur␣ in 5-day-old mouse brain. Thus, our data suggest that PU-RNA functions as a co-factor by determining the binding activity of Pur␣ to the MBP promoter sequence and indirectly participates in the developmental activation of the MBP promoter in mouse brain.

EXPERIMENTAL PROCEDURES
Nuclear Extract Preparation-Nuclear extracts from mouse brains at different stages of development were prepared according to the method of Dignam et al. (13) except that the second wash with buffer A was eliminated. The concentration of protein extract was measured using the Bradford assay (Bio-Rad), and the extracts were stored at Ϫ70°C.
Immunoprecipitation and RNA Extraction-For immunoprecipitation, approximately 100 g of nuclear extract were incubated with 1 g of either nonimmune serum or clone 9C12 anti-Pur␣ antibody for 2 h at 4°C in buffer containing 12 mM HEPES (pH 7.9), 4 mM Tris-HCl (pH 7.5), 150 mM KCl, 2.5 mM CaCl 2 , 2.5 mM MgCl 2 , 0.1% Nonidet P-40, and 0.1 mM dithiothreitol. The immunocomplex was collected with 50 l of prewashed Pansorbin, and the pellets were washed three times in the above buffer. Subsequently, the pellets were resuspended in phosphatebuffered saline and either analyzed by the Western blot technique for detection of Pur␣ or used for extraction of RNA. For RNA extraction, total nucleic acids (DNA and RNA) were extracted from the pellets and the supernatants of the immunoprecipitate using phenol/chloroform extraction followed by ethanol precipitation. The nucleic acid pellets were redissolved in buffer containing 10 mM Tris-HCl (pH 7.0) and 30 mM NaCl and treated with RNase-free DNase for 15 min at room temperature in the presence of RNase inhibitor. The reaction mixture was subjected to phenol/chloroform extraction and ethanol precipitation. This step was repeated twice to ensure complete removal of the DNA molecules. The RNA samples were stored at Ϫ70°C in 75% ethanol until use.
Immunoprecipitation Band Shift Analysis-One hundred micrograms of mouse brain nuclear extracts were precipitated using anti-Pur␣ antibody according to the procedure described above with the exception that CaCl 2 and MgCl 2 were removed from the buffer. The immunoprecipitates were washed three times, and the protein-antibody complexes were eluted in 50 l of high salt buffer C (13) for 1 h at 4°C. Pansorbin was removed by centrifugation at 14,000 rpm for 1 min, and the supernatants that contained the eluted protein were collected for use. The eluate was diluted with low salt buffer B (13) to a final KCl concentration of 150 mM. For band shift studies, 5 l of immunopurified protein and 1 g of total nuclear extract were incubated for 30 min on ice in binding buffer containing 12 mM Hepes (pH 7.9), 4 mM Tris-HCl (pH 7.5), 2.5 mM CaCl 2 , 2.5 mM MgCl 2 , 50 mM KCl, 0.1 mM dithiothreitol, and 30,000 cpm of the 5Ј-32 P-labeled MB1A oligonucleotide DNA (5Ј-TCAGAGGGCCTGTCTTTGAAGGTG-3Ј) or MB1A ribonucleotides in the presence or absence of various amounts of DNA and RNA competitors, as indicated in each experiment. PU-RNA competitor was prepared by in vitro transcription using T7 polymerase. For RNase treatment, the protein and/or extracts were incubated with or without RNase for 15 min at room temperature prior to the addition of the labeled probe. All incubations for binding reaction were performed in the presence of 5 g of bovine serum albumin and 1 g of poly(dI/dC). The protein-DNA complexes were resolved on 6% native polyacrylamide gels (19:1) (Bio-Rad) in 0.5ϫ TBE buffer at constant (150 V) voltage.
Reverse Transcription-Polymerase Chain Reaction (RT-PCR)-The RNAs obtained from the immunocomplexes and the supernatant of the immunoprecipitation reactions using 100 g of nuclear extracts were used as a template for reverse transcription assay. The RNA was denatured in transcription buffer provided by the manufacturer (Boehringer Mannheim) containing the specific BC200-derived downstream primer (5Ј-AGAGAACGGGGTCTCGC-3Ј) at 75°C for 15 min. The primer was then allowed to anneal to the RNA molecules as the reaction was gradually adjusting to 42°C. AAV Reverse Transcriptase (Boehringer Mannheim) was used to synthesize the cDNAs. The cDNA synthesis was carried out at 42°C for 1 h in the presence of the mixture of 0.1 mM dNTP and the RNase inhibitor. The reaction was terminated by heat inactivation of the enzyme at 95°C for 10 min. For low cycle PCR, one-fifth of the total cDNA was used as a template in the PCR. The amplification reaction was carried out using 10 cycles of PCR with the following cycle parameters: denaturation at 95°C for 30 s, annealing at 55°C for 30 s, and synthesis at 72°C for 1 min. The RT primer along with the upstream primer (5Ј-CGTGCCTGTAGTCCCAG-3Ј) derived from BC200 was used in the PCRs. For cloning of the amplified DNA in the TA vector (Invitrogen), the PCR was performed at 35 cycles, and a 10-min extension at 72°C was included at the end of the last PCR cycle. The PCR product was cloned into the TA vector according to the manufacturer's instructions (Invitrogen). Direct sequencing of the PU-RNA cDNA was performed using both upstream and downstream BC200derived primers. Sequencing was performed by the Sanger methods (14). The Sequenase version kits were used according to the manufacturer's instructions (version 2.0, Amersham Pharmacia Biotech). The [␣-32 P]dATP or the [␣-35 S]dATP labeled DNA fragments were resolved on a 6% denaturing (urea) acrylamide gel (19:1) (Bio-Rad). The data from sequencing reactions were verified in at least five independent experiments. For Southern blot hybridization, the PCR products were resolved on a 2% agarose gel, and the DNAs were transferred onto Hybond membranes (Amersham Pharmacia Biotech). After cross-linking, the membranes were prehybridized at 65°C in buffer containing 6ϫ SSC, 10 mM EDTA, 0.5% SDS, 5ϫ Denhardt's solution, and 0.1 mg/ml salmon sperm DNA. After 6 h, the DNA probe that was generated from PU-RNA cDNA using the Redi-prime kits (Amersham Pharmacia Biotech) was added directly to the prehybridization solution, and hybridization was carried out for 16 h. The membranes were washed twice in 2ϫ SSC, 0.1% SDS at 65°C and with 0.1ϫ SSC, 0.1% SDS at 65°C. The membrane was air-dried and exposed overnight to Kodak x-ray film.
Northwestern Analysis-PU-RNA probe was synthesized using a linearized pCDNA3 vector containing the PU-RNA cDNA as a template and the T7 RNA polymerase (Boehringer Mannheim). Labeling of the RNA and transcription reaction was carried out by the method provided by the supplier (Boehringer Mannheim), where the newly synthesized PU-RNA was uniformly labeled with [ 32 P]UTP. The RNA probe was treated with RNase-free DNase to remove the template DNA, and the unincorporated isotope was removed by G50 spin columns (Life Technologies, Inc.). The full-length GST fusion Pur␣ and the various Pur␣ deletion mutants were resolved on 10% SDS-polyacrylamide gel electrophoresis and were transferred to the transblot membrane (Bio-Rad) for 16 h at 4°C using semidry transfer at a constant current of 125 mA. After transfer, the membrane was treated with binding buffer containing 12 mM Hepes (pH 7.9), 4 mM Tris-HCl (pH 7.5), 2.5 mM CaCl 2 , 2.5 mM MgCl 2 , 50 mM KCl, 0.1 mM dithiothreitol, and 5% nonfat dry milk for 30 min at room temperature. The membrane was incubated for an additional 30 min in binding buffer containing 0.5% nonfat dry milk and 50 g/ml tRNA. The labeled RNA probe (10 7 cpm) was added to the binding buffer, and the blot was incubated for 3-5 h at room temperature. All washes (usually 4 -6 washes, for approximately 10 min each) were carried out in the presence of 0.5% dry milk and 50 g/ml tRNA in binding buffer. The membranes were exposed to Kodak x-ray film overnight.
Northern Blot Analysis-Total RNA was extracted from mouse brain at various stages of development by a modified guanidine isothiocyanate method (14). Approximately 20 g of total RNA isolated from mouse brain tissue were analyzed on denaturing formaldehyde-agarose gel and hybridized to the PU-RNA cDNA probe after transfer to nitrocellulose as described previously (14).

RESULTS
Previously, we identified and molecularly cloned a gene whose product, Pur␣, increased the level of transcription from the MBP promoter both in vitro and in vivo (10,11). Pur␣ associates with the MBP regulatory motif, MB1, which is located between nucleotides Ϫ14 and Ϫ50, with respect to the transcription start site and contains the putative Pur␣ binding site, GGNGGN. Our preliminary observations, along with the results from other laboratories, suggested that Pur␣ has strong binding affinity to single-stranded DNA and associates with RNA molecules (11,12). In order to further investigate the association of Pur␣ with RNA and to determine the importance of such an interaction in binding of Pur␣ to the MB1 DNA sequence, mouse brain nuclear extract was treated with anti-Pur␣ antibody, and the ability of the immunopurified Pur␣ to bind to the MB1 DNA probe was examined by band shift assay. As shown in Fig. 1A, the immunopurified Pur␣ bound to the MB1 DNA probe and formed a complex with electrophoretic mobility similar to that seen upon binding of Pur␣ from the unfractionated nuclear extract and MB1 probe (compare lanes  2 and 3). The protein content of the immunocomplexes obtained by the control serum showed no binding ability to the MB1 probe (Fig. 1A, lane 4). Fig. 1A (bottom) illustrates results from Western blot analysis of total brain nuclear extract (lane 1) and the immunopurified Pur␣ protein obtained from brain nuclear extract (lane 3). The band corresponding to the 39-kDa Pur␣ was detected in total nuclear extract, and the immunocomplex was obtained from the extract treated with anti-Pur␣ antibody. To gain some evidence regarding the association of RNA mol-ecules with the immunopurified Pur␣ and to evaluate the effect of such an association on the Pur␣-MB1 assembly, total ribonucleic acid was extracted from the Pur␣ immunocomplex and was used as a competitor in an MB1-directed band shift assay. As shown in Fig. 1B, inclusion of 1 ng of PU-RNA in the DNA binding reaction decreased the interaction of immunopurified Pur␣ with the MB1 DNA probe (compare lanes 2 and 3). At a higher concentration (3 ng) of PU-RNA, no signal corresponding to the Pur␣-MB1 complex was detected (Fig. 1B, lane 4). The addition of RNase to the binding reaction containing 3 ng of PU-RNA restored the assembly of the Pur␣-MB1 complex (Fig. 1B, lane 5), suggesting that the observed inhibitory effect is dispensable and mediated by the RNA molecules. The addition of PU-RNA and RNase to the MB1 DNA alone had no effect on the electrophoretic mobility of the MB1 probe (Fig. 1B, lane  6). To further examine the specificity of this observation, the competition experiment was repeated with 3 ng of PU-RNA (PU) or an equal amount of RNA obtained from the supernatant of the immunoprecipitate (S). As shown in Fig. 1C, whereas the PU-RNA effectively blocked association of Pur␣ to the MB1 probe, the supernatant RNA did not abrogate the formation of the Pur␣-MB1 complex. These observations indicate that RNA species are associated with the immunopurified Pur␣ and that the participant RNA molecules have the ability to control DNA binding activity of Pur␣.
The computer-assisted search for RNAs with the Pur␣ binding site led us to suspect that Pur␣ may be associated with the RNAs encompassing Alu core sequences, the GGAGGC repeat (15). The 7 S RNAs represent the only known abundant RNA type that contains the Alu-like core repeat sequence (16 -18). Of interest, in neural cells brain-specific RNAs, named BC1 and BC200, represent the transcripts of transcriptionally active Alu genes and contain a high degree of homology to 7 S RNA (19 -22). In order to examine whether immunopurified Pur␣ is associated with 7 S or BC200 RNAs, we designed oligonucleotides derived from the conserved regions of these transcripts and used them as primers for cDNA priming and PCR amplification in a reaction containing Pur␣-associated RNA molecules. Under optimum conditions, no DNA fragment corresponding to the expected 95-base pair species indicative of 7 S or BC200 sequences was amplified. However, under similar conditions, an approximately 290-base pair fragment was obtained, indicating that the PCR primers were able to hybridize to the regions further upstream from the designated region and amplify a larger DNA fragment. Fig. 2 illustrates the primary sequence of the central region of the amplified fragment, which exhibits significant homology to 7 SL RNA. Despite the unexpected size of the amplified DNA fragment, an advanced gapped BLASTN search confirmed that the amplified cDNA corresponding to the PU-RNA sequence may belong to the 7 SL family of cellular RNA and has the highest homology score of 613 and a probability event of 4.2e Ϫ 78. Fig. 2 illustrates the alignment of PU-RNA to the neuron-specific BC200 (BC) and left arm of 7 SL RNAs (7S), where the primers for RT-PCR were generated and to the 7 SL (SL) RNA with the highest homology score. PU-RNA differs significantly from the BC200 and left arm of 7 SL, which were used to generate primers, as examined by an advanced BLASTN search, which was optimized to find only the nearly identical sequences. The @ symbol depicts the nucleotides that are conserved. As shown in Fig. 2, only two single base substitutions and one single base deletion distinguish PU-RNA from the highest homology human 7 SL sequence. In order to identify putative Pur␣ binding sites positioned in the single-stranded areas of PU-RNA, secondary prediction was determined. PU-RNA shows extensive stemloop structure with the stability of Ϫ69.7 kcal/mol at 25°C. Through this analysis, we were able to identify at least two putative Pur␣ binding sites (GGN) that are located within the loops of the PU-RNA (data not shown).
In a different series of studies, we assessed the fraction of Pur␣-associated RNA that encompasses sequences with homology to 7 SL RNA. Results from sequencing of 12 clones obtained by RT-PCR of Pur␣-associated RNAs utilizing random primers revealed that five clones have homology to the Alu repeat, which is common in 7 SL RNAs (data not shown).
In the next series of studies, we examined the level of Pur␣ and the associated PU-RNA at various stages of brain development. As shown in Fig. 3A, the 39-kDa Pur␣ protein was detected by Western blot analysis of the brain protein extracts from mouse at 5, 10, 12, and 45 days after birth (lanes 1-4, respectively) with the notion that its levels increased at day 10 and remained virtually constant thereafter. In parallel, we performed immunoprecipitation of the extract from mouse brain at various ages utilizing anti-Pur␣ antibody, and the RNA molecules were extracted from Pur␣ immunocomplexes and were used in an RT-PCR assay for detection of the 295-bp PU-RNA. As shown in Fig. 3B, the 295-bp PU-RNA was detected in the Pur␣ immunocomplex derived from 5-day-old mouse brain. In order to examine the level of PU-RNA total RNAs were obtained from mouse brain at various stages of development and examined by Northern blot technique. As shown in Fig. 3C, strong signals indicative of a high level of PU-RNAs were detected at the early, middle, and late stages of mouse brain development (lane 1). The integrity of the RNAs is demonstrated in Fig. 3D, where the levels of 28 and 18 S RNAs remain unchanged. Our observation on the association of PU-RNA with Pur␣ at the early stage of brain development is interesting in light of our earlier studies (10) demonstrating that association of Pur␣ with MB1 DNA is developmentally regulated, with extremely low levels at the early stages of brain development (2-5 days) and maximizing in adults. The observed discrepancy in the abundance of Pur␣ in newborn mice (shown in Fig. 3A) and its extremely low binding ability to MB1 suggested that binding of Pur␣ to MB1 DNA at the early stage of brain development may be regulated by a distinct mechanism. The inhibitory action of PU-RNA on the DNA (MB1) binding activity of Pur␣ and its association with Pur␣ produced in 3-5-day-old mouse brain suggested that PU-RNA may play a role in down-modulating binding of Pur␣ to MB1 DNA at the early stage of brain development. Thus, one may envision a model whereby PU-RNA, by inhibiting DNA binding activity of Pur␣ may indirectly affect MBP gene transcription at early, but not late, stages of brain development. In accord with this concept, results from band shift studies indicated that RNase treatment of the nuclear extract from 5-day-old mouse brain induces binding of nuclear proteins to the MB1 DNA probe. In this study, nuclear extracts from 5-day-old, 10-day-old, and adult (60-day-old) mouse brain were mixed with the oligonucleotide probe containing the Pur␣ binding site, and the nucleoprotein complex was analyzed by native gel. Consistent with   4) were mixed with the oligonucleotide DNA probe containing a Pur␣ binding site as described previously (10). The arrow indicates the position of the Pur␣-DNA complex. B, DNA binding reaction mixture derived from 5-day-old brain extract (lanes 1 and 2) and adult brain (lanes 3 and 4) were incubated with and without RNase prior to gel electrophoresis. our previous data (10), a significant increase in Pur␣ binding activity was detected in the mature brain extract (Fig. 4A,  compare lanes 2 and 4). Treatment of the extract with RNase enhanced association of Pur␣ from 5-day-old mouse brain extract to the DNA probe and exhibited an insignificant effect on binding activity of Pur␣ from adult brain extract. These observations, along with the data presented in Fig. 1, strongly suggest that the association of Pur␣ with the RNA molecule modulates DNA binding activity of Pur␣.
To directly investigate the effect of PU-RNA on the association of Pur␣ to the MB1 DNA sequence, various amounts of in vitro transcribed PU-RNA were used as competitors in DNA binding studies. As shown in Fig. 5A, inclusion of PU-RNA in the binding reaction decreased the intensity of the band corresponding to Pur␣ association with the MB1 DNA probe. In a separate series of studies, we compared the relative binding affinity of Pur␣ to DNA and RNA molecules with similar nucleotide compositions. Toward this end, synthetic MB1 ribonucleic acid was used as a probe in the binding reaction in the absence and presence of unlabeled MB1 RNA, MB1 DNA, and nonspecific RNA competitors. As illustrated in Fig. 5B, both MB1 RNA and MB1 DNA were able to inhibit PU-RNA-Pur␣ complex formation. From the intensity of the ribonucleocomplex, however, it is evident that under the identical conditions, MB1 DNA is more effective than the MB1 RNA in dissociating the PU-RNA-Pur␣ complex.
The primary amino acid sequence of Pur␣ demonstrates the well defined modular nature of the protein, which has no strong homologies to the known proteins (23,24). The structural domains within Pur␣ include (i) the stretch of 18 glycines, interrupted by one serine, in the N-terminal region of the protein, (ii) three basic-aromatic (class I) and two acidic-leucine (class II) repeats in the middle portion of the protein, and (iii) the region of limited homology to T-antigen, which includes the putative phosphorylation site for casein kinase II and is fol-lowed by glutamine-and glutamate-rich domains (23). In an attempt to determine the region within Pur␣ that is responsible for its association with PU-RNA, we performed Northwestern experiments utilizing PU-RNA as a probe. To carry out these experiments, we created and utilized recombinant cDNA clones expressing different regions of Pur␣ in the prokaryotic system. Fig. 6A illustrates the structural organization of wild type and the various mutants of Pur␣ that were used in this experiment. The glutathione S-transferase (GST) fusion proteins containing the various regions of Pur␣ as illustrated in Fig. 6A were produced in bacteria and resolved by SDS-polyacrylamide gel electrophoresis. The fractionated proteins were transferred to nitrocellulose, and the transblot was reacted with the riboprobe containing PU-RNA. The equal loading of proteins was verified by SDS-polyacrylamide gel electrophoresis followed by Coomassie Blue staining of the proteins (data not shown). As shown in Fig. 6B, PU-RNA bound to the fulllength Pur␣. Deletion of the first 85 N-terminal amino acids resulted in a severe reduction of PU-RNA binding to the N-85 mutant. None of the subsequent N-terminal Pur␣ deletion mutants were able to bind to PU-RNA. The first C-terminal mutant (mutant 1-215), which lacks amino acids 216 -322, exhibited binding activity to PU-RNA. Further C-terminal deletion mutant (mutant 1-174), which deletes amino acids 174 -322, completely lost its ability to bind to PU-RNA. Neither the internal deletion mutant nor the 85/230 mutant, which consists of the sequence deleted in mutant ⌬72-231, was able to bind to PU-RNA. These results suggest that at least two distinct regions of Pur␣, located at the N terminus and between amino acids 174 and 215, are required for binding of Pur␣ to RNA; however, neither of these regions alone is sufficient for binding. Cooperation between these two regions of the protein is probably responsible for the binding of Pur␣ to PU-RNA. Also, our data indicate that the DNA binding site of Pur␣, which is located between amino acid residues 66 and 246, overlaps with the region responsible for its binding to RNA. time, the 7 S RNA has been recognized as a component of the signal recognition particle, which is involved in transport of the nascent peptides into the endoplasmic reticulum (25,26). 7 SL RNA is highly conserved between species, and from an evolutionary perspective, it is thought to serve as an RNA intermediate in the retrotransposition of the middle repetitive genomic Alu sequences (16,17). Here we report that the transcription factor Pur␣, which regulates expression of the differentiationspecific MBP gene in oligodendrocytes (10,11) engages in a complex with RNAs containing the 7 SL sequence, called PU-RNA, at the early stages of brain development. We speculate that in addition to its lower level, association of Pur␣ with the PU-RNA, overproduced in the premyelinating brain, may account for the functional inactivation of this transcription factor before the onset of myelination. Accordingly, a decrease in the level of PU-RNA at the later stages of development and its dissociation from Pur␣, which permits interaction of Pur␣ with MB1, may represent a mechanism for the maximum transcription of MBP expression during the peak of myelination. Thus, one may hypothesize that the 7 SL RNA sequence, by associating with or dissociating from a regulatory protein, may function as a co-factor for control of gene transcription. Of note, earlier studies indicated that 7 SL RNA may be associated with the glucocorticoid receptor, a protein that can regulate transcription of a variety of genes (27,28).
In a previous report, Hereault et al. (12) reported that RNA molecules can stimulate interaction of a single-stranded DNA binding protein with the GC-rich DNA motif. Note that while the reported protein and Pur␣ both target a GC-rich singlestranded DNA, the former has a molecular mass (29 kDa) different from that of Pur␣.
As mentioned earlier, Pur␣ has a complex structure with two series of interspersed repeats, a glutamine-rich domain (a potential candidate for an activation domain), a region of amphipathic helix, and a glycine-rich domain. Earlier studies indicated that a region positioned between amino acids 66 and 246 is critical for its DNA binding activity (29). Our observation of the requirement for the glycine rich motif in order to bind to the PU-RNA is consistent with the presence of similar glycinerich motifs within the known RNA binding proteins (23). Class I repeats were previously implicated in the DNA binding ability of Pur␣ (23). These regions, however, do not seem to be important for binding to RNA, since the deletion of class I repeats did not abolish binding to RNA. The class II repeats were shown to be involved in Pur␣ interaction with human immunodeficiency virus-1 regulatory protein, Tat (30). Surprisingly, one of the class II repeats appeared to be required for the binding of Pur␣ to RNA. These results raise the possibility that Pur␣-Tat interaction may be mediated by RNA. Perhaps it should be noted that involvement of RNA molecules in the interaction of proteins such as dimerization of estrogen and progesterone receptors has been previously demonstrated (31)(32)(33).
Repetitive RNA species are differentially expressed in cells in response to various conditions, such as the following: release by serum (34), stress (35), insulin (36), treatment with carcinogens (37), brain pathology (38) cell transformation (40 -43), and in response to viral regulatory proteins (44). Interestingly, 7 SL RNA and the Alu transcripts can inhibit cell proliferation when overexpressed in HeLa cells (45). The possible mechanism of growth inhibitory activity stems from recent work where in vitro selected RNAs were shown to suppress cell growth via inhibition of binding of E2F1 to the DNA (46). It has long been assumed that promoter and enhancer binding factors recognize their cognate site on duplex DNA. Results from several laboratories have demonstrated that several transcription regulatory proteins, including homeodomain proteins, can bind to RNA and exhibit a regulatory effect on both transcription and translation (47)(48)(49). Earlier studies have demonstrated that Alu transcripts and 7 SL RNA can also stimulate transcription of several genes, including SV40, c-myc (41,42), and the heat-stable enterotoxin receptor (50) promoters via unknown mechanisms. Based on the data presented in this paper, we propose that repetitive RNAs may also be involved in the regulation of expression of differentiation-specific MBP gene through their binding to transcriptionactivating protein Pur␣.