Kon Yew Kwek ¹, Shona Murphy ¹, Andre Furger ¹, Benjamin Thomas ¹, William O'Gorman ¹, Hiroshi Kimura ², Nick J. Proudfoot ¹ and  Alexandre Akoulitchev ^{1, @}

¹ Sir William Dunn School of Pathology, University of Oxford, South Parks Road, Oxford OX1 3RE, UK.
² Medical Research Institute, Tokyo Medical and Dental University, Yushima 1-5-45, Bunkyo-ku, Tokyo 113-8510, Japan.

^@ Correspondence should be addressed to A. Akoulitchev.
e-mail: alexandre.akoulitchev@path.ox.ac.uk

Diverse classes of noncoding RNA, including small nuclear RNAs (snRNAs), play fundamental regulatory roles at many stages of gene expression. For example, recent studies have implicated 7SK RNA and components of the splicing apparatus in the regulation of transcriptional elongation. Here we present the first evidence of the involvement of an snRNA in the regulation of transcriptional initiation. We demonstrate that TFIIH, a general transcription initiation factor, specifically associates with U1 snRNA, a core-splicing component. Analysis of the TFIIH-dependent stages of transcription in a reconstituted system demonstrates that U1 stimulates the rate of formation of the first phosphodiester bond by RNA polymerase II. In addition, a promoter-proximal 5' splice site recognized by U1 snRNA stimulates TFIIH-dependent reinitiation of productive transcription. Our results suggest that U1 snRNA functions in regulating transcription by RNA Polymerase II in addition to its role in RNA processing.

Noncoding RNAs, particularly small nuclear RNAs (snRNAs), play fundamental roles in regulating transcription by RNA polymerase II (Pol II) and in processing of the transcripts [1]. For example, 7SK RNA regulates elongation of transcription [2, 3], U1–U6 are required for splicing and elongation [4] U7 directs 3' processing of histone mRNA [5]. However, until now no noncoding RNA was shown to influence transcriptional initiation, a critical regulatory stage of gene expression.

Initiation of transcription by Pol II requires a minimal set of general transcription factors and can be reconstituted in vitro with highly purified components. In such a minimal system, initiation remains largely insensitive to transcriptional activation and multiple rounds of reinitiated transcription do not occur, two main features of regulated gene expression in vivo [6].

Here we describe the specific association of U1 snRNA with one of the general transcription factors, TFIIH. TFIIH is a multisubunit factor directly implicated in Pol II initiation [6-9]. Because U1 participates in recognition of the 5' splice site [5], the association between TFIIH and U1 snRNA provides a potential connection between initiation of transcription and mRNA splicing.

We demonstrate that the association of TFIIH with U1 stimulates the rate of initiation by Pol II. Moreover, in a reinitiation assay following the first round of transcription, the presence of a 5' splice site increases efficiency of reinitiation from the scaffold [6] because of the presence of U1 and TFIIH. This result is particularly important in view of the in vivo data linking the strength of eukaryotic promoters with the position of the promoter proximal intron [10, 11].

TFIIH associates with RNA

To determine whether any noncoding RNAs plays a role in initiation of transcription, we purified the general transcription factors that are required and sufficient for reconstituted initiation on nucleosome-free ('naked') DNA templates [8] and then tested them for association with RNA. Transcription factors TFIID, TFIIB, TFIIF, TFIIE, TFIIH and Pol II were prepared as described [12]. When these factors were incubated with T4 RNA ligase and [5'-³²P]-pCp², a distinct labeled product of ~160 nucleotides is present in the general transcription factor TFIIH purified by both conventional [12] (Fig. 1b, lane 1) and affinity [13] (Fig. 1a,b, lane 3) methods. No other factors required for initiation of transcription in this minimal system contained any pCp-labeled products (data not shown). Treatment of TFIIH by RNase A, but not by RNase-free DNase I (Fig. 1c, lanes 1 and 2), abolished pCp labeling, confirming the presence of RNA in the sample.

Figure 1: Association of TFIIH with RNA.

a, Coomassie blue staining of affinity-purified TFIIH (lane 1).

b, pCp labeling of RNA in TFIIH purified by conventional (lane 1) and immunoaffinity (lane 3) methods. Lanes 2 and 4 are negative controls for each of the purification methods. Lane 5 represents RNA marker (Sigma) labeled with [³²P]-pCp. Samples were separated on a 6% sequencing gel and scanned on a Storm 840 phosphorimager.

c, TFIIH was pCp-labeled following treatment with DNase I (lane 1) or RNase A (lane 2).

We subsequently cloned and sequenced the TFIIH-associated RNA. All clones proved identical to the GenBank database entry #J00318 corresponding to the mature, 3'-processed U1 snRNA. To further demonstrate the identity of the associated RNA, we used RNase protection assay with a probe complementary to nucleotides 12–75 of U1 snRNA. The result confirmed that our preparations of TFIIH contained U1 snRNA (Fig. 2a, lane 2).

Figure 2: U1 snRNA specifically associates with TFIIH.

a, RNase protection of U1 snRNA. Lane 1 contains undigested probe; lane 2, total RNA; lane 3, TFIIH; and lane 4, negative control.

b, Western blot against cyclin H. Titration of purified recombinant His-tagged cyclin H (pmol) and TFIIH (ul) that were used in RNase protection assays.

c, Western blot of the pull down with U1 snRNA-complementary biotinylated oligonucleotides (lane 3) and control oligonucleotides (lane 4). Nuclear extracts (lane 1) and TFIIH (lane 2) represent positive controls.

d, UV crosslinking of U1 and TFIIH. Internally labeled synthetic sense U1(+) and anti-sense U1(-) oligonucleotides were crosslinked to TFIIH, treated with T1 RNase, resolved on 10% SDS PAGE gel and transferred to PVDF membranes. The membrane was autoradio graphed (lanes 1 and 2) and assayed by western blot for the presence of the p34 subunit of THIIH (lanes 3 and 4). Immunoprecipitation of crosslinked TFIIH with antibodies against subunits p34 and p62 were performed following TFIIH disruption in 1% (w/v) SDS (lanes 5 and 6).

e, Western blot of purified TFIIH (lane1) and nuclear extracts (lane 2) against U1 70K and SmB components of U1 snRNP.

U1 snRNA–TFIIH interaction is specific

We tested the specificity of the association between U1 and TFIIH using a pull-down assay. A biotinylated DNA oligo complimentary to the region of nucleotides 12–75 of U1 was incubated with purified TFIIH (Fig. 2c). A pull down with streptavidin-coated magnetic beads specifically precipitated the p62 and cdk7 subunits of TFIIH. Furthermore, these subunits were precipitated only when the U1-complementary oligonuleotide, but not a control oligonucleotide, was used (Fig. 2c, lane 3).

To further test the specificity of the interaction, we crosslinked affinity purified TFIIH with recombinant U1 snRNA (Fig. 2d). We used ³²P-internally labeled sense and anti-sense full-length U1 snRNA probes (Fig. 2d, lanes 1 and 2). After crosslinking with the U1 sense probe followed by T1 RNase digestion, a 34 kDa product was detected by autoradiography. Mobility of the band corresponded to that of the cyclin H subunit of TFIIH (lanes 1 and 2 are autoradiographs of lanes 3,4). To confirm the nature of the subunit, the crosslinked TFIIH was disrupted by boiling in 1% SDS and individual subunits were immunoprecipitated directly. Antibodies against cyclin H (p34) (lane 5), but not against p62 (lane 6), precipitated the crosslinked subunit. Therefore, we conclude that U1 snRNA interacts specifically with the cyclin H subunit of TFIIH.

We tested affinity-purified TFIIH containing U1 snRNA for the presence of several spliceosome components, known to associate with U1 in a mature U1 snRNP5. Protein gel staining (Fig. 1a; data not shown) did not reveal any polypeptide components present in stoichiometric amounts apart from the nine subunits of TFIIH itself. Moreover, western blot analysis and nanospray ionization mass spectrometry of the affinity-purified TFIIH sample confirmed the presence of p89, p82, p62, cdk7 and cyclin H subunits of TFIIH, but failed to detect U1 70K and Sm B splicing components (Fig. 2e). We therefore conclude that purified TFIIH is not associated with the mature U1 snRNP.

U1 snRNA stimulates abortive initiation

To assess the functional relevance of the interaction between the U1 snRNA and TFIIH during transcription, we tested the effect of U1 snRNA on reconstituted TFIIH-dependent abortive [14] and productive transcription. When transcription is assayed on the adenovirus major late promoter (AdML) in the presence of ATP and [a-³²P]-CTP, Pol II can advance only to the second position from the start site because the third position requires GTP. As a result, Pol II can catalyze the formation of only the first phosphodiester bond, and an ApC dinucleotide is released as a product of the abortive initiation cycle. Levels of released dinucleotides reflect the initiation potential of the pre-initiation complex. In the presence of all four ribonucleoside triphosphates (productive transcription assay), Pol II is able to proceed past the second nucleotide of the template. A small fraction of Pol II molecules eventually escape the promoter and advance along the template in its elongation mode, and TFIIH dissociates from these populations of elongating polymerases [8]. The advance of Pol II along the G-less transcriptional templates is readily detected, because the produced transcripts are resistant to treatment by RNase T1.

To analyze the role of U1 in initiation, we dissociated U1 from TFIIH by gel filtration under high salt conditions. Samples of TFIIH containing U1 snRNA were loaded onto a Superose 6HR 10/30 column in the presence of either 1 M or 300 mM NaCl (Fig. 3).

Figure 3: Dissociation of U1 snRNA from TFIIH.

Western blot of the p89, p62 and cdk7 subunits of TFIIH from the Superose 6 HR column profile run at 1 M or 300 mM NaCl. Transcriptional activity of TFIIH fractions was assayed in TFIIH-dependent productive transcription on the AdML promoter, and the peaks of activity were tested in the U1 RNase protection assay (lanes 1 and 2).

Analysis of the eluate for the p89, p62 and cdk7 subunits of TFIIH by western blot demonstrated slight difference in the elution profile, most likely resulting from increased dissociation of the CAK and CAK-ERCC3 subcomplexes from TFIIH at 1 M NaCl conditions, as decribed [15]. The western blot and a standard TFIIH-dependent productive transcription assay identified the fractions of TFIIH from the preparations at the different salt conditions with the same polypeptide composition and specific transcription activity. Transcription assays were conducted in the presence of T1 RNase (Fig. 3). RNase protection indicated that the U1 snRNA is displaced from transcriptionaly active TFIIH in the presence of 1 M NaCl (Fig. 3).

A comparison of the activity of TFIIH with and without U1 snRNA  in the TFIIH-dependent transcription assay indicates that removal of U1 snRNA reduces the activity of TFIIH (Fig. 4a). A similar comparison of TFIIH activity in the abortive initiation assay indicates that initiation is reduced >10-fold in the absence of U1 snRNA (Fig. 4a).

Figure 4: U1 snRNA stimulates TFIIH-dependent abortive initiation by RNA polymerase II.

a, Productive and abortive transcription with the titration of TFIIH fractions with and without U1 (lanes 2–6 and 7–11, respectively) in TFIIH-dependent (no TFIIH in lane 1) productive (AdML) and abortive (ApC) initiation assays. Transcripts were resolved on 6% and 15% sequencing gels and scanned using a phosphorimager.

b, Abortive initiation assay with TFIIH treated with T1 RNase (lanes 3 and 4) and micrococcal nuclease (lanes 1, 2 and 5–9). TFIIH pre-treated with micrococcal nuclease was complemented with synthetic U1 snRNA (lanes 7–8), E. coli tRNA (lane 9) or synthetic U2 snRNA (lanes 5–6).

c, Productive transcription assays of U1-containing and U1-free TFIIH preparations (lanes 1 and 2, respectively) pre-treated with T1 RNase at the stage of pre-initiation complex formation (lanes 3 and 4).

To further investigate the direct role of U1 snRNA in induced abortive initiation, we took advantage of the observation that U1 snRNA was the only snRNA detected by pCp labeling in the reconstituted system (see above). We destroyed U1 snRNA by digesting TFIIH with T1 RNase, which does not target the ApC product of the abortive initiation assay (Fig. 4b, lanes 3 and 4), or by digesting TFIIH with micrococcal nuclease in the presence of Ca²⁺. The nuclease was inhibited with EGTA prior to the abortive initiation assay (Fig. 4b, lanes 1 and 2). In both cases, levels of abortive initiation were significantly reduced. Consistent with this result, treatment with RNase T1 before transcription also abolished the difference observed earlier in the productive transcriptional assay (Fig. 4a,c). The role of U1 in abortive initiation was further confirmed when, following treatment with micrococcal nuclease, addition of synthetic U1 snRNA restored high rates of initiation (Fig. 4b, lanes 7 and 8). Neither Escherichia coli tRNA (lane 9) nor synthetic U2 snRNA (lanes 5 and 6) had any effect. These results indicate that U1 snRNA specifically stimulates the rate of TFIIH-dependent abortive initiation.

Role of U1 snRNA in transcriptional reinitiation

The role of TFIIH in transcription is not limited to abortive initiation. TFIIH has been implicated in promoter escape by Pol II [16, 17], as well as in transcriptional reinitiation [6]. Hahn and co-workers [6] have demonstrated that, upon initiation of the first round of transcription, a TFIIH-dependent reinitiation scaffold is formed on immobilized templates, with an important role in the subsequent round of transcription. In vivo evidence links transcription efficiency and the positioning of the first intron [10, 11]. Therefore, we tested whether the efficiency of reinitiated transcription was dependent on the presence of U1 snRNA and the promoter proximal 5' splice site.

The b-globin intron sequences [18] containing 5' (SD) and 3' (SA) splice sites were introduced into the AdML promoter template. We compared the in vitro efficiency of the first and second rounds of transcription on immobilized templates containing the wild type and mutant 5'/3' splice sites (mut SD/SA). In full agreement with the earlier data [6], western blot of the immobilized templates (Fig. 5b) confirmed that TFIIH was retained on the templates after the first round of transcription in a complex functionally equivalent to the described reinitiation
scaffold [6].

In the first round of transcription in the presence of the nuclear extracts, all the constructs containing the promoter supported similar levels of transcription (Fig. 5a).

Figure 5: Reinitiation of transcription is dependent upon U1 snRNA and the proximal 5' splice site.

a, The templates containing AdML promoter (lane 2–10) and intron sequences with wild type or mutant SD/SA sites (lane 3–10) were tested in the first and the second rounds of productive transcription (lanes 1–5) as described. The template containing the wild type SD/SA sites (lane 3 and 6–10) was tested in the second round after RNase H treatment in the presence of U1-complementary (nucloetides 12–75) (lanes 7–8) and control (lanes 9–10) DNA oligonucleotides.

b, The reinitiation scaffolds formed on the immobilized templates were assayed in western blot for the presence of TFIIH.

To test whether high levels of transcription depend on U1 snRNA, we targeted its destruction within the reinitiation scaffold with RNase H by using a DNA oligonucleotide complementary to nucleotides 12–75 of U1 snRNA. The efficiency of reinitiated transcription in the presence of the wild type intron was reduced after treatment with the U1-complementary, but not with the control, oligonucleotide, indicating a role for U1 snRNA in transcription reinitiation (Fig. 5a). The same treatment of the reinitiation scaffold did not affect the already low levels of reinitiated transcription obtained when the 5' splice site was mutated (Fig. 5a, lane 4; data not shown). We consider the observed effects of up to three-fold as significant, because the effect likely accumulates with each round of transcription. The role of TFIIH as a reinitiation scaffold component and the effect of U1 snRNA and the 5' splice site on the efficiency of reinitiation suggests that reinitiation is coordinated by the components of the splicing machinery.

Discussion

In our earlier studies (A.A. and D. Reinberg, unpub. results), we observed that an unidentified factor associated with TFIIH strongly stimulated the rates of TFIIH-dependent abortive initiation in a fully reconstituted system. Under chromatographic separation on a sizing column, this factor dissociated from TFIIH with a significant loss in abortive initiation activity. We were unable to identify the factor, partly because we assumed that it was a protein, not a nucleic acid. Thus, the data presented here satisfactorily explain our earlier findings.

Well-documented features of U1 snRNA provide an interesting insight into possible functional interconnections between transcription initiation and splicing. Interaction of the U1 snRNA with the 5' splice site initiates assembly of U2, U4, U5 and U6 snRNAs and >60 associated proteins into the spliceosome. Recent studies, however, indicate a more flexible order of assembly and a more diverse role of U1 snRNA [19]. For instance, U1 snRNA associates with a distinct but as yet uncharacterized complex apart from the penta-snRNPs [20]. Notably, U1 snRNP can be dispensable for in vitro splicing when Ser-Arg (SR) proteins are present [21]. SR proteins have been reported to interact with the C-terminal domain of the largest subunit of Pol II [22]. All these data suggested an additional function for U1 snRNA. In this context, it is interesting to note that a selection of RNA ligands interfering with early stages of Pol II transcription produced an 11-nucleotide sequence consensus complimentary to U1 snRNA [23].

TFIIH plays a critical role during the initiation of transcription by Pol II by controling the early stages of promoter melting and formation of the open complex in an ATP-dependent manner [24-26]. In reconstituted systems, the levels of productive transcription are significantly reduced compared to the levels of abortive initiated transcription [16]. Introduction of auxiliary factors such as certain activators improves coordination between initiation and productive transcription with a higher fraction of initiated polymerases successfully clearing the promoter and advancing into elongation [16, 27]. On the basis of our results, we suggest that exposure to a promoter proximal 5' splice site via TFIIH–U1 interaction alters the reinitiation scaffold. This could, in turn, make initiated Pol II advance into productive elongation with higher efficiency. Whether a promoter proximal intron may alter the composition of the reinitiation scaffold remains to be explained.

The detailed mechanism of U1–TFIIH interaction and its effect on the enzymatic activities of TFIIH during transcriptional initiation, promoter escape and reinitiation remain to be determined. It will be important to identify, step by step, the interactions between U1 snRNA and other components of the transcription and splicing machinery in the dynamic context of initiation and mRNA processing.

Earlier in vivo data demonstrated strong enhancement of transcription efficiency by promoter proximal introns [10, 11]. Removal of promoter proximal splicing signals from a mammalian gene or the excision of introns from yeast genes results in a marked reduction of nascent transcription. In vivo recognition of the proximal 5' splice site by U1 snRNA increases transcriptional efficiency [28]. Altogether, these results further emphasize the importance of the interaction of U1 snRNA with TFIIH as a molecular mechanism that underlies the coordinated and efficient control of transcriptional initiation and early mRNA processing.

Methods

Chromatographic purification.
All general transcription factors were purified as described [12] from HeLa extracts (Computer Cell Culture Center (C4), Belgium) or were produced as recombinant polypeptides in E. coli BL21 strain. Conventional TFIIH was purified using phosphocellulose, DEAE-52, DEAE-5PW, phenyl-Superose, Mono S and Superose 6 columns on an AKTA Purifier 10 FPLC system (Amersham Bioscience). Immunopurified TFIIH was affinity-purified as described [13]. Gel filtration chromatography of TFIIH was performed in the presence of 300 mM or 1 M NaCl on a Superose 6 HR 10/30 column at a flow rate of 0.5 ml min ^-1.

TFIIH–U1 pull-down experiments.
Purified TFIIH (50 ul) was incubated at 30 °C for 30 min with 2 mM of the biotinylated oligo complimentary to nucleotides 12–75 of U1 snRNA. The mixture was then incubated for 1 h at 4 °C with streptavidin magnetic beads; washed three times in 0.3 M KCl, 0.1 % (v/v) NP-40 and 10 mM Tris-HCl, pH 8.0; and analyzed by western blot.

Transcription assays.
Productive transcription assays from pG5 MLP, a 400-nucleotide G-less template, and abortive initiation assays in the presence of ATP and [³²P]-CTP were monitored as described [14]. T1 RNase treatment was performed post-transcriptionally unless specified. TFIIH was treated with micrococcal nuclease in 1 mM CaCl ₂ for 15 min at 20 °C. The reaction was terminated with 4 mM EGTA. U1 and U2 were synthesized as described [29].

Template constructs for reinitiation assay were based on the pSLG407 construct containing the adenovirus major late (AdML) promoter followed by the 100-nucleotide G-less cassette [30]. The fragment corresponding to nucleotides 110–1,360 of the template was amplified by PCR with the 5' biotinylated primer. Deletion of the core promoter at nucleotides 360–420 generated the 'delta promoter' template. 'WT intron' template contained the b-globin intron sequence [18] inserted downstream from the G-less cassette, 139 nucleotides from the start site. Mutant templates (SD/SA) had the 5' splice site changed from GTTGGT to CTTGTA (mut SD), and the mutant 3' splice site changed from TTAGG to AAACT (mut SA). The template containing the promoter and G-less cassette, but no intron sequences was marked as ' intron'.

pCp labeling, RNA cloning and RNase protection.
[³²P]-pCp labeling and RNase protection were performed as described [2]. RNase protection assays were carried out using a probe containing a sequence complementary to nucleotides 12–75 of U1 snRNA. For cloning RNA associated with TFIH, RNA was prepared from TFIIH after proteinase K digestion and following phenol extraction. A poly(A) tail was added using poly(A) polymerase (Amersham Bioscience), and first and second strands synthesis was carried out using an oligo d(T) primer and a 5' RACE kit (Invitrogen). The resulting PCR product was cloned into the pGEM-T EASY vector and 10 individual clones were sequenced.

Reinitiation assay.
The first round of transcription was conducted in the presence of 5 ug of transcriptionally active nuclear extracts (C4) in transcription buffer (20 mM HEPES, pH 7.9, 100 mM KCl, 3 mM MgCl₂, 0.2 mM EDTA, 0.5 mM dithiothreitol (DTT) and 10% (v/v) glycerol). The reinitiation scaffold formed on the immobilized templates was then washed, and the reinitiation assay was performed by adding TFIIF, TFIIB, TFIIE, NTPs, 32P-CTP and Pol II (DEAE-5PW fraction) as described [6].

Mass spectrometry.
Analysis of TFIIH was performed by nanospray ionization of the samples into a quadrupole time-of-flight mass spectrometer (Q-Tof Micro, Micromass).

UV crosslinking.
Internally ³²P-labeled U1 or anti-sense U1 snRNA were incubated with affinity purified TFIIH pretreated with micrococcal nuclease for 30 min in 10 l of 10 mM Tris-HCl, pH 8.0, 150 mM KCl, 4 mM MgCl₂ and 1 mM EGTA. UV crosslinking was performed at 254 nm wavelength, with exposure pre-set at 120 mJ cm ^-2 for 5 min on ice.

Cleavage by RNase H.
RNase H (1 U) was incubated at 30 °C for 45 min with the transcriptional reinitiation intermediate in the presence of up to 200 ng of the oligo-DNA complementary to nucleotides 12–75 of U1 snRNA or 200 ng of nucleotides 69,214–69,270 of pET28b (Novagene), as control, in transcription buffer (20 mM HEPES, pH 7.9, 100 mM KCl, 3 mM MgCl₂, 0.2 mM EDTA, 0.5 mM DTT and 10% (v/v) glycerol).

Acknowledgments:

We thank A. Catchpole and A. Taylor for help with snRNA cloning, and T. Harder, P. Uguen, P. Cook, O. Bensaude, J. Manley and D. Reinberg for helpful discussions and reagents. A.A. is supported by the Wellcome Trust Career Development Fellowship and by grants from Cancer Research UK and the Edward Abraham Research Fund.

References:

1. Eddy, S.R. Nature Rev. Genet. 2, 919-929 (2001).

2. Nguyen, V.T., Kiss, T., Michels, A.A. & Bensaude, O. Nature 414, 322-325 (2001).

3. Yang, Z., Zhu, Q., Luo, K. & Zhou, Q. Nature 414, 317-322 (2001).

4. Fong, Y.W. & Zhou, Q. Nature 414, 929-933 (2001).

5. Proudfoot, N.J., Furger, A. & Dye, M.J. Cell 108, 501-512 (2002).

6. Yudkovsky, N., Ranish, J.A. & Hahn, S. Nature 408, 225-229 (2000).

7. Feaver, W.J. et al. Cell 79, 1103-1109 (1994).

8. Orphanides, G., Lagrange, T. & Reinberg D. Genes Dev. 10, 2657-2683 (1996).

9. Coin, F. & Egly, J.M. Cold Spring Harb. Symp. Quant. Biol. 63, 105-110 (1998).

10. Brinster, R.L. et al. Proc. Natl. Acad. Sci. USA 85, 836-840 (1988).

11. Palmiter, R.D. et al. Proc. Natl. Acad. Sci. USA 88, 478-482 (1991).

12. Maldonado, E., Drapkin, R. & Reinberg, D. Methods Enzymol. 274, 72-100 (1996).

13. LeRoy, G. et al. J. Biol. Chem. 273, 7134-7140 (1998).

14. Akoulitchev, S., Makela, T.P., Weinberg, R.A. & Reinberg, D. Nature 377, 557-560 (1995).

15. Drapkin, R. et al. Proc. Natl. Acad. Sci. USA 93, 6488-6493 (1996).

16. Kumar, K.P., Akoulitchev, S. & Reinberg, D. Proc. Natl. Acad. Sci. USA 95, 9767-9772 (1998).

17. Moreland, R.J. et al. J. Biol. Chem. 274, 22127-22130 (1999).

18. Efstratiadis, A. et al. Cell 21, 653-668 (1980).

19. Johnson, T.L. & Abelson, J. Genes Dev. 15, 1957-1970 (2001).

20. Stevens, S.W. et al. Mol. Cell 9, 31-44 (2002).

21. Crispino, J.D., Blencowe, B.J. & Sharp, P.A. Science 265, 1866-1869 (1994).

22. Hirose, Y. & Manley, J.L. Genes Dev. 14, 1415-1429 (2000).

23. Tian, H. FEBS Lett. 509, 282-286 (2001).

24. Holstege, F.C., Fiedler, U. & Timmers, H.T. EMBO J. 16, 7468-7480 (1997).

25. Tirode, F., Busso, D., Coin, F. & Egly, J.M. Mol. Cell 3, 87-95 (1999).

26. Kim, T.K., Ebright, R.H. & Reinberg, D. Science 288, 1418-1422 (2000).

27. Liu, J. et al. Cell 104, 6353-6327 (2001).

28. Furger, A., O'Sullivan, J.M. Binnie, A., Lee, B.A. & Proudfoot, N.J. Genes Dev. in the press (2002).

29. Cuello, P. et al. EMBO J. 18, 2867-2877 (1999).

30. Lee, J.M. & Greenleaf, A.L. J. Biol. Chem. 272, 10990-10993 (1997). 

0. Frenster JH, "Nuclear RNA Species Activate DNA Transcription Within Chromatin", FASEB Journal, Vol. 13, No. 7, A1506 (April 23, 1999).

1. Hovsepian JA, and Frenster JH, "RNA-Induced Melting of DNA during Selective Gene Transcription".

2. Frenster JH, and Hovsepian JA, "RNA Feedback Mechanisms during Eukaryotic Gene Regulation".

3. Frenster JH, "Yeast  RNA  Re-Programming  of  Already-Active  Mammalian Chromatin".

4. Frenster JH, "Activation of DNA Transcription within Repressed Chromatin".

5. Frenster JH, "Selective Gene De-Repression by De-Repressor RNA".

6. Frenster JH, "Ultrastructural Probes of Active DNA Sites, and the RNA Activators of DNA".