InteractiveFly: GeneBrief

Negative elongation factor E: Biological Overview | References

Transcriptional elongation regulators NELF and DSIF collaborate to inhibit elongation by RNA polymerase IIa in extracts from human cells. A multifaceted approach was taken to investigate the potential role of these factors in promoter proximal pausing on the hsp70 gene in Drosophila. Immunodepletion of DSIF (FlyBase term: Spt5) from a Drosophila nuclear extract reduces the level of polymerase that pauses in the promoter proximal region of hsp70. Depletion of one Negative elongation factor E (NELF) subunit in salivary glands using RNA interference also reduces the level of paused polymerase. In vivo protein-DNA cross-linking shows that NELF and DSIF associate with the promoter region before heat shock. Immunofluorescence analysis of polytene chromosomes corroborates the cross-linking result and shows that NELF, DSIF, and RNA polymerase IIa colocalize at the hsp70 genes, small heat shock genes, and many other chromosomal locations. Finally, following heat shock induction, DSIF and polymerase but not NELF are strongly recruited to chromosomal puffs harboring the hsp70 genes. It is proposed that NELF and DSIF cause polymerase to pause in the promoter proximal region of hsp70. The transcriptional activator, HSF, might cause NELF to dissociate from the elongation complex. DSIF continues to associate with the elongation complex and could serve a positive role in elongation (Wu, 2003).

It is proposed that promoter proximal pausing occurs when the nascent transcript emerges from the RNA exit channel of the Pol II and is grabbed by the NELF-E subunit. Tethering of the NELF-E to the elongation complex would generate a rigid body that could restrict the movement of the Pol IIa. This model is supported by several observations. The paused polymerase is in the Pol IIa state, and NELF and DSIF only inhibit elongation by Pol IIa. In vitro transcription analysis indicates that the elongation complex is not receptive to inhibition by NELF and DSIF until the nascent transcript is ~30 nucleotides long. This length coincides approximately to the distance polymerase elongates on hsp70 before it pauses. In vitro transcription analyses indicate that DSIF and NELF associate with polymerase shortly after initiation but probably before the polymerase reaches the region of pausing. Finally, NELF-E has an RNA-binding motif that is essential for its inhibitory action (Wu, 2003 and references therein).

Although NELF and DSIF are sufficient to slow the elongation rate of purified Pol IIa, it is suspected that additional proteins are involved in stably pausing Pol II on the hsp70 promoter. In cell-free transcription reactions done with other promoters, the pausing caused by DSIF and NELF appears to be transient -- the polymerase eventually moves forward if given enough time. In contrast, several observations indicate that the Pol II on hsp70 is stably paused. The paused Pol II remains associated with the hsp70 promoter when nuclei are isolated from uninduced cells, and sarkosyl or high salt must accompany addition of nucleotides to cause the Pol II to resume elongation. In a cell-free system, Pol II remains stably paused on the hsp70 promoter for at least 30 min. GAGA factor might be involved in stabilizing the pause because mutations in the GAGA element result in a loss of paused Pol II (Wu, 2003).

Heat shock rapidly induces transcription as a result of the association of HSF with sites located upstream from the TATA element. The data suggest that HSF may activate transcription in part by causing NELF to dissociate from the Pol II. How HSF might cause the release of NELF is unclear. Phosphorylation of Pol IIa is likely to be an important step because the Pol II found in the body of the gene during heat shock is hyperphosphorylated. Phosphorylation of DSIF is another possibility as this has been observed to occur early in elongation in vitro. It is also unclear which kinase might be responsible for phosphorylating the Pol II. P-TEFb (see Cdk9) is a candidate because it associates with the hsp70 gene during heat shock induction, and HSF can be bypassed by directing a Gal4/P-TEFb fusion protein to the hsp70 promoter. No interaction, however, has been detected between P-TEFb and HSF. Recent results show that HSF associates with the mediator. Drosophila mediator contains a kinase that phosphorylates the CTD, and phosphorylation can occur synergistically with the TFIIH kinase. Perhaps HSF recruits the mediator and in turn the mediator releases the paused polymerase by phosphorylating the CTD (Wu, 2003).

The strong immunofluorescence staining observed for DSIF at heat shock loci during heat shock indicates that DSIF is associated with many of the polymerase molecules transcribing the gene. RNA polymerase initiates at a rate of once every few seconds during heat shock resulting in a train of elongation complexes traversing the gene. In the absence of NELF, DSIF might act as a positive elongation factor. Shortly after DSIF was discovered, another investigation identified DSIF as a cofactor required for reconstituting tat-dependent transcription. In this situation, DSIF appears to be stimulating elongation. DSIF has been found in a complex with another positive elongation factor called Tat-SF1. Tat-SF1 was first identified as a stimulatory factor for Tat, but subsequent results indicate that Tat-SF1 may promote elongation on cellular genes. In yeast, DSIF appears to act as either a positive or negative regulator of elongation depending on circumstances. A hypothesis that unites the positive and negative activities of DSIF considers this factor an adaptor that connects other modulators to the elongation complex. In this regard, DSIF has been shown to bind on its own to Pol II, whereas the stable association of NELF with Pol II requires the presence of DSIF (Wu, 2003 and references therein).

NELF and DSIF appear to associate with several hundred interbands in polytene chromosomes. Each interband could contain many genes. The weak staining of interbands by Hoecsht suggests that the DNA in the interbands is in a decondensed state. Residing in these decondensed regions could be genes whose primary control mechanism does not involve a disruption of chromatin structure or even assembly of the initiation complex. Instead, alleviating repression by NELF and DSIF could underlie the mechanism of activation (Wu, 2003).

RNA polymerase II initiates transcription in slp1-repressed cells and pauses downstream from the promoter in a complex that includes the negative elongation factor NELF

The simple combinatorial rules for regulation of the sloppy-paired-1 (slp1) gene by the pair-rule transcription factors during early Drosophila embryogenesis offer a unique opportunity to investigate the molecular mechanisms of developmentally regulated transcription repression. Initial repression of slp1 in response to Runt and Fushi-tarazu (Ftz) does not involve chromatin remodeling, or histone modification. Chromatin immunoprecipitation and in vivo footprinting experiments indicate RNA polymerase II (Pol II) initiates transcription in slp1-repressed cells and pauses downstream from the promoter in a complex that includes the negative elongation factor NELF. The finding that Negative elongation factor E also associates with the promoter regions of wingless (wg) and engrailed (en), two other pivotal targets of the pair-rule transcription factors, strongly suggests that developmentally regulated transcriptional elongation is central to the process of cell fate specification during this critical stage of embryonic development (Wang, 2007).

DNase I hypersensitivity was used to probe the chromatin structure of the slp1 locus. These assays revealed the presence of a DNase I-hypersensitive site near to the 5'-end of the slp1 transcription unit. Chromatin immunoprecipitation (ChIP) experiments with antiserum against histone H3 provide an explanation for this DNase I hypersensitivity. There is significantly reduced association of H3 with the slp1 promoter region compared with both the structural gene as well as sequences upstream of the promoter. These observations strongly suggest that the promoter region is nucleosome free. Importantly, matched collections of wild-type and Runt + Ftz (R+F) embryos show both the same pattern of DNase I hypersensitivity and histone H3 association. These results indicate that the 40-fold decrease in mRNA expression in slp1-repressed embryos is not due to gross changes in the accessibility of the slp1 promoter region (Wang, 2007).

Histone acetylation and deacetylation are important for transcriptional regulation with a general correlation between histone acetylation and active transcription. Indeed, prior work has demonstrated that the Rpd3 histone deacetylase is important for maintaining the Runt-dependent repression of the segment-polarity gene en. ChIP experiments reveal no significant difference in the H3 acetylation pattern of slp1 chromatin from wild-type versus R+F embryos. Although no differences were detected in H3 or Ac-H3 association that correlate with slp1 repression, there are interesting differences in the H3 acetylation levels at different genomic locations. The slp1 structural gene shows stronger Ac-H3 association than the upstream region. This difference is not observed for the association of H3 with these same intervals, suggesting that H3 acetylation marks genomic regions that are permissive for transcription. The relative levels of H3 and Ac-H3 association with Brother (Bro), a gene that is not transcribed in the early embryo (as measured by RT-PCR), provide additional evidence for this trend. Although H3 association with the Bro gene is greater than for any region of the slp1 locus, the level of Ac-H3 association with Bro is lower than for any region of slp1. Based on the observation that differences in H3 and Ac-H3 association can be detected that correlate with transcriptional potential and yet no differences was detected between wild-type and slp1-repressed embryos, it is concluded that H3 acetylation plays a negligible role in the establishment of slp1 repression (Wang, 2007).

The above observations led to a characterization the interactions of the transcriptional machinery with slp1. Association of the TATA-box-binding protein (TBP) is a first step in assembly of the transcriptional machinery on a promoter. As expected, TBP association is detected with a promoter-proximal interval centered 6 base pairs (bp) upstream of the slp1 transcript initiation site in chromatin from wild-type embryos. A weaker signal is detected for an interval within the 5' untranslated region (UTR), centered 124 bp downstream from the start site, whereas all other intervals give background level signals. Very similar levels of TBP association were found in chromatin from R+F embryos. More surprising is the finding that there is almost no difference in the level of Pol II association with the slp1 promoter-proximal interval in chromatin from wild-type and R+F embryos. Pol II is also associated with the slp1 structural gene in wild-type embryos, but at lower levels than at the promoter. In contrast, Pol II association with the slp1 structural gene is markedly reduced in R+F embryos and near to background levels for regions downstream from the 5'-UTR. Based on these results, it is concluded that promoter recruitment of Pol II is not blocked in slp1-repressed embryos. slp1-associated Pol II was further characterized using an antibody that recognizes the Phospho-Ser-5 form of the heptad repeats that comprise the C-terminal domain (CTD) of the largest Pol II subunit. Phospho-Ser-5 modification of the CTD is associated with transcription initiation. This antiserum also gives the strongest signals with the slp1 promoter-proximal interval in wild-type chromatin, and this signal is not reduced in chromatin from R+F embryos. This result indicates that slp1 repression occurs at a step downstream from transcription initiation (Wang, 2007).

The Drosophila hsp70a promoter is an extensively studied example of regulated transcriptional elongation. Pol II initiates transcription at the hsp70a promoter, and then, in the absence of a heat shock, pauses immediately downstream from the promoter. All somatic cells in 3-4-h-AED embryos are capable of activating the hsp70a gene, and as expected, Phospho-Ser-5-modified Pol II is readily detected on the hsp70a promoter in chromatin preparations from non-heat-shocked embryos. The paused Pol II complex on the hsp70a promoter is also readily detected using permanganate footprinting due to the increased sensitivity of thymine residues in single-stranded regions. This same technique was used to carry out footprinting studies on the slp1 promoter region. The results reveal strong hyperreactivity of thymine residues at +15, +28, +30, +38, and +50 downstream from the transcription start site in blastoderm. This interval is similar, though perhaps somewhat larger than the interval detected for hsp70a, within which the most prominent increases in reactivity are at residues +22 and +30. The pattern of reactivity on slp1 is extremely similar in both wild-type and slp1-repressed embryos, indicating that the hyperreactivity is not due to active transcription of the slp1 gene. Importantly, this pattern is not observed in nuclei from Drosophila tissue culture cells. Thus, unlike hsp70a, the footprint on the slp1 5'-UTR is developmentally regulated (Wang, 2007).

The negative elongation factor NELF is thought to play a key role in establishing the paused Pol II complex on the hsp70a promoter. Indeed, NELF association provides a marker for the paused complex as heat-shock-induced transcriptional elongation involves release of NELF (Wu, 2005). In agreement with the results of footprinting studies, ChIP experiments reveal the NELF-D and NELF-E subunits are associated with the slp1 promoter region in chromatin from wild-type embryos, but not in chromatin from Drosophila tissue culture cells. Strong signals are obtained in chromatin from embryos with both the promoter-proximal and 5'-UTR intervals, whereas background level signals are obtained with other intervals of the slp1 locus. It is notable that the promoter-proximal signal is less than or equal to the signal detected for the 5'-UTR interval. This pattern of association contrasts that obtained with TBP, which shows a threefold stronger signal with the promoter-proximal primer pair. These association patterns suggest is that NELF is bound downstream from the slp1 transcription start site, presumably as a component of the paused Pol II complex. Consistent with this interpretation, a similar differential pattern was found of TBP and NELF association with promoter-proximal and 5'-UTR intervals of hsp70a. These results strongly suggest that NELF plays a key role in regulating slp1 elongation in the blastoderm-stage Drosophila embryo (Wang, 2007).

The initial indications that slp1 expression was regulated at a step downstream from transcription initiation came from ChIP experiments on chromatin from a homogeneous population of embryos that uniformly repress slp1. Localized association of NELF in a region downstream from the transcription start site is a hallmark of promoter-proximal pausing. Importantly, this association provides a method for detecting paused Pol II complexes in chromatin from embryos that contain a mixture of cells, some of which are expressing full-length mRNA transcripts. ChIP assays were used to determine whether NELF associates with the promoter regions of wg and en, two pivotal segment-polarity gene targets of the pair-rule transcription factors. The results reveal specific association of NELF with the promoter-proximal and 5'-UTR regions of both genes in 3-4-h-AED embryos. Furthermore, the differential association pattern of TBP and NELF with these two intervals indicates that NELF is localized to a region immediately downstream from the initiation sites for both genes. These findings indicate that regulation of transcriptional elongation is likely to be central in generating the initial patterns of segment-polarity gene expression in the Drosophila embryo (Wang, 2007).

Regulation of transcriptional elongation has been described for several genes in addition to the Drosophila heat-shock genes, including human c-myc, c-myb, c-fos, junB, and p21. A feature shared by these previously characterized examples is rapid induction of gene expression in response to external stimuli. The initial establishment of segment-polarity gene-expression patterns in response to the pair-rule transcription factors occurs within a relatively brief developmental window of ~30 min, spanning the completion of cellularization and the beginning of germ band extension. The temporal advantages offered by regulating these genes at a transcriptional elongation step as compared with chromatin remodeling and/or Pol II initiation complex assembly may be essential for the timely establishment of differing gene expression programs during cell fate specification in the Drosophila blastoderm embryo. The observations that Pol II molecules are enriched at the 5'-ends of a number of genes, coupled with findings that defects in transcriptional elongation factors produce specific developmental defects, strongly suggest that regulation of transcriptional elongation is a hitherto overlooked, but potentially widespread strategy for controlling gene expression during development (Wang, 2007).

Molecular characterization of Drosophila NELF

NELF and DSIF act together to inhibit transcription elongation in vitro, and are implicated in causing promoter proximal pausing on the hsp70 gene in Drosophila. Drosophila NELF has four subunits similar to subunits of human NELF. The amino acid sequences of NELF-B and NELF-D are highly conserved throughout their lengths, while NELF-A and NELF-E contain nonconserved regions inserted between conserved N- and C-terminal regions. Immunodepletion of NELF or DSIF from a nuclear extract desensitizes transcription in vitro to the nucleoside analog DRB. Immunodepletion of NELF also impairs promoter proximal pausing on the hsp70 promoter in vitro without affecting initiation. Chromatin immunoprecipitation analyses detect NELF at the promoters of the hsp70 and ß1-tubulin genes where promoter proximal pausing has been previously detected. Heat shock induction of hsp70 results in a marked decrease in NELF at the hsp70 promoter. Immunofluorescence analysis of polytene chromosomes shows extensive colocalization of the NELF-B and NELF-D subunits at hundreds of interbands. Neither subunit appears to be recruited to puffs. These results provide a foundation for genetic and biochemical analysis of NELF in Drosophila (Wu, 2005; full text of article).

The amino acid sequences of the various NELF subunits were aligned to learn more about NELF. The sequences of NELF-B and NELF-D orthologs exhibit several regions with >50% identity distributed over their entire lengths. Analysis of subunit interactions has lead to a proposal that NELF-B and NELF-D form a central core that brings together NELF-A and NELF-E. NELF-A associates with NELF-D while NELF-E associates with NELF-B. The high degree of homology between NELF-B and NELF-D orthologs is consistent with constraints that might be dictated by the requirements that these two proteins interact with each other and with other subunits of NELF. The extensive colocalization of NELF-B and NELF-D on polytene chromosomes is also consistent with these two subunits forming the core of the NELF complex (Wu, 2005).

When NELF-A was initially identified through a BLAST search, it was of concern that the genome annotation because dNELF-A was over twice the size of human NELF-A. Immunoblot analysis verifies that dNELF-A is approximately twice the size of hNELF-A. The alignment shows that the size difference is due to a large nonconserved region located approximately between amino acids 300 and 1100. Inspection of the amino acid sequence in this region of Drosophila NELF-A reveals numerous tracts of poly-glutamine, poly-asparagine, poly-threonine and poly-alanine that are absent from hNELF-A. This lack of conservation raises the possibility that this region serves as a linker between two distinct functional domains defined by the regions of homology spanning the first 250 amino acids and the last 100 amino acids of NELF-A. The N-terminal region of homology in hNELF-A has been shown to interact with Pol II and NELF-D. Moreover, this region is required for NELF-mediated repression in vitro. Deletion of amino acids 431-528 of hNELF-A, which corresponds to the last 100 amino acids of dNELF-A, does not impair the ability of the NELF complex to inhibit transcription in vitro. Future analysis using genetic and molecular genetic approaches in Drosophila may help elucidate the function of the large nonconserved region and the C-terminal domain (Wu, 2005 and references therein).

hNELF-E binds RNA, and the region responsible for this activity exhibits 30% identity to amino acids 167-232 in dNELF-E. This RNA recognition motif has been shown to be required for NELF inhibition in human extracts. Two features found in hNELF-E are notably absent in dNELF-E. A 46 amino acid stretch spanning from 196 to 242 in hNELF-E that is composed primarily of alternating arginine and aspartic acid residues is absent -- the function of this region in hNELF-E is not known. Also absent from dNELF-E are serines that correspond to amino acids 181, 185, 187 and 191 in hNELF-E. These serines in hNELF-E are phosphorylated in vitro by P-TEFb, and these modifications reduce the binding between hNELF-E and RNA. It will be interesting to learn if phosphorylation of dNELF-E regulates its RNA binding activity (Wu, 2005 and references therein).

Two results implicate NELF in promoter proximal pausing on the hsp70 heat shock gene. Depleting salivary glands of NELF using RNA interference diminishes the level of promoter proximal pausing occurring on hsp70 in salivary glands. Also, NELF was found to cross-link to the hsp70 promoter region prior to heat shock induction. This study provides additional data strengthening the conclusion that NELF is involved in promoter proximal pausing (Wu, 2005).

Immunodepletion of NELF from nuclear extracts impairs formation of the paused Pol II. Importantly, cell-free transcription reactions with 3' O-methyl GTP indicate that NELF does not contribute to steps in the transcription process encompassing initiation and elongation to +15. It has also been determined that NELF is associated with the ß1-tubulin promoter. Genomic footprinting with permanganate and nuclear run-on analysis indicate that promoter proximal pausing occurs on the ß1-tubulin gene (Wu, 2005).

Based on strong immunofluorescence staining detected for DSIF and the weak staining for NELF at heat shock puffs, it is proposed that NELF dissociates from elongation complexes during activation but DSIF remains associated with the elongation complexes. The observation that the level of NELF cross-linking to the hsp70 promoter region after heat shock is ~5-fold less than before heat shock is consistent with this model. However, a 2-fold decrease in the level DSIF cross-linking to the hsp70 promoter region was also observed after heat shock, in agreement with a report from Saunders (2003). The decrease in DSIF cross-linking near the promoter was unexpected since heat shock puffs on polytene chromosomes stain very strongly with DSIF antibody. However, much of the signal detected on the polytene chromosomes probably originates from DSIF associated with elongation complexes in the body of the gene. In addition, the interaction of DSIF with Pol II in the promoter proximal region might be transient or not occur at all on some elongation complexes in the promoter proximal region. The rate of reinitiation at the hsp70 promoter is exceptionally high, once every few seconds, so a significant portion of the Pol II molecules might escape into the body of the gene before associating with DSIF (Wu, 2005).

The broad distribution of NELF-B and NELF-D on polytene chromosomes could mean that hundreds of genes harbor paused Pol II similar to that found on hsp70. Alternatively, NELF could have additional functions. mRNA capping is coupled to transcription, and recent results lead to the hypothesis that pausing induced by NELF might be important for coupling transcription and mRNA capping (Mandal, 2004). A subset of genes like hsp70 and ß1-tubulin might use cofactors such as GAGA factor to generate a stably paused state (Wu, 2005).

NELF-mediated stalling of Pol II can enhance gene expression by blocking promoter-proximal nucleosome assembly

The Negative Elongation Factor (NELF) is a transcription regulatory complex that induces stalling of RNA polymerase II (Pol II) during early transcription elongation and represses expression of several genes studied to date, including Drosophila Hsp70, mammalian proto-oncogene junB, and HIV RNA. To determine the full spectrum of NELF target genes in Drosophila, a microarray analysis was performed of S2 cells depleted of NELF; NELF RNAi affects many rapidly inducible genes involved in cellular responses to stimuli. Surprisingly, only one-third of NELF target genes are, like Hsp70, up-regulated by NELF-depletion, whereas the majority of target genes showed decreased expression levels upon NELF RNAi. These data reveal that the presence of stalled Pol II at this latter group of genes enhances gene expression by maintaining a permissive chromatin architecture around the promoter-proximal region, and that loss of Pol II stalling at these promoters is accompanied by a significant increase in nucleosome occupancy and a decrease in histone H3 Lys 4 trimethylation. These findings identify a novel, positive role for stalled Pol II in regulating gene expression and suggest that there is a dynamic interplay between stalled Pol II and chromatin structure (Gilchrist, 2008).

These results establish that the NELF complex regulates transcription of a significant number of genes in Drosophila S2 cells (241 transcripts. As exemplified by Hsp70, the only previously defined target of Drosophila NELF, promoter-proximally stalled Pol II is a common feature of these genes. However, in contrast to Hsp70, where NELF-mediated Pol II stalling represses transcription, this study found that nearly 70% of NELF target genes are down-regulated upon NELF RNAi. NELF is critical for maintaining Pol II stalling at these down-regulated genes, and stalled Pol II helps preserve a nucleosome-free region around these promoters. Although this work is the first to investigate the mechanisms of NELF's positive effects on gene expression, it is noted that previous microarray analyses of NELF-depleted human cells also detected significant numbers of transcripts that were down-regulated upon NELF depletion (Aiyar, 2007; Narita, 2007; Gilchrist, 2008 and references therein).

The current data also identify many genes in which depletion of NELF produces no change in transcriptional output but does lead to a detectable reduction in Pol II promoter occupancy, in agreement with previous results (Muse, 2007). It is suggested that at these genes, early elongation is not rate-limiting for transcription under the experimental conditions that were used. However, it is expected that NELF could influence transcriptional output of some of these genes under other conditions, in other cell types, or at different developmental stages (Gilchrist, 2008).

Importantly, many NELF target genes are involved in responses to stimuli. This finding is consistent with a recent genome-wide search for genes with promoter-proximally stalled Pol II, which also found a significant enrichment in stimulus-responsive genes among the ~1000 Drosophila genes that were determined to possess stalled polymerase (Muse, 2007). It is proposed that inducible genes possess a poised Pol II and open chromatin architecture in order to facilitate rapid gene activation in response to environmental signals (Gilchrist, 2008).

A key result of this study is that NELF-mediated stalling of polymerase can have both positive and negative effects on gene expression. Pol II stalling can function either to attenuate the expression of genes like Hsp70 under noninducing conditions or, alternatively, to maintain basal expression levels by preventing transcription inactivation and the assembly of promoter-proximal nucleosomes. Based on the results and on the previously defined role of NELF in inhibiting transcription elongation, the following model is proposed for NELF activity: NELF binds to the Pol II complex and collaborates with DSIF and/or other factors to induce promoter-proximal stalling. Depletion of NELF reduces the duration of Pol II stalling, allowing the polymerase to resume elongation and move away from the promoter region (Gilchrist, 2008).

At up-regulated genes such as Hsp70, NELF-mediated Pol II stalling represses expression and NELF RNAi allows Pol II to be released into the gene more rapidly after initiation, resulting in an increase in transcriptional output. Significant Pol II ChIP signal remains at up-regulated promoters upon NELF RNAi (Muse, 2007), and permanganate footprinting of Pol II complexes shows similar levels of reactivity within the initially transcribed regions of these genes, demonstrating that the released Pol II is efficiently replaced by newly recruited polymerase molecules. Furthermore, permanganate reactivity was observed at sites slightly further downstream from the promoter of the up-regulated genes in both NELF-depleted S2 cells and flies, reflecting Pol II elongation into these genes. Interestingly, these results suggest that the depletion of NELF reduces, but does not completely abolish, Pol II stalling near these promoters and that inhibition of early elongation by NELF is not the only mechanism that affects promoter-proximal transcription efficiency at up-regulated genes (Gilchrist, 2008).

Consistent with the continued presence of promoter-proximal Pol II at up-regulated genes following NELF RNAi, no increase is detected in nucleosome occupancy of these promoters or a change in local histone H3-K4-me3 levels. In addition, no evidence was found that the accessibility of the upstream region of the up-regulated genes is altered by NELF RNAi; it is concluded that the chromatin architecture at these genes is not substantially affected by a reduction in Pol II stalling duration (Gilchrist, 2008).

In contrast, at down-regulated genes, the Pol II released upon NELF RNAi is not replaced by recruitment of additional Pol II. A striking reduction was observed in promoter-proximal Pol II occupancy and stalling at these promoters. Moreover, this decrease in polymerase occupancy is accompanied by an increase in nucleosome density and nuclease protection near the down-regulated promoters and a loss of the active chromatin mark H3-K4-me3. Taken together, these data suggest that the presence of a promoter-proximally stalled Pol II complex might positively influence transcription of these genes by preventing nucleosome assembly near their promoters. It remains to be determined if the stalled polymerase achieves this task by sterically blocking the promoter region, thereby protecting it from encroaching nucleosomes, or if the stalled Pol II complex recruits or stabilizes other protein complexes near these promoters. In either case, it is envisioned that the resulting permissive chromatin structure would enhance subsequent transcription by facilitating recruitment of the general transcription machinery and additional Pol II molecules (Gilchrist, 2008).

Previous work has shown that mutations within the Hsp70 promoter sequence that reduced Pol II stalling also decreased the accessibility of the upstream DNA region to the heat-shock activator HSF, and leads to a slower, less efficient heat-shock response. These results indicated that the stalled Pol II plays a critical role in establishing the normal, open promoter architecture at Hsp70 that is required for both activator binding and Pol II recruitment during heat induction. It is proposed that preventing promoter-proximal nucleosome assembly is a general role for stalled Pol II. In this way, it is envisioned that Pol II stalling can play a positive role in gene expression by 'poising' the promoter for further or future activation (Gilchrist, 2008).

The idea that there is interplay between the stalled polymerase and the +1 nucleosome has recently received support from a whole-genome analysis of nucleosome positioning in Drosophila (Mavrich, 2008). These data revealed the presence of a highly positioned +1 nucleosome located in proximity to the promoter, in a position to influence Pol II stalling. The +1 nucleosome was shifted slightly downstream at genes with stalled Pol II, suggesting that Pol II engages this nucleosome and influences its location. These data are consistent with this possibility and, furthermore, suggest that stalled Pol II can affect the occupancy as well as the positioning of promoter-proximal nucleosomes (Gilchrist, 2008).

It is noted that although an increased signal is detected for histone transcripts in the microarray analysis, this arises from aberrant polyadenylated histone transcripts produced in the absence of NELF rather than an actual increase in transcription of histone genes. These results are consistent with the recently demonstrated role of human NELF in stimulating histone mRNA processing and indicate that this function is conserved in metazoans (Narita, 2007). Importantly, this change in histone mRNA processing does not lead to observable changes in histone protein levels (Gilchrist, 2008).

While it is currently unknown what dictates whether a given gene would be repressed, stimulated, or unaffected by NELF-mediated Pol II stalling, this is likely to involve interactions among the stalled polymerase, transcription activators like P-TEFb, chromatin-modifying complexes, and general transcription factors. It is envisioned that differences in the inherent rate of Pol II recruitment and initiation at individual genes would affect for how long the promoter was left unoccupied following release of stalled Pol II. Under conditions in which the Pol II recruitment rate is fast, the gene might maintain high Pol II occupancy and a nucleosome-free structure even in the absence of stalling. In contrast, a gene with a slow rate of recruitment might be more susceptible to encroaching nucleosomes when stalled Pol II is released artificially, such as through depletion of NELF. Likewise, the duration of Pol II stalling at a given gene, and how this compared with the rate of Pol II recruitment, would play an important role in determining the transcription outcome in response to NELF depletion (Gilchrist, 2008).

Both this work and previous ChIP-chip analyses demonstrated that promoters with stalled Pol II exhibit variable responses to NELF RNAi. Whereas the majority of promoters with stalled Pol II on partial genome arrays exhibited dramatic decreases in Pol II levels upon NELF depletion, approximately one-third of these genes retained promoter-proximal enrichment of Pol II in the absence of NELF (Muse, 2007). These data suggested that while most genes require NELF to establish stalled Pol II, a subset of genes possess mechanisms for recruiting and retaining Pol II within the promoter region that are independent of NELF. Many up-regulated genes, including Hsp70, fall into this latter class. Based on the wealth of data describing promoter-proximal Pol II stalling at Hsp70, it is suggested that aspects of the initially transcribed sequence and the binding of GAGA factor upstream would help to establish stalling at these genes (Gilchrist, 2008).

To address the possibility that inherent sequence properties of the promoter may play a role in determining the role of NELF-mediated stalling, promoter-proximal sequences of up-regulated and down-regulated genes (from -250 to +50 bp) were analyzed. It was found that up-regulated promoters are significantly more likely to contain TATA elements than are down-regulated genes or Drosophila promoters genome-wide. This result is intriguing, since TATA-containing promoters as a class have been shown to exhibit a less canonical nucleosome organization and lack a well-positioned +1 nucleosome centered near +150. Instead, these genes appear to establish gene-specific chromatin structures with the help of chromatin remodeling complexes (Albert, 2007; Mavrich, 2008). Thus, the TATA-containing, up-regulated genes might possess redundant mechanisms to maintain the appropriate chromatin structure. This is clearly the case at heat-shock genes, where GAGA factor, TFIID and the stalled polymerase have all been shown to contribute to maintaining a nucleosome-depleted upstream region and positioned +1 nucleosome located well downstream of +200. Although other up-regulated genes have not been studied as extensively, MNase footprinting performed on the promoter regions of mfas, oaf, and rho showed no evidence of a well-positioned +1 nucleosome positioned upstream of +200, suggesting that they too lack a canonical nucleosome distribution. It is proposed that these up-regulated genes, like Hsp70, might use chromatin remodeling complexes to help maintain their gene-specific nucleosome structure, making them less reliant on NELF-mediated Pol II stalling for maintenance of promoter accessibility (Gilchrist, 2008).

In contrast, the down-regulated promoters investigated have tightly positioned nucleosomes that are located close to the transcription start site and a canonical nucleosome pattern, typical of TATA-less genes. A positioned +1 nucleosome is found centered at approximately +150 bp downstream from the TepII, Tl, and nocturnin promoters, as well as defined nucleosome-free promoter and upstream regions. It is suggested that, in the absence of specific chromatin remodeling activities, the down-regulated genes depend on promoter-proximally stalled Pol II as a place-holder in order to maintain a nucleosome free promoter region and potentiate gene expression (Gilchrist, 2008).

In summary, this study has found that NELF-mediated Pol II stalling is not a strictly repressive process, but, unexpectedly, can also function to enhance transcription. The fact that many NELF target genes are involved in responses to stimuli suggests that Pol II stalling plays a role in dynamic, signal-dependent activation of transcription. Consistent with this idea, it was found that the stalled Pol II maintains a chromatin environment around NELF target genes that could facilitate rapid recruitment of additional Pol II to these genes. It will be interesting in future studies to determine in detail how the stalled Pol II impacts local nucleosome architecture and promoter accessibility (Gilchrist, 2008).

Stalled Hox promoters as chromosomal boundaries

Many developmental control genes contain stalled RNA Polymerase II (Pol II) in the early Drosophila embryo, including four of the eight Hox genes. Evidence is presented that the stalled Hox promoters possess an intrinsic insulator activity. The enhancer-blocking activities of these promoters are dependent on general transcription factors that inhibit Pol II elongation, including components of the DSIF (Spt4, and Spt5) and NELF complexes. The activities of conventional insulators are also impaired in embryos containing reduced levels of DSIF and NELF. Thus, promoter-proximal stalling factors might help promote insulator-promoter interactions. It is proposed that stalled promoters help organize gene complexes within chromosomal loop domains (Chopra, 2009b).

Hox genes are responsible for the anterior-posterior patterning of most metazoan embryos. They are typically organized in gene complexes containing a series of cis-regulatory DNAs, including enhancers, silencers, and insulator DNAs . In Drosophila, the eight Hox genes are contained within two gene complexes: the Antennapedia complex (ANT-C), which controls the patterning of anterior regions, and the Bithorax complex (BX-C), which controls posterior regions. The proper spatiotemporal transcription of Hox genes is achieved by the coordinated action of linked cis-regulatory DNAs that are organized in a colinear fashion across the ANT-C and BX-C complexes (Chopra, 2009b).

Chromosomal boundary elements, or insulators, are essential for the orderly regulation of Hox gene expression. They are thought to ensure proper cis-regulatory 'trafficking,' whereby the correct enhancers interact with the appropriate target promoters. Insulators might also help control the levels of transcription by attenuating enhancer-promoter interactions. Insulators are sometimes associated with promoter targeting sequences (PTS), which can facilitate enhancer-promoter interactions by modulating the activities of neighboring insulators (Chopra, 2009b).

Recently, long-range cis-regulatory interactions have been mapped in Drosophila Hox complexes using the DamID technique, chromosomal conformation capture (3C) assays, and transgenic approaches. These studies suggest that the Fab7 and Fab8 insulators are associated with the Abd-B promoter under repressed conditions, even though they map >30-50 kb downstream from the promoter. These long-range interactions depend on the CTCF boundary-binding protein, thereby raising the possibility that insulators interact with one another and organize Abd-B cis-regulatory DNAs within chromosomal loop domains. Similarly, the prototypic insulators flanking the heat-shock puff locus, scs and scs', have also been shown to interact with one another. Additional insulator-insulator loops have also been documented. These loops are thought to facilitate the interactions of remote enhancers and silencers with appropriate target promoters. This study presents evidence that Hox promoters with stalled RNA Polymerase II (Pol II) possess an intrinsic insulator activity, which might help foster the formation of insulator-promoter chromosomal loop domains (Chopra, 2009b).

Four of the eight Hox genes contained in the ANT-C and BX-C contain stalled Pol II. Interestingly, all four stalled genes map at the boundaries of the two Hox complexes. In contrast, internal Hox genes (pb, Dfd, and Scr within the ANT-C, and abd-A within the BX-C) lack stalled Pol II. This arrangement of stalled Hox genes raises the possibility that stalling contributes to the chromosomal organization of Hox complexes. All four stalled Hox genes (lab, Antp, Ubx, and Abd-B) were tested for enhancer-blocking activity in transgenic embryos, along with the promoter regions of two nonstalled genes (Scr and abd-A). Test promoters were placed 5' of lacZ and inserted between a divergent white reporter gene and 3' iab-5 enhancer (IAB5) (Chopra, 2009b).

IAB5 regulates Abd-B expression in posterior regions of the early embryo, corresponding to the primordia for parasegments 10-14. IAB5 is a robust enhancer, and can activate lacZ and white even when positioned far from the reporter genes. This assay was used to reveal an intrinsic enhancer-blocking activity of the eve promoter region. eve/lacZ fusion genes block the ability of IAB5 to activate a distal CAT reporter gene. However, mutagenized eve promoter sequences lacking a critical proximal GAGA element failed to block IAB5-white interactions. Similarly, the Abd-B proximal promoter (Abd-Bm) and Ubx promoter regions block activation of distal white expression, whereas the abd-A promoter does not interfere with the activation of white expression in the presumptive abdomen by the IAB5 enhancer (Chopra, 2009b).

These results suggest that the stalled Abd-B proximal promoter and Ubx promoters possess an enhancer-blocking activity, whereas abd-A does not. A similar trend was observed for Hox promoter sequences from the ANT-C. The Antp and lab promoters block IAB5-white interactions, whereas the Scr promoter (which lacks stalled Pol II) does not interfere with the activation of white expression in the presumptive abdomen. Stalled genes from the tinman complex (Tin-C), which encode NK homeobox proteins responsible for patterning mesodermal lineages, were also examined. All of the stalled promoters from the Tin-C contain insulator activities. In contrast, nonstalled promoters from lbl and C15 lack such activities when tested in similar transgenic assays. Even the Hsp70 promoter, the classic example of Pol II pausing, displayed insulator activity when tested in similar enhancer-blocking transgenic assays (Chopra, 2009b).

The preceding experiments suggest that stalled Hox gene promoters contain enhancer-blocking activities. However, an alternative possibility is that stalled promoters are 'stronger' than the white promoter, and are able to sequester the shared IAB5 enhancer. To distinguish between competition and insulator activities, the IAB5 enhancer was placed between the divergently transcribed white and lacZ reporter genes. When the white promoter sequence was placed 5' of the lacZ reporter gene, the shared IAB5 enhancer worked equally well to activate both white and lacZ expression. Similar results were obtained when the leftward lacZ reporter gene was placed under the control of either the stalled Abd-B or Ubx promoters. In all of these cases, both white and lacZ are expressed equally well in the presumptive abdomen. These results suggest that stalled promoters do not block enhancer-promoter interactions by a competition mechanism. Rather, they work like insulators and block such interactions only when positioned between the distal enhancer and target promoter (Chopra, 2009b).

To determine whether stalled Pol II is important for the enhancer-blocking activities of Ubx and Abd-B, mutant embryos were examined with reduced levels of critical Pol II elongation factors. Ubx and Abd-B were selected for further studies since optimal expression of both genes depends on the Pol II elongation factors Cdk9 (pTEFb) and Elo-A (Chopra, 2009a). It was reasoned that destabilization of stalled Pol II might reduce the enhancer-blocking activities of the Ubx and Abd-B promoter regions. However, reductions in Cdk9 and Elo-A are expected to stabilize, not destabilize, Pol II stalling since both are positive factors that promote elongation (Saunders, 2006). Indeed, reductions in Cdk9 or Elo-A activity do not alter the enhancer-blocking activities of the Ubx and Abd-B promoters (Chopra, 2009b).

To investigate the link between Pol II stalling and enhancer blocking, two negative elongation factors were examined: NELF (Lee, 2008) and DSIF (Wada, 1998; Yamaguchi, 1998; Kaplan, 2000). The NELF-E protein binds to the short nascent transcripts protruding from the active site of Pol II after transcription initiation and promoter clearance, and thereby inhibits Pol II elongation (Wu, 2005; Lee, 2008). Both NELF and DSIF are thought to help stabilize Pol II at the pause site, typically 20-50 base pairs (bp) downstream from the transcription start site (Saunders, 2006; Gilchrist, 2008; Lee et al. 2008). Since Pol II elongation factors are encoded by essential genes, it is not possible to examine the lacZ/white reporter genes in homozygous mutant embryos. Instead, the transgenes were expressed in embryos derived from heterozygous females, and thereby contain half the normal levels of NELF and DSIF (Spt) subunits. Reductions in Nelf-E, Nelf-A, Spt4, and Spt5 cause clear disruptions in the enhancer-blocking activities of both the Ubx and Abd-B promoters, as seen by the strong activation of the distal white reporter gene. In contrast, white expression is blocked when the same transgenes are expressed in a wild-type background. The simplest interpretation of these results is that reduced levels of the NELF and DSIF inhibitory complexes destabilize stalled Pol II at the pause site. Reduced pausing results in diminished enhancer-blocking activities. There is a similar loss in the enhancer-blocking activities of the eve promoter and Fab7 insulator when the transgenes are expressed in embryos containing reduced levels of the GAGA factor, Trl. It is conceivable that the GAGA factor also contributes to the enhancer-blocking activity of the Ubx promoter since Trl/+ embryos display augmented expression of white (Chopra, 2009b).

In principle, the augmented expression of the white reporter gene might not result from the impaired function of the stalled insulators, but might arise from enhanced activity of the white promoter. To investigate this issue, Pol II chromatin immunoprecipitation (ChIP) assays were performed, coupled with quantitative PCR (qPCR) assays. In DSIF and NELF mutant embryos, there is no increase in Pol II levels at either the white promoter or intronic regions as compared with wild-type embryos. These results suggest that augmented expression of white is due to diminished insulator activities of stalled promoters in embryos containing reduced levels of negative Pol II elongation factors (Chopra, 2009b).

It has been suggested that insulators might work, at least in part, via promoter mimicry. To explore this issue, the impact of reductions in NELF and DSIF on the activities of two known insulators, Fab7 and Fab8, from the BX-C, were examined. Previously published transgenic lines were used that contain Fab7 or Fab8 inserted between the IAB5 and 2XPE (twist) enhancers attached to a leftward lacZ reporter gene and rightward white reporter. In wild-type embryos, the reporter genes are activated only by the proximal enhancer. Thus, white is activated solely in the mesoderm by the 2XPE enhancer, while lacZ is activated in the presumptive abdomen by IAB5. The distal enhancers are blocked by the Fab7 or Fab8 insulators. Consequently, IAB5 fails to activate white and the 2XPE enhancer fails to activate lacZ (Chopra, 2009b).

Very different results are observed when the transgenes are crossed into mutant embryos containing reduced levels of NELF or DSIF (Spt) subunits. There is a loss in the enhancer-blocking activities of the Fab7 and Fab8 insulators and, as a result, white and lacZ display composite patterns of expression in the mesoderm and abdomen since they are now activated by both enhancers. These results suggest that negative Pol II elongation factors are required for the enhancer-blocking activities of the Fab7 and Fab8 insulators (Chopra, 2009b).

It is proposed that insulators interact with stalled promoters to form higher-order chromatin loop domains, similar to those created by insulator-insulator interactions. Perhaps proteins that bind insulators interact with components of the Pol II complex at stalled genes. Indeed, the recent documentation that the BEAF insulator protein binds to many of the same sites as NELF is consistent with a physical link between stalled Pol II and insulators (Jiang, 2009). The resulting chromatin loops can prevent the inappropriate activation of stalled genes by enhancers associated with neighboring loci. As discussed earlier, stalled Hox genes are located at the boundaries of the ANT-C and BX-C. This arrangement might help ensure that cis-regulatory sequences located outside the complexes do not fortuitously interact with genes contained inside the complex and vice versa. The demonstration that stalled Hox promoters possess an intrinsic insulator activity adds to the intricacy of the chromosomal landscapes that control Hox gene expression in both arthropods and vertebrates (Chopra, 2009b).

Stalled Hox promoters may help promote higher-order chromatin organization within the Hox loci (see illustration). These results suggest that the stalled promoters contain intrinsic insulator activity that requires NELF and DSIF proteins, and this may help define higher-order loops within gene complexes such as the Hox complex. The stalled Pol II along with the NELF and DSIF complex may interact with putative insulator sequences, as seen for the Abd-B promoter and the Fab7. These experiments also suggest that that putative insulator sequences also require NELF and DSIF proteins, and this could be due to sharing of these proteins via the formation of higher-order loops. Such loop domains may help in proper regulation of genes and prevent any aberrant activation from neighboring enhancers, thus favoring proper gene regulations at the higher-order level (Chopra, 2009b).

NELF potentiates gene transcription in the Drosophila embryo

A hallmark of genes that are subject to developmental regulation of transcriptional elongation is association of the negative elongation factor NELF with the paused RNA polymerase complex. This study used a combination of biochemical and genetic experiments to investigate the in vivo function of NELF in the Drosophila embryo. NELF associates with different gene promoter regions in correlation with the association of RNA polymerase II (Pol II) and the initial activation of gene expression during the early stages of embryogenesis. Genetic experiments reveal that maternally provided NELF is required for the activation, rather than the repression of reporter genes that emulate the expression of key developmental control genes. Furthermore, the relative requirement for NELF is dictated by attributes of the flanking cis-regulatory information. It is proposed that NELF-associated paused Pol II complexes provide a platform for high fidelity integration of the combinatorial spatial and temporal information that is central to the regulation of gene expression during animal development (Wang, 2010).

A principle conclusion that emerges from this study is that NELF can play a positive role in supporting transcription in the Drosophila embryo. This finding is somewhat surprising based on NELF's well-characterized properties as a transcriptional inhibitor in vitro and the current view of its role in regulating the hsp70 gene in vivo. So how does a factor that antagonizes transcription elongation play a positive role in promoting gene expression? Depletion of NELF in Drosophila S2 cells leads to reduced expression of a number of genes, and this drop in expression levels correlates with the re-positioning of nucleosomes around the promoter. The idea that NELF stabilizes the local architecture at the promoter that supports transcription is attractive, but the current results indicate these presumptive architectural contributions are not essential for transcription of several endogenous loci in the early embryo (Wang, 2010).

Central to understanding the requirement for NELF in promoting transcription is defining the key differences between the endogenous eve, ftz, rho and slp1 loci and the NELF-dependent reporter genes containing different specific cis-regulatory enhancers from these genes. The results strongly suggest that it is not the basal promoter per se that dictates the requirement for NELF. The basal promoter region contained in the composite slp1[DESE+PESE]-lacZ reporter extends from 260 bp upstream to 121 bp downstream of the transcription start site and includes the entire 5′ untranslated region of the slp1 mRNA. Even more telling is the observation that expression of the ftz-lacC reporter is lost in NELF-A GLC embryos. This reporter contains 6.5 kb of contiguous upstream cis-regulatory DNA extending to 120 bp downstream of the transcription start site. This stands in contrast to the NELF-independent expression of the P{PZ}slp1[05965] enhancer trap inserted 44 bp upstream of the slp1 transcription start site. Transcription of lacZ mRNA from this enhancer trap presumably initiates at the P-element promoter located at the 5′ terminus of this transgene insertion (Wang, 2010).

Although the requirement for NELF appears not to be dictated by the basal promoter, the observations that NELF is specifically localized to promoter regions and does not travel with elongating Pol II complexes strongly suggest the requirement involves NELF-associated Pol II complexes paused downstream of the promoter. The differential effect of NELF-E depletion on expression of the different slp1-lacZ reporters further indicates that the relative requirement for NELF is a function of the extent of flanking cis-regulatory information. Taken together these observations suggest that the relative requirement for NELF depends on interactions involving these flanking cis-regulatory DNA regions and NELF-associated paused Pol II complexes. It is proposed that NELF interacts with Pol II complexes that have initiated transcription but that are not fully competent to enter productive elongation and helps to stabilize these complexes in a form that is competent for responding to activating (or repressing) cues from enhancer-bound transcription factors. In this model the relative requirement for NELF in allowing for active transcription would depend on the strength of the interaction between a promoter and an enhancer and the relative efficiency of generating productive elongation complexes. Although the enhancers contained in the different lacZ reporters used in these experiments are all clearly capable of communicating with the promoter it would certainly be expected that this communication would be less efficient than for endogenous loci that contain the full complement of flanking DNA that has evolved to optimize the regulation of gene expression at this stage. Importantly, the NELF-dependent expression of these reporters strongly suggests that the generation of a productive Pol II elongation complex is the key step in the transcription cycle that is targeted for integrating the regulatory cues that drive the patterned expression of these genes in the early embryo (Wang, 2010).

NELF clearly has a pleiotropic role during Drosophila development. Perturbations in maternally provided NELF lead to two distinct embryonic lethal phenotypes. The observation that both phenotypes, albeit with different penetrance are produced either by elimination of maternal NELF-A or by reduction of maternal NELF-E strongly suggests both phenotypes are due to decreased activity of the NELF complex. The early arrest phenotype occurs prior to the onset of transcription in the embryo and thus is most likely due to defects that occur during oogenesis. The maternally provided histone mRNAs are one likely candidate as a prospective target of NELF activity during oogenesis. NELF is required for the proper processing of the 3′ termini of replication-dependent histone mRNAs in HeLa cells, and Drosophila embryos with defects in the processing of maternally provided histone mRNAs arrest during the nuclear division cycles that precede the blastoderm stage. Further studies should reveal whether the early arrest of NELF-A and NELF-E GLC embryos reflects a conserved role for NELF in the 3′-end processing of histone mRNAs (Wang, 2010).

The finding that many genes have paused Pol II complexes at their 5′ end strongly suggests that the regulation of transcription elongation is a widespread phenomenon in higher eukaryotes. Recent studies indicate that more than one third of all genes in Drosophila S2 cells generate short, 5′-capped RNAs similar to those produced by stalling of Pol II. The results of Pol II chromatin immunoprecipitation whole genome microarray assays suggest that paused Pol II complexes are formed on approximately 10% of genes in the blastoderm stage Drosophila embryo. This is almost certainly an underestimate as five of the seven genes for which NELF association has been demonstrated were not identified as having paused Pol II complexes at this stage. Indeed, the stringent cut-off used in this study led to the assignment of slp1 as a member of the 27% of genes that have uniform Pol II association in the blastoderm embryo (Wang, 2010).

It is furthermore clear that NELF association is developmentally regulated as neither srya nor any of the five segmentation genes for which NELF association in the early embryo was demonstrated in this study are also associated with NELF in S2 cells. Amongst these six genes with early embryonic association of NELF there are differences in the level of association at different developmental stages. The two genes with the most rapid loss of NELF, ftz and srya show little to no expression after four hours of development. Thus NELF is not involved in the stable maintenance of repression at these later stages, which involves instead other mechanisms such as epigenetic maintenance by the Polycomb group proteins and specific histone methylation marks. The observation that NELF association is also reduced on genes such as en and wg that have increased expression levels at later stages may suggest that NELF is not involved in the ongoing expression of these genes at later stages. However, as the embryo is comprised of a mixture of expressing and non-expressing cells it will be important to examine NELF association specifically in cells expressing these genes before coming to this conclusion (Wang, 2010).

The high levels of NELF association with the promoter regions of a number of genes involved in segmentation and other early developmental processes serves to emphasize the unique and pivotal aspects of this critical stage of Drosophila embryogenesis. Pre-blastoderm nuclei are totipotent and come to be specified in response to maternally-provided positional information and the action of the genetic systems that respond to this information. The regulation of gene transcription is central to the initial specification of cell fates along both the anterior-posterior and dorsal-ventral axes of the early embryo, and it is clear that regulation of transcription elongation is central to this process. Similar to Drosophila blastoderm nuclei, the pluripotent properties of human embryonic stem cells are reflected by the presence of paused Pol II complexes on a wide number of genes, including many key developmental regulators. Further studies on the mechanisms of developmentally regulated transcription elongation are clearly of great importance for understanding the initial programming of cell fates expression during animal embryogenesis (Wang, 2010).

Interactions between DSIF (DRB sensitivity inducing factor), NELF (negative elongation factor), and the Drosophila RNA polymerase II transcription elongation complex

Negative elongation factor (NELF) and 5,6-dichloro-1-beta-D-ribofuranosylbenzimidazole sensitivity-inducing factor (DSIF) are involved in pausing RNA Polymerase II (Pol II) in the promoter-proximal region of the hsp70 gene in Drosophila, before heat shock induction. Such blocks in elongation are widespread in the Drosophila genome. However, the mechanism by which DSIF and NELF participate in setting up the paused Pol II remains unclear. The interactions were analyzed among DSIF, NELF, and a reconstituted Drosophila Pol II elongation complex to gain insight into the mechanism of pausing. The results show that DSIF and NELF require a nascent transcript longer than 18 nt to stably associate with the Pol II elongation complex. Protein-RNA cross-linking reveals that Spt5, the largest subunit of DSIF, contacts the nascent RNA as the RNA emerges from the elongation complex. Taken together, these results provide a possible model by which DSIF binds the elongation complex via association with the nascent transcript and subsequently recruits NELF. Although DSIF and NELF were both required for inhibition of transcription, no NELF-RNA contact is detected when the nascent transcript was between 22 and 31 nt long, which encompasses the region where promoter-proximal pausing occurs on many genes in Drosophila. This raises the possibility that RNA binding by NELF is not necessary in promoter-proximal pausing (Missra, 2010).

DSIF and NELF are key factors in pausing Pol II in the promoter-proximal region of genes in Drosophila and human cells. To gain insight into the mechanism by which DSIF and NELF contribute to promoter-proximal pausing, a system was developed in which the physical interaction of DSIF and NELF with a Pol II elongation complex could be monitored using a native gel electrophoresis assay. Previously, it was demonstrated that DSIF alone could associate with the Pol II elongation complex. In this study a method was developed to purify Drosophila NELF, thus allowing exploration of the interplay of DSIF and NELF with the elongation complex (Missra, 2010).

The results show that the association of NELF with the elongation complex is dependent on the presence of DSIF. Previous work provided evidence that NELF associated with preformed complexes of DSIF and Pol II in nuclear extracts but the interaction of DSIF and Pol II was not dependent on NELF. These interactions were likely occurring outside the context of an elongation complex and were relatively weak because the bulk of DSIF, NELF, and Pol II exist independent of each other in nuclear extracts. In contrast, the current results show that NELF can significantly influence the binding of DSIF to Pol II within the context of an elongation complex when limiting amounts of DSIF are present. Since Pol II, DSIF, and NELF have been shown to interact individually with each other, it is likely that this network of interactions contributes to stable association of these proteins in the context of the elongation complex (Missra, 2010).

The binding assays show that the length of the nascent transcript affects the association of DSIF and NELF with the elongation complex. While binding of DSIF alone or in combination with NELF to the elongation complex was evident for an elongation complex with a nascent transcript of 22 nt, no binding was detected when the nascent transcript was 18 nt long. These results are consistent with the finding that human DSIF and NELF require transcripts ≥18 nt long to inhibit transcription, and also a recent study showed human DSIF preferentially bound elongation complexes containing transcripts that were at least 25 nt long. The 5′ end of an 18-nt-long nascent transcript just begins to emerge from the surface of Pol II. Exposure of four additional nucleotides appears to be sufficient for binding of DSIF alone or with NELF. Notably, the association of DSIF with the elongation complex is not simply due to nonspecific interaction with the RNA or DNA because previous experiments show that binding of DSIF to the elongation complex requires specific contacts with Pol II (Missra, 2010).

One way in which nascent transcript length could affect the association of DSIF and NELF is by providing an additional binding site in the elongation complex. Previous results have directed attention at an RRM in NELF-E. Mutations in this RRM impair the capacity of NELF to inhibit elongation in the presence of DSIF. However, these experiments focused on elongation over distances greater than 100 nt. The finding that DSIF associates with elongation complex containing a 22-nt-long radioactive transcript (EC22) but not EC18 suggests that DSIF rather than NELF might be interacting with the nascent transcript, and RNA-protein cross-linking data support this hypothesis. The Spt5 subunit of Drosophila and human DSIF contains five Kyprides, Ouzounis, Woese (KOW) domains. An isolated KOW domain from Aquifex aeolicus NusG has been shown to associate with RNA, so it is possible that one of these domains in Spt5 is contacting the nascent transcript as it emerges from the elongation complex (Missra, 2010).

Cross-linking analysis detected contact between NELF-E and the nascent transcript in EC70 but not in EC31. The 5′ end of the nascent transcript contacts Rpb7 when its length is between 26 to 32 nt. Therefore it is possible that a longer nascent transcript is required to allow contact with NELF. Given that promoter-proximal pausing can occur before Pol II transcribes 30 nt, it is proposed that the RRM of NELF-E is not involved in promoter-proximal pausing. Its role could be limited to processes involving longer nascent transcripts such as regulation that appears to involve the transactivation response element of HIV or 3′ end formation of histone mRNAs (Missra, 2010).

The finding that DSIF and NELF associate with EC22 but not with EC18 is very relevant to the process of promoter-proximal pausing. Permanganate genomic footprinting of over 60 different promoters reveals that Pol II pauses in the promoter-proximal region 20 to 50 nt downstream from the transcription start site. Those cases where the Pol II appeared to be pausing closer to a transcription start site were found to have the start sites inaccurately mapped. Thus, the promoter-proximal limit for the range where Pol II pauses is likely to be dictated by the minimum length of RNA required for DSIF to associate with the elongation complex (Missra, 2010).

From the results presented in this study, it is proposed that the first step in promoter-proximal pausing involves binding of DSIF to the nascent transcript. NELF subsequently associates to form a stable complex. Importantly, this complex alone is not sufficient to stably pause the Pol II as the results show that elongation is slowed but not halted in reactions involving only these three proteins. Hence other factors that remain to be identified are likely to act in concert with this core complex of DSIF, NELF, and Pol II to stably pause Pol II in the promoter-proximal region of genes. Since transcription in vivo occurs on chromatin, nucleosomes may cooperate with DSIF and NELF in setting up the paused polymerase. The experimental approach described in this study could serve as a way to identify additional factors involved in pausing (Missra, 2010).

Negative elongation factor accelerates the rate at which heat shock genes are shut off by facilitating dissociation of heat shock factor

Promoter-proximal pausing of RNA polymerase II (Pol II) occurs on thousands of genes in animal cells. This pausing often correlates with the rapid induction of genes, but direct tests of the relationship between pausing and induction rates are lacking. hsp70 and hsp26 in Drosophila are rapidly induced by heat shock. Contrary to current expectations, depletion of negative elongation factor (NELF), a key factor in setting up paused Pol II, reduced pausing but did not interfere with rapid induction. Instead, depletion of NELF delayed the time taken for these genes to shut off during recovery from heat shock. NELF depletion also delayed the dissociation of HSF from hsp70 and hsp26, and a similar delay was observed when cells were depleted of the histone acetyltransferase CBP. CBP has been reported to associate with Pol II, and acetylation of HSF by CBP has been implicated in inhibiting the DNA-binding activity of HSF. It is proposed that NELF-mediated pausing allows Pol II to direct CBP-mediated acetylation of HSF, thus causing HSF to dissociate from the gene. Activators are typically viewed as controlling Pol II. These results reveal a possible reciprocal relationship in which paused Pol II influences the activator (Ghosh, 2011).

Reduction of promoter proximal pausing on hsp70 does not alter the rate of heat shock induction. Correlations between the presence of paused Pol II and rapid induction of gene expression have led to the hypothesis that promoter proximal pausing provides the basis for rapid induction. However, none of these studies have actually tested if disruption of promoter proximal pausing delays expression of genes. The current analysis reveals that reduction of promoter proximal pausing at hsp70 does not detectably alter the rate of heat shock induction. This was evident by monitoring the association of Pol II at both the 5’ and 3’ ends of hsp70 using permanganate genomic footprinting and also by measuring synthesis of hsp70 RNA. The two types of assays are in good agreement with each other in that a substantial increase in the level of transcript and in the presence of Pol II at the 3’ end of the gene occurs between 2 and 4 minutes after heat shock. Also, the elongation rate of approximately 1.25 kb/min for transcription of hsp70 that can be deduced from permanganate data is in excellent agreement with other types of measurements, thus validating the use of permanganate footprinting as a means for analyzing the kinetics of transcription induction and elongation. It has been proposed that the establishment of paused Pol II at a promoter allows for rapid induction because steps that precede initiation such as chromatin remodeling and Pol II recruitment are bypassed. In the case of the heat shock genes, these steps are still likely to occur prior to heat shock induction even when promoter proximal pausing is impaired. TFIID, which provides the foundation for assembling a preinitiation complex, makes multiple sequence-specific contacts with hsp70 including the TATA box, the initiator, and downstream sequences. In addition, another DNA binding protein called GAGA factor maintains the promoter region in an accessible state even when TFIID binding is impaired. Since HSF associates with a transgenic hsp70 promoter deleted of its TATA box, the accessible state of the promoter in lieu of TFIID appears to be sufficient for HSF binding. Thus access of DNA by the general transcriptional machinery and HSF may not be limiting for hsp70 even when promoter proximal pausing is disrupted. The extent to which promoters retain an accessible state upon loss of paused Pol II is likely to vary. The Drosophila heat shock genes could represent one end of a spectrum where accessibility is retained without sustaining full occupancy of the promoter with paused Pol II because of the binding of GAGA factor and TFIID. At the other end of the spectrum may be genes that require the paused Pol II to maintain a potentially inducible state. Cases have been identified where the paused Pol II appears to contribute to induction by preventing a nucleosome from assembling over the promoter (Ghosh, 2011).

Depletion of NELF impacts the ability of hsp70 to shut-off during recovery from heat shock. Permanganate genomic footprinting provides a high-resolution view of the behavior of Pol II. Permanganate reactivity on hsp70 at positions +7 and +8 are indicative of newly initiated Pol II whereas reactivity downstream from +34 corresponds to Pol II that has read through the pause. Similarly for hsp26, permanganate reactivity at +9 and +10 are indicative of newly initiated Pol II whereas reactivity downstream from +45 corresponds to Pol II that has read through the pause. The patterns of permanganate reactivity that were observed indicate that both initiation and read-through persist for a longer time in NELF-depleted glands than in control glands during recovery from heat shock for both hsp70 and hsp26. Corroborating this conclusion is the demonstration that a brief heat shock causes these two genes to be more highly expressed in NELF-depleted glands than in control glands. Limiting the production of hsp70 protein to a level appropriate to the degree of stress could be critical for proper cell development, since ectopic expression of hsp70 in salivary glands under non-heat shock conditions inhibits growth of the cells in this tissue (Ghosh, 2011).

These results now expand the types of scenarios in which NELF and promoter proximal pausing serve to attenuate gene expression. In the case of the estrogen-mediated activation of the pS2 gene, direct interaction between the NELF-B subunit and the estrogen receptor causes estrogen-mediated association of NELF with the target gene. Depletion of NELF results in hyper-activation of the pS2 gene, indicating that the estrogen receptor is mediating both activation and repression to achieve a particular level of expression in the presence of stimuli. In the case of junB, depletion of NELF increased the expression of junB both before and after serum-mediated activation. The current results now reveal a role for NELF in shutting off gene expression upon withdrawal of the stimulus - this case being heat shock (Ghosh, 2011).

To investigate why the shut off of the heat shock genes was delayed by depletion of NELF, the association of HSF with hsp70 and hsp26 in salivary glands was measured using chromatin immunoprecipitation. Significantly more HSF was detected on the heat shock genes after 45 minutes of recovery in NELF-depleted than control glands indicating that the delay in shutting off the heat shock genes could be due to a delay in the dissociation of HSF and that NELF is somehow involved in this dissociation. It is unlikely that the dissociation of HSF is due to stress induced by depletion of NELF, since there is no evidence of this depletion inducing a stress response in salivary glands. Prior to heat shock, both control and NELF-depleted glands exhibit low levels of heat shock gene expression, low levels of HSF at heat shock gene promoters, and undetectable levels of HSF DNA-binding activity in amount of HSF detected in glands nor did it alter a heat shock-dependent shift in the mobility of HSF on SDS-PAGE. Approximately 2-fold higher HSF binding activity was detected in extracts from NELF-depleted glands than control glands after recovery from heat shock, but the standard deviations in the measurements indicate that this difference is not significant. Since the results showed that depletion of NELF resulted in a delay in the dissociation of HSF from heat shock genes during the recovery from heat shock, recent indirect evidence in human cells that acetylation of HSF by CBP regulates HSF binding activity was intriguing. The finding that depletion of CBP delays the dissociation of HSF from the heat shock genes provides the first direct evidence that CBP is involved in regulating HSF’s interaction with heat shock genes in vivo. Since depletion of CBP and NELF each delays the dissociation of HSF from the heat shock genes during recovery, it is proposed that their actions are linked. It is posited that NELF-mediated pausing of Pol II could position CBP to acetylate HSF, thus causing HSF to dissociate from the heat shock gene. Several observations provide support for this model. CBP associates with Pol IIa and not Pol IIo and the Pol II paused in the promoter proximal region is in the IIa state. Also, CBP associates with the heat shock genes in Drosophila during heat shock induction. Further investigation of this proposal will be aided by biochemical analyses of interactions between CBP and reconstituted paused elongation complexes. Multiple mechanisms can attenuate the activity of HSF during heat shock, so the possibility cannot be ruled out that the actions of NELF and CBP are independent of each other or that these proteins are acting indirectly on HSF by influencing other regulators of HSF. The heat shock proteins such as hsp70 produced during heat shock repress the activation domain of HSF thus providing a negative feedback loop that limits heat shock gene transcription. These chaperons have also been implicated in attenuating the DNA binding activity of HSF in Drosophila during extended periods of a moderate heat shock. Changes in phosphorylation of HSF have also been implicated in regulating its activity, but such changes do not appear to effect the DNA binding activity of HSF in Drosophila. The mechanisms by which heat shock genes are returned to their uninduced state during recovery from heat shock remain to be fully elucidated. The use of permanganate footprinting to directly monitor the behavior of Pol II at heat shock genes during recovery from heat shock in cells depleted of specific proteins provides a way to identify candidates involved in shutting off transcription of the heat shock genes. A significant advantage of this approach is that it allows one to directly monitor the reestablishment of the paused state (Ghosh, 2011).

Paused Pol II could function as a regulator of gene regulatory factors. Many stimulus-responsive and developmental genes are found to have paused Pol II, and attenuation of transcription of these genes is likely to be critical for establishing appropriate physiological or developmental programs of gene expression. The finding that paused Pol II influences dissociation of an activator from the promoter provides a mechanism for rapidly shutting off gene expression. The pausing of Pol II could function more broadly by allowing the Pol II to remain stationary at a promoter while it serves to recruit modulators of chromatin structure and gene expression (Ghosh, 2011).

Search PubMed for articles about Drosophila Nelf

Aiyar, S. E., Blair, A. L., Hopkinson, D. A., Bekiranov, S. and Li, R. (2007). Regulation of clustered gene expression by cofactor of BRCA1 (COBRA1) in breast cancer cells. Oncogene 26: 2543-2553. PubMed Citation: 17043641

Albert, I., Mavrich, T.N., Tomsho, L.P., Qi, J., Zanton, S.J., Schuster, S.C., and Pugh, B.F. (2007). Translational and rotational settings of H2A.Z nucleosomes across the Saccharomyces cerevisiae genome. Nature 446: 572-576. PubMed Citation: 17392789

Chopra, V. S., Hong, J. W. and Levine, M. (2009a). Regulation of Hox gene activity by transcriptional elongation in Drosophila. Curr. Biol. 19: 688-693. PubMed Citation: 19345103

Chopra, V. S., Cande, J., Hong, J. W. and Levine, M. (2009b). Stalled Hox promoters as chromosomal boundaries. Genes Dev. 23(13): 1505-9. PubMed Citation: 19515973

Gilchrist, D. A., et al. (2008). NELF-mediated stalling of Pol II can enhance gene expression by blocking promoter-proximal nucleosome assembly. Genes Dev. 22(14): 1921-33. PubMed Citation: 18628398

Ghosh, S. K., Missra, A., Gilmour, D. S. (2011). Negative elongation factor accelerates the rate at which heat shock genes are shut off by facilitating dissociation of heat shock factor. Mol. Cell Biol. 31(20): 4232-43. PubMed Citation: 21859888

Jiang, N., Emberly, E., Cuvier, O. and Hart, C. M. (2009). Genome-wide mapping of BEAF binding sites in Drosophila links BEAF to transcription. Mol. Cell. Biol. 29(13): 3556-68. PubMed Citation: 19380483

Kaplan, C. D., Morris, J. R., Wu, C. and Winston, F. (2000). Spt5 and spt6 are associated with active transcription and have characteristics of general elongation factors in D. melanogaster. Genes Dev. 14: 2623-2634. PubMed Citation: 11040216

Lee, C., et al. (2008). NELF and GAGA factor are linked to promoter-proximal pausing at many genes in Drosophila. Mol. Cell. Biol. 28: 3290-3300. PubMed Citation: 18332113

Mandal, S.S., Chu, C., Wada, T., Handa, H., Shatkin, A.J., Reinberg, D. (2004). Functional interactions of RNA-capping enzyme with factors that positively and negatively regulate promoter escape by RNA polymerase II Proc. Natl Acad. Sci. 101: 7572-7577. Medline abstract: 15136722

Mavrich, T. N., et al. (2008). Nucleosome organization in the Drosophila genome. Nature 453: 358-362. PubMed Citation: 18408708

Missra, A. and Gilmour, D. S. (2010). Interactions between DSIF (DRB sensitivity inducing factor), NELF (negative elongation factor), and the Drosophila RNA polymerase II transcription elongation complex. Proc. Natl. Acad. Sci. 107(25): 11301-6. PubMed Citation: 20534440

Muse, G. W., Gilchrist, D. A., Nechaev, S., Shah, R., Parker, J. S., Grissom, S. F., Zeitlinger, J. and Adelman, K. (2007). RNA polymerase is poised for activation across the genome. Nat. Genet. 39: 1507-1511. PubMed Citation: 17994021

Narita, T., Yung, T. M., Yamamoto, J., Tsuboi, Y., Tanabe, H., Tanaka, K., Yamaguchi, Y. and Handa, H. (2007). NELF interacts with CBC and participates in 3' end processing of replication-dependent histone mRNAs. Mol. Cell 26: 349-365. PubMed Citation: 17499042

Saunders, A., Werner, J., Andrulis, E.D., Nakayama, T., Hirose, S., Reinberg, D., Lis, J.T. (2003) Tracking FACT and the RNA polymerase II elongation complex through chromatin in vivo Science. 301: 1094-1096. Medline abstract: 12934007

Saunders, A., Core, L. J. and Lis, J. T. (2006). Breaking barriers to transcription elongation. Nat Rev Mol Cell Biol 7: 557-567. PubMed Citation: 16936696

Wada, T., et al. (1998). DSIF, a novel transcription elongation factor that regulates RNA polymerase II processivity, is composed of human Spt4 and Spt5 homologs. Genes Dev. 12: 343-356. PubMed Citation: 9450929

Wang, X., Lee, C., Gilmour, D. S. and Gergen, J. P. (2007). Transcription elongation controls cell fate specification in the Drosophila embryo. Genes Dev. 21(9): 1031-6. Medline abstract: 17473169

Wang, X., Hang, S., Prazak, L. and Gergen, J. P. (2010). NELF potentiates gene transcription in the Drosophila embryo. PLoS One 5(7): e11498. PubMed Citation: 20634899

Wu, C.-H., et al. (2003). NELF and DSIF cause promoter proximal pausing on the hsp70 promoter in Drosophila. Genes Dev. 17: 1402-1414. Medline abstract: 12782658

Wu, C. H., Lee, C., Fan, R., Smith, M. J., Yamaguchi, Y., Handa, H., Gilmour, D. S. (2005). Molecular characterization of Drosophila NELF. Nucleic Acids Res. 33(4): 1269-79. Medline abstract: 15741180

Yamaguchi, Y., Wada, T. and Handa, H. (1998). Interplay between positive and negative elongation factors: Drawing a new view of DRB. Genes Cells 3: 9-15. PubMed Citation: 9581978

date revised: 28 December 2011

Home page: The Interactive Fly © 2007 Thomas Brody, Ph.D.