Repression of RNA Polymerase II Elongation In Vivo Is Critically Dependent on the C-Terminus of Spt5

The stalling of RNA polymerase II (RNAPII) at the promoters of many genes, including developmental regulators, stress-responsive genes, and HIVLTR, suggests transcription elongation as a critical regulatory step in addition to initiation. Spt5, the large subunit of the DRB sensitivity-inducing factor (DSIF), represses or activates RNAPII elongation in vitro. How RNAPII elongation is repressed in vivo is not well understood. Here we report that CTR1 and CTR2CT, the two repeat-containing regions constituting the C-terminus of Spt5, play a redundant role in repressing RNAPII elongation in vivo. First, mis-expression of Spt5 lacking CTR1 or CTR2CT is inconsequential, but mis-expression of Spt5 lacking the entire C-terminus (termed NSpt5) dominantly impairs embryogenesis in zebrafish. Second, NSpt5 de-represses the transcription of hsp70-4 in zebrafish embryos and HIVLTR in cultured human cells, which are repressed at the RNAPII elongation step under non-inducible conditions. Third, NSpt5 directly associates with hsp70-4 chromatin in vivo and increases the occupancy of RNAPII, positive transcription elongation factor b (P-TEFb), histone H3 Lys 4 trimethylation (H3K4Me3), and surprisingly, the negative elongation factor A (NELF-A) at the locus, indicating a direct action of NSpt5 on the elongation repressed locus. Together, these results reveal a dominant activity of NSpt5 to de-repress RNAPII elongation, and suggest that the C-terminus of Spt5 is critical for repressing RNAPII elongation in vivo.


Introduction
The production of mRNA is a multi-step process that involves transcription initiation, elongation, and termination [1][2][3].For decades, the major mechanism of gene regulation in higher eukaryotes is thought to reside at the level of recruitment of RNAPII by sequence-specific DNA binding factors, despite the presence of RNAPII proximal to the promoters of a few genes including hsp70 [4] and HIVLTR [5] in un-induced conditions.But recently, it has been observed that many transcriptionally repressed genes have promoter-proximal paused or stalled RNAPII [6][7][8][9].Moreover, biochemical studies have identified over a dozen proteins and a small nuclear RNA, which regulate RNAPII elongation in vitro [2,10,11].
Despite these advancements, how Spt5 regulates RNAPII elongation in vivo is not well understood.Although genetic analyses in yeast [22], C. elegans [23], Drosophila [24], and zebrafish [25][26][27] have revealed an essential role of Spt5 in cell growth and embryonic development, these studies do not address whether these phenotypes are caused by a defect in RNAPII elongation, since Spt5 has also been implicated in regulating mRNA capping [28,29], splicing [30], 39 end processing [31,32], and mRNA export [33].Despite that point mutations located in the Cterminus of zebrafish [25) or Drosophila {Jennings, 2004 #2125] Spt5 are shown to disrupt the elongation repressive activity in vitro, a loss of repressive activity of these mutant proteins has not been demonstrated in vivo.In fact, neither the zebrafish point mutation nor the zygotic null mutation of spt5 significantly de-represses hsp70-4, a gene that is repressed at the elongation level (Chen and Guo, unpublished observations, and this study).
The positive role of Spt5 in transcription elongation in vivo has been supported by its occupancy detected at many transcriptionally active chromosomal sites and its rapid recruitment to endogenous and transgenic heat shock loci upon heat shock, whereas its repressive activity has been suggested by its presence at the promoter proximal region in un-induced heat shock gene loci [34,35].Recently, expression profiling together with in vivo chromatin immunoprecipitation (ChIP) studies in the zebrafish spt5 mutant reveal essential target genes that are occupied and regulated by Spt5, hence providing direct evidence that Spt5 indeed has dual activity in regulating RNAPII elongation in vivo [36].
In this study, we sought to determine how such in vivo dual activity of Spt5 is encoded in the protein.Through a structurefunction study in zebrafish, we found that deleting either CTR1 or CTR2CT does not significantly affect Spt5's in vivo activity, suggesting that partially redundant functions reside in the Spt5 Cterminus.However strikingly, deleting the entire C-terminus yielded NSpt5, which had a dominant activity that impaired embryogenesis in zebrafish.Using hsp70-4 and HIVLTR, two examples of RNAPII elongation -repressed genes, we found that NSpt5 acted to de-repress their transcription.In vivo ChIP further uncovered that NSpt5 directly associated with hsp70-4 chromatin in vivo and increases the occupancy of RNAPII, P-TEFb, H3K4Me3, and surprisingly, NELF-A at the locus, indicating a direct action of NSpt5 on the elongation repressed hsp70-4 gene in vivo.Together, these results reveal a dominant activity of NSpt5 to de-repress RNAPII elongation, and suggest that the C-terminus of Spt5 is critical for repressing RNAPII elongation in vivo.

Redundant Activity of CTR1 and CTR2CT Domains of Spt5 in zebrafish development
To dissect the activity of Spt5 (Figure 1A) in vivo, we employed a RNA rescue assay in zebrafish (Figure 1).The fog s30 allele (Figure 1B), which harbors a deletion of the entire locus of spt5, was used as the ground state of spt5 activity (a zygotic null with residual maternal Spt5 activity) [26].FLAG epitope-tagged wild type (WT) spt5 (F-spt5) mRNA rescued fog s30 mutant phenotypes fully, as assessed by the overall normal morphology (Figure S1, and Table 1) and the proper development of dopaminergic (DA) neurons at 30 hours post fertilization (hpf)(Figure S2B).F-spt5 had no discernible effect in WT embryos, suggesting that overexpression of spt5 alone does not interfere with its function in vivo (Table 1).
Using this in vivo assay, the activity of a series of spt5 deletion variants was tested.Notably, removal of the RNAPII-binding domain [15,37] in Spt5 (F-spt5DRNAPII-BD) abolished its ability to rescue fog s30 embryos (Table 1 and Figure S2C), despite the variant protein being detected readily in the embryo (data not shown).This observation suggests that the activity of Spt5 in vivo is mediated via its interaction with RNAPII.The expression of F-spt5DRNAPII-BD had no discernible effect in WT embryos (Figure S2C and Table 1).
Next, we turned our attention to the C-terminus of Spt5 (CSpt5), which is composed of two repeat-containing regions named CTR1 and CTR2, and a small new domain called CT (Figure 1A).CTR1 contains multiple hepta-peptide repeats that are phosphorylated by P-TEFb [21], and is considered as an important domain for Spt5's positive elongation activity, as the prevention of CTR1 phosphorylation impairs epidermal growth factor (EGF)-induced c-fos expression but not its basal transcription in Hela cells [19].However, CTR1-deleted Spt5 (F-spt5DCTR1) fully rescued the morphological defects of fog s30 embryos (Figure 1C) and the development of DA neurons (Figure S2D).Spt5 with a deletion of the other half of CSpt5 (F-spt5DCTR2CT) also fully rescued the embryonic morphology (Table 1) but not DA neuron development (Figure S2E).This is consistent with the previous published observation, in which a point mutation in CT affects DA neuron development but not embryonic morphology [25].Taken together, CTR1 and CTR2CT appear to carry out partially redundant functions in vivo.
Removal of the entire C-terminus (CSpt5) unleashes a dominant activity that resides in the N-terminus of Spt5 (NSpt5) We next determined the functional effect of removing the entire C-terminus (CSpt5).Spt5 with both CTR1 and CTR2CT deleted (NSpt5) not only failed to rescue fog s30 mutant phenotypes, including the body shape, brain morphology, eyes, heart, and blood circulation, but also exacerbated them.Moreover, NSpt5 impaired dominantly the development of WT embryos (Figure 1D, and Table 1).WT embryos injected with F-Nspt5 RNA (F-Nspt5-expressing embryos) exhibited a similar phenotype to NSpt5-expressing fog s30 embryos, with severe deformity including small size, deformed brain and eyes, heart edema, lack of blood circulation, and dorsally curved body axis (Figure 1D), with the phenotypes worsening when more F-Nspt5 RNA was injected.No significant increase in cell death was detected in NSpt5-expressing embryos (data not shown).These results suggest that Nspt5 has no rescuing activity, but exhibits a dominant negative effect on the endogenous Spt5 (both the normal levels of Spt5 in WT and residual maternal Spt5 in the fog s30 mutant).

NSpt5 de-represses hsp70 expression in the absence of heat shock but does not affect the induction of hsp70-4 upon heat shock
To understand NSpt5's dominant activity, we carried out gene expression profiling on WT and NSpt5-expressing embryos.We identified a number of genes whose transcript levels were significantly increased in NSpt5-expressing embryos (H. Chen, unpublished data).One of the most prominent is hsp70-4: its transcript level was increased ,10 fold in NSpt5-expressing embryos.Since hsp70 is a model gene whose transcription elongation is regulated by Spt5 [2], we focused our analysis on hsp70-4 in this study.Multiple members of the hsp70 gene family were identified in zebrafish, and hsp70-4 was shown to be heat inducible [38].A GFP reporter line driven by hsp70-4 regulatory elements mimics the endogenous hsp70-4 expression [39], and moreover, the induction of GFP upon heat shock in this line was attenuated in fog s30 embryos [26], suggesting that its expression is under Spt5 regulation.
We first examined the impact of NSpt5 on hsp70-GFP expression.In the absence of heat shock, no GFP signal was detected in WT (Figures 2A and D) or in fog s30 embryos (Figure 2B).Although our previous studies did detect ,2 fold increase of basal hsp70-4 transcripts in the fog sk8 mutant [36], such modest increase might not cause sufficient accumulation of GFP protein to be above the detection sensitivity of the fluorescent microscope.However, in WT embryos expressing F-Nspt5, a strong GFP signal was detected (Figures 2C and E), often in patches of embryonic cells.The patchiness of the signal could possibly be due to transient and mosaic expression of the injected F-Nspt5 RNA, or alternatively, due to unknown spatial and temporal constraints on hsp70-4 expression in developing embryos.Double labeling with FLAG and GFP antibodies revealed that the GFP signal was detected in cells that co-expressed the FLAG epitope, indicating a cell-autonomous effect of F-NSpt5 (Figure 2E).Upon heat shock for 1 hour, hsp70-GFP was induced equally well in F-spt5or F-Nspt5-expressing embryos (Figure 2F).
The effect of NSpt5 on endogenous hsp70-4 expression was further confirmed by quantitative RT-PCR and in situ hybridization.About 14 fold increase of hsp70-4 transcripts were detected in Nspt5-expressing embryos (Figure 2G).In situ hybridization further revealed the increase and the spatial distribution of hsp70-4 transcripts in Nspt5-expressing embryos (Figure S3).With heat shock, hsp70-4 expression was dramatically increased, and such heat shock-induced hsp70-4 expression was unaffected by Nspt5 mis-expression (Figure 2H).Together, these results suggest that NSpt5 can interfere with the elongation repressive activity of Spt5 (revealed by the increase of basal hsp70-4 expression), but not the elongation stimulatory activity of Spt5 (revealed by the unalteration of heat shock-induced hsp70-4 expression).

NSpt5 de-represses HIVLTR transcription under basal conditions but does not affect Tat trans-activation in human cells
To determine whether NSpt5's effect is more general than one gene (hsp70-4) and one species (zebrafish), we examined HIVLTR, which harbors a promoter-proximal stalled RNAPII [5] and Spt5 [40] in human Hela cells.In the absence of the viral transactivator Tat, NELF and DSIF cooperatively promote the arrest of RNAPII on the HIVLTR, resulting in the accumulation of short transcripts [41].However, when Tat is recruited to the transactivation response (TAR) RNA, Spt5 helps to increase rates of productive elongation [41][42][43].Since Spt5 from zebrafish is homologous in structure and function to human Spt5 protein [25], F-Spt5 or F-NSpt5 were co-expressed with RD (a component of NELF) in HeLa cells in the absence or presence of Tat (Figure 2I), and the activity of chloramphenicol acetyltransferase (CAT) fused to HIVLTR was examined.In the absence of Tat, the expression of F-NSpt5 resulted in almost a four fold increased expression from the HIVLTR (Figure 2I, bar 2) as compared to that of F-Spt5 control (Figure 2I,, bar 1).In the presence of Tat, there was no significant difference between F-Spt5-or F-NSpt5-mediated CAT activity (Figure 2I, bars 3 and 4).These results suggest that NSpt5 again interferes with Spt5's repressive but not its stimulatory activity in HIVLTR transcription, similar to its action on hsp70-4.

NSpt5 impairs the repressive but not the stimulatory activity of DSIF in a dominant manner in vitro
To further test the idea that NSpt5 interferes with Spt5's repressive but not its stimulatory activity, we employed an in vitro system, which measures Spt5's activating activity in the presence of constitutively active P-TEFb and its repressive activity upon P-TEFb inactivation via the kinase inhibitor DRB [12].The elongation stimulation activity of DSIF was assayed using pSLG402 as a template, which generates short (promoterproximal) and long (promoter-distal) RNase T1-resistant products under the control of the adenovirus major-late promoter.The long (promoter-distal) transcripts are dependent on the elongation stimulatory activity of DSIF.Nuclear extracts (NE) containing the constitutively active P-TEFb and WT DSIF led to a timedependent increase of the long (promoter-distal) RNase T1resistant products, reflecting the elongation stimulatory activity of Spt5 (Figure 3A, first three lanes).Increasing amounts of NSpt5 were added to normal HeLa nuclear extract and were found to modestly enhance elongation efficiency (Figure 3A).This may be due to its dominant-negative effect on the repressive activity of endogenous DSIF, but this could also be explained by its residual elongation activation potential.To clarify this point, we used a different transcription system, in which synthesis of a 380-nt RNase T1-resistant product was examined in the presence or absence of the transcriptional inhibitor DRB (Figure 3B).DRB blocks P-TEFb kinase activity, and therefore the repression activity is clearly seen in the presence of DRB (Figure 3B, first two lanes), Table 1.The ability of deletion variants in rescuing fog s30 mutant embryos and producing dominant phenotypes in WT embryos.

RNA variants injected
Ability to rescue fogs30 morphology (%, n = a/bˆ) Ability to produce dominant interference  whereas its activation activity is not clearly seen because of relatively short transcript length.Consequently, while NSpt5 had only a modest effect in the absence of DRB, NSpt5 strongly interfered with the transcription inhibition by DRB (Figure 3B), suggesting that NSpt5 indeed acts in a dominant-negative manner, capable of inhibiting the repressive activity of endogenous DSIF.These findings provide biochemical evidence that NSpt5's dominant activity in vivo is due to its ability to interfere with endogenous Spt5's repressive but not its stimulatory activity.

NSpt5 increases RNAPII occupancy on the HIVLTR and hsp70-4 chromatin in human cells and zebrafish
To further examine NSpt5's mechanism of action in intact cells, we carried out chromatin immunoprecipitation (ChIP) studies, first in Hela cells, to determine whether NSpt5 impacted the occupancy of RNAPII on HIVLTR chromatin in the absence of Tat (Figure 4A).To investigate promoter proximal and distal transcription, we used anti-RNAPII antibodies and PCR primers specific for the HIVLTR and CAT coding sequences (Figure 4A, top panel).F-NSpt5, but not F-Spt5, significantly increased the presence of RNAPII on the HIVLTR and CAT coding regions (Figure 4A, compare bars 2 and 5 to bars 3 and 6).Taken together, these results suggest that NSpt5 significantly enhances RNAPII occupancy on the HIVLTR in the absence of the Tat trans-activator.
We next carried out ChIP studies to examine the occupancy of RNAPII on the hsp70-4 chromatin in zebrafish embryos (Figure 4B).The hsp70-4 locus in zebrafish is ,3.4 kb in length, and contains two exons and one intron (Figure 4C).We first determined the patterns of RNAPII on hsp70-4 in WT embryos with or without heat shock.In WT embryos without heat shock, RNAPII (Figure 4D) exhibited a higher occupancy at the 59 than the 39 end, which is similar to the occupancy pattern of RNAPII on the hsp70 locus in Drosophila, indicating that there is an elongation block at the promoter proximal region.Upon heat stimulus of various lengths of time, occupancies of RNAPII (Figure 4D) are significantly increased at the 59end and moderately increased at the downstream region, which is consistent with the observed persistence of elongation pause under heat shock condition in Drosophila [44].In Nspt5-expressing embryos, the increased occupancy pattern of RNAPII was similar to the pattern observed under heat shock, although the occupancy level is lower than that with heat shock (Figure 4E).In spt5expressing embryos, RNAPII occupancy was not increased (Figure 4E).Together, these results suggest that NSpt5 derepresses the expression of HIVLTR and hsp70-4 in the noninduced conditions, through increasing RNAPII occupancy at these loci.

NSpt5 directly interacts with the hsp70-4 chromatin in vivo
To determine whether NSpt5 directly or indirectly causes the enhanced RNAPII occupancy on the hsp70-4 chromatin in vivo, we examined the chromatin occupancy of NSpt5, in comparison with Spt5.ChIP was carried out using the anti-FLAG antibody in F-spt5or FNspt5expressing embryos.F-NSpt5 was found to associate with the hsp70-4 chromatin (Figure 5).However, F-Spt5 did not exhibit any detectable association with the hsp70-4 chromatin (Figure 5).This observation suggests that NSpt5 directly interacts with the hsp70-4 chromatin to cause the increased RNAPII occupancy at the locus.The fact that the exogenously provided Spt5 has no association with the hsp70-4 chromatin suggests that NSpt5 has better access to the hsp70-4 chromatin than the full length Spt5.

NSpt5 increases the occupancy of Cdk9, histone marks for active transcription, and NELF-A at the hsp70-4 promoter region in vivo
We next determined the in vivo chromatin occupancy of other proteins involved in regulating transcription elongation at the hsp70-4 locus in vivo.Cdk9 is the kinase subunit of P-TEFb.In the presence of NSpt5, the occupancy of Cdk9 is significantly increased at the 59 end, middle, and 39 end (Figure 6A).Inhibition of Cdk9 activity, either through a specific morpholino antisense oligonucleotide or using the kinase inhibitor flavoperidol (FP), significantly suppressed NSpt5-induced dorsalization phenotype and the de-repression of hsp70-4 (Figure S4 and S5).Several lines of evidence suggest that such a suppression of NSpt5 effects is not due to a general developmental delay but rather, is likely due to a direct requirement of CDK9 in the manifestation of NSpt5's dominant negative effect: First, we checked later stages of Nspt5expressing and Cdk9-impaired embryos, and still did not see a dorsalization phenotype, suggesting that it is not due to a simple developmental delay.Second, at the molecular level, NSpt5mediated, increased expression of hsp70-4, which is a marker gene directly regulated by Spt5, is significantly impaired by Cdk9 The elongation stimulation activity of DSIF was assayed using pSLG402 as a template, which generates short (promoterproximal) and long (promoter-distal, dependent on the elongation stimulatory activity of DSIF) RNase T1-resistant products under the control of the adenovirus major-late promoter.Transcription initiation/ elongation was allowed to proceed for the indicated times.Timedependent increase of distal transcripts is observed in the control (first three lanes), while NSpt5 slightly enhanced transcription at 2X or 4X concentration but not at 8X concentration.(B) pTF3-6C2AT, which generates a 380-nt RNase T1-resistant product under the control of the adenovirus E4 promoter, was used as a template, and transcription was allowed to proceed for 10 minutes.This product is sensitive to the elongation repressive activity of DSIF (in the presence of the P-TEFb inhibitor DRB)(first two lanes).NSpt5 inhibits the repression activity of endogenous DSIF at 8X concentration.doi:10.1371/journal.pone.0006918.g003knockdown.Third, Cdk9, being a subunit of P-TEFb, is known to phosphorylate RNAPII CTD, Spt5 C-terminal domains, and NELF, which together form protein complexes, suggesting a possible direct requirement of RNAPII or NELF phosphorylation for the manisfestation of NSpt5's effect.Taken together, we suggest that Cdk9 activity and hence the phosphorylation of RNAPII or NELF is likely to be directly required for NSpt5's dominant negative effect in vivo.Consistent with the increased occupancy of Cdk9, we also observed increased occupancy of H3k4Me3 (Figure 6B), and a slight increase in occupancy of H3K79Me2 (Figure 6C), both of which represent histone marks for active transcription [45].
DSIF collaborates with RNAPII and NELF complex to stall transcription elongation in vitro [13].Since upon heat shock induction, DSIF and RNAPII but not NELF are strongly recruited to chromosomal puffs harboring the hsp70 genes [46], NELF has been viewed to have only a repressive role on RNAPII elongation.To determine the occupancy of NELF on hsp70-4 chromatin in the presence of NSpt5, we carried out CHIP with the antibody recognizing NELF-A, a subunit of NELF (antibodies against other NELF subunits do not work in zebrafish).In control embryos, NELF-A was detected mainly at the 59 end of hsp70-4 chromatin, consistent with the role of NELF in the promoter proximal pause of hsp70 (Figure 6D).However strikingly, in NSpt5-expressing embryos, NELF-A was dramatically and exclusively increased at the 59end of hsp70-4 (Figure 6D).This surprising result suggests that NSpt5 promotes the recruitment of NELF-A.Thus, NELF appears to behave very differently in transcriptional increases mediated by NSpt5 (this study) versus that mediated by heat shock [46], which is consistent with the significant difference of the expression level of hsp70-4 under these two conditions (see Figure 2G and H).

Discussion
Accumulating biochemical evidence suggests that the evolutionarily conserved protein Spt5 plays a critical role in regulating RNAPII elongation, by acting as a repressor or an activator under different in vitro assay conditions [12].The conversion of Spt5 from a repressor to an activator involves P-TEFb, which phsosphorylate RNAPII CTD [20,37] as well as the CTR1 domain of Spt5 [15][16][17][18][19], leading to the hypothesis that CTR1 acts as a mini-CTD for assembling active elongation complexes [19].Despite these findings, regulation of RNAPII elongation by Spt5 in vivo is not well understood.
In this study, we demonstrate that CSpt5 composed of partially redundant CTR1 and CTR2CT domains is a repressive module, by showing that NSpt5 lacking this module exerts a dominant negative effect on the repressive but not the stimulatory activity of the endogenous Spt5, while has no stimulatory activity on its own.First, NSpt5 has no rescuing activity but instead dominantly impair the development of WT embryos, suggesting that NSpt5 is likely to be a dominant negative form.Second, the increase of hsp70 expression in NSpt5-expressing WT is much lower compared to heat shock induced hsp70 expression, suggesting that Nspt5 ''de-represses'' hsp70 due to dominant negative interference with the endogenous Spt5's repressive activity, while has no elongation stimulatory activity on its own.This is consistent with a previous study, which shows that phosphorylated Spt5 C-terminus is critical for elongation stimulatory activity [19].Third, biochemical analysis shows that NSpt5 can dominantly interfere with Spt5's repressive but not stimulatory activity.The fact that Nspt5 does not affect heat-induced hsp70 expression or TAT transactivation at HIVLTR provides in vivo evidence that NSpt5 does not affect the stimulatory activity of endogenous Spt5.Finally, consistent with the dominant nature, in vivo ChIP analysis shows that NSpt5 has a preferential access to chromatins than WT Spt5.Because mis-expression of NSpt5 significantly impairs embryogenesis, this observation suggests that other important developmental genes are also de-repressed by NSpt5 possibly via similar mechanisms uncovered for hsp70-4 and HIVLTR.
Taken previous studies and our new findings into consideration, we propose the following model to explain the role of Spt5 in regulating RNAPII elongation in vivo (Figure 7). 1) In the uninduced state, Serine-5 phosphorylated RNAPII is stalled after the synthesis of a short stretch of RNA, via an Spt5-dependent mechanism.It is conceivable that low amount of P-TEFb may be present near the locus, either because of its constitutive presence in the nucleoplasm, or due to ''spill-over'' from neighboring, actively transcribing loci in vivo.We propose that CSpt5 may play an important role to prevent the RNAPII CTD from being phosphorylated by P-TEFb, thereby repressing RNAPII elongation and at the same time allowing critical processes such as mRNA capping to occur.2) The incorporation of mis-expressed NSpt5 into the stalled RNAPII complex removes CSpt5's repressive activity, thereby allowing Serine-2 phosphorylation by P-TEFb on the RNAPII CTD, hence transcription elongation.Alternatively, since NSpt5 has a stretch of acidic residues from aa. 3 to aa. 105, it may stimulate RNAPII transcription, through recruiting P-TEFb either directly or indirectly to the RNAPII complex.However, it is important to note that the increased transcription by NSpt5 appears much less than that in the induced state.3) Upon induction mediated by sequence-specific DNA binding proteins, a much larger amount of P-TEFb is recruited to the locus, which can phosphorylate CSpt5 (and NELF) to remove the repressive effect, which in turn allows the phosphorylation of RNAPII CTD, thereby allowing assembly of additional accessory factors for productive elongation.Taken together, our findings reveal a previously unknown role of CSpt5 in repressing RNAPII elongation in vivo.It remains possible that the phosphorylated CSpt5 may have a role in positively regulating transcription elongation through recruiting active elongation complexes.Future experiments are needed to test this and further verify the validity of this model.
We have made an interesting observation that enhanced NELF-A presence coincides with increased transcription caused by NSpt5.Since NELF has been considered to inhibit transcription elongation based on biochemical studies in vitro [13,[47][48][49] and the colocalization of NELF with an elongation incompetent form of RNAPII on polytene chromosomes [46], it is an unexpected finding.However, NELF is recently found to be critical for enhancing gene expression by blocking promoter proximal nucleosome assembly [50], and is also found to associate with many highly expressed genes [51].Our finding of the increased enrichment of NELF-A at the promoter region of hsp70-4 chromatin in Nspt5-expressing embryos thus supports such positive role of NELF in transcription elongation.
CDK9 Morpholino (MO) antisense oligonucleotide (Gene Tools, Corvallis OR) was designed to complement the exon 2/intron 2 junction.The MO sequence was: ACATCAAATACTCACC-CAAAGGTGC.1-2 nl of the morpholino oligonucleotide was injected at a concentration of 1.25 mM.

Immunohistochemistry
Immunohistochemistry was performed as previously described [54].The following primary antibodies were used: rabbit polyclonal anti-GFP (Abcam), mouse anti-Flag M2 antibody (Sigma).The following secondary antibodies were used: anti-rabbit IgG Alexa Fluor 488 and anti-mouse Alexa Fluor 568 (Molecular Probes).Differential interference contrast microscopy was performed on a Zeiss Axiophot 2 microscope and fluorescence microscopy was performed using a Leica TCS SP2 confocal microscope.

Chromatin immunoprecipitation (ChIP) in zebrafish embryos
De-chorionated embryos (about 150 at 9 hpf) were collected and cross-linked with 1% formaldehyde for 20 min at room temperature and quenched by addition of glycine to 0.125 M final concentration for 5 min.After being washed twice with PBS, the embryos were snap-frozen and stored at 270uC.The embryos were homogenized and resuspended in 300 ml lysis buffer (10 mM Tris-HCl pH 8.0, 200 mM NaCl, 10 mM EDTA, 0.5% NP-40, 0.1% SDS and protease inhibitor cocktail from Roche).The suspension was then sonicated on ice to generate approximately 500 base-pair (bp) fragments.The lysates were centrifuged, precleared with protein-A or protein-G agarose beads (Sigma), and then divided into 0.15 ml aliquots per immunoprecipitation (5% of the lysate was kept as 'input' before the addition of the antibody).Antibodies for immunoprecipitation were as follows: mouse anti-RNA Pol II (8WG16, Covance Research Products); mouse anti-Flag M2 antibody (Sigma), and rabbit anti-Spt5 polyclonal antibody raised against the zebrafish C-terminal region (CTR1CTR2CT).After addition of the antibody, lysates were incubated at 4uC overnight, and then incubated with protein-A or protein-G agarose beads (30 ml resin) for 1 hr at 4uC.The beads were washed once with DNA wash buffer (10 mM Tris-HCl pH 8.0, 200 mM NaCl, 10 mM EDTA, 0.5% NP-40), once with DNA wash buffer containing 100 mg/ml salmon sperm DNA (Invitrogen), once with 56RIPA buffer (10 mM Tris-HCl pH 8.0, 500 mM NaCl, 1 mM EDTA, 0.5 mM EGTA, 0.1% SDS, 1% Triton, 0.1% Sodium deoxycholate), and once with LiCl buffer (10 mM Tris-HCl pH 8.0, 250 mM LiCl, 1 mM EDTA, 0.5 mM EGTA, 1% Triton, 1% sodium deoxycholate).Finally, the beads were washed once with DNA wash buffer and pelleted, and chromatin was eluted from the beads by adding 160 ml elution buffer (25 mM Tris-HCl pH 7.5, 10 mM EDTA and 0.5% SDS) and incubated at 65uC for 20 min, then eluted with 100 ml elution buffer again.The inputs were also added with the elution buffer to 260 ml.After addition of equal volume digestion buffer (50 mM Tris-HCl pH 8.0, 1 mM EDTA, 100 mM NaCl and 0.5% SDS, 0.09 mg/ml proteinase K and 0.1 mg/ml RNase A), all samples were incubated at 42uC for 1 hr and 65uC overnight to reverse cross-linking.Chromatin DNA was purified by phenol extraction followed by alcohol precipitation, and used for real-time PCR.

Analysis of hsp70-4 mRNA levels
The control or injected embryos were collected at 6 hpf and total RNA was extracted using Trizol reagent (Invitrogen) and treated with Turbo DNA-free DNase (Ambion).First-strand cDNA was reverse transcribed using oligo-dT primers and Superscript reverse transcriptase (Invitrogen).Real-time PCR amplifications of hsp70-4 and b-actin were carried out with SYBR Green PCR Master Mix (Applied Biosystems).Primers used were Actin-forward (F), 59-TGAGCGCAAATACTCCGTCTGGAT-39, Actin-reverse (R), 59-GTTCGAGAGTTTAGGTTGGTCGTTCG-39 and hsp70-4 39end primers that were used in the ChIP experiment.

Transfections and CAT assays
HeLa cells were co-transfected with of pEF-RD (0.5 mg) and/or pCMV-SPT5 (wild-type or mutant version, 0.5 mg) and pHIV-CAT (0.1 mg), in the absence or presence of pTat (0.1 mg) using Fugene 6 according to the manufacturer's instruction (Roche).48 hours after the transfection, the activity of chloramphenicol acetyltransferase (CAT) was measured in the cell lysate by using a Lumitech ReportaLight Bioassay kit (Cambrex Bioscience).

ChIP in Hela cells
ChIP was carried out essentially as described previously [55].Primers used for LTR were described previously [56] and primers for CAT are: forward-59-atcccaatggcatcgtaaag-39; reverse-59tcgtcgtggtattcactcca-39.Standard curves for each primer pair were obtained first to determine their amplification efficiencies.Products were quantified using Brilliant SYBR Green qPCR according to manufacturer's directions (Stratagene).Relative enrichment was calculated and normalized to the input.
Additional methods can be found online as ''Supplementary Methods S1''

Figure 3 .
Figure 3. NSpt5 interferes with the repressive but not the stimulatory activity of endogenous DSIF in a dominant manner in vitro.(A)The elongation stimulation activity of DSIF was assayed using pSLG402 as a template, which generates short (promoterproximal) and long (promoter-distal, dependent on the elongation stimulatory activity of DSIF) RNase T1-resistant products under the control of the adenovirus major-late promoter.Transcription initiation/ elongation was allowed to proceed for the indicated times.Timedependent increase of distal transcripts is observed in the control (first three lanes), while NSpt5 slightly enhanced transcription at 2X or 4X concentration but not at 8X concentration.(B) pTF3-6C2AT, which generates a 380-nt RNase T1-resistant product under the control of the adenovirus E4 promoter, was used as a template, and transcription was allowed to proceed for 10 minutes.This product is sensitive to the elongation repressive activity of DSIF (in the presence of the P-TEFb inhibitor DRB)(first two lanes).NSpt5 inhibits the repression activity of endogenous DSIF at 8X concentration.doi:10.1371/journal.pone.0006918.g003

Figure 4 .
Figure 4. RNAPII elongates on the HIVLTR and hsp70-4 chromatin in the presence of F-NSpt5.(A) ChIP and qRT-PCR were performed with RNAPII antibodies and indicated primers (arrows) in Hela cells.Data are presented as percent of input material immunoprecipitated with specific antibodies over those with the IgG control.Error bars represent S.E.M. of triplicate measurements from three independent experiments.(B) A scheme of the in vivo ChIP analysis in zebrafish.(C) The gene structure of hsp70-4, highlighting the location of the hsp70-4 primer sets that are used during real-time PCR amplification of the immuoprecipitated material.The long arrow indicates the start of transcription.(D) The relative enrichment of RNAPII on hsp70-4 with RNAPII antibody over the IgG control under heat shock condition.(E) The relative enrichment of RNAPII on hsp70-4 in NSpt5 expressing embryos.Error bars represent S.E.M. of duplicate measurements from two independent experiments.doi:10.1371/journal.pone.0006918.g004

Figure 5 .Figure 6 .
Figure 5. NSpt5 directly interacts with the hsp70-4 chromatin in vivo.Charts show the percent of input material immunoprecipitated in different regions of hsp70-4 chromatin.The relative enrichment of ChIP and qRT-PCR values obtained with Flag antibody over the IgG control.Error bars represent S.E.M. of duplicate measurements from two independent experiments.doi:10.1371/journal.pone.0006918.g005