To determine if dPcf11 was involved in transcription termination in Drosophila cells, a combination of RNA interference (RNAi) and chromatin immunoprecipitation (ChIP) was used. ChIP revealed that dPcf11 was concentrated at the 3′ end of hsp70. All five copies of hsp70 are nearly identical and induced by heat shock. The presence of dPcf11 at the 3′ end of hsp70 was revealed by the marked enrichment of the +2433 to +2601 region over a control region in the dPcf11 immunoprecipitate. In addition, lower levels of dPcf11 were detected near the 5′ end of the gene and in a region preceding the polyadenylation site but appeared to be absent from the region in the middle of the gene. Treatment of cells with RNAi targeting dPcf11 mRNA reduced the level of dPcf11 detected on the hsp70 gene (Zhang, 2006).

The distribution of Pol II on hsp70 was monitored by ChIP using antibody against the Rpb3 subunit (Adelman, 2005). Although depletion of dPcf11 had no significant impact on the density of Pol II across the body of hsp70, it increased the level of Pol II crosslinking downstream from the polyadenylation signal by ~3-fold. This increase in Pol II downstream from the polyadenylation signal indicates that the efficiency of termination was impaired by depletion of dPcf11. This conclusion is further supported by RT-PCR, which detects RNA from the +2433 to +2601 region in cells depleted of dPcf11, but not in control cells (Zhang, 2006).

An EC was assembled with purified Drosophila Pol II and tested the effect of adding various derivatives of dPcf11 produced in E. coli. Pol II was allowed to initiate transcription on a tailed, G less cassette template (Zhang, 2004). In the presence of the UpG dinucleotide, ATP, CTP, and radioactive UTP, Pol II initiates transcription at the tailed end of the template and pauses at the end of the G less cassette. These ECs remain active, as they will resume transcription upon addition of GTP (Zhang, 2004). Also, the nascent transcript is sensitive to RNase A, but not RNase H, indicating that the nascent transcript has not hybridized to the template strand of DNA along its entire length, as sometimes occurs with other tailed templates (Zhang, 2004; Zhang, 2006).

When the ECs were analyzed on a native gel, two bands were detected. The upper band contains Pol IIA with an intact CTD; the lower band contains Pol IIB that has lost its CTD during the purification of the Pol II (Zhang, 2004). The Pol IIB EC provides an internal control for monitoring CTD-dependent disruption of the EC. Treatment of the ECs with full-length dPcf11 (1–574) selectively eliminated the Pol IIA EC. The same thing occurred when the ECs were treated with a CTD antibody. By testing various derivatives of dPcf11, dismantling activity was mapped to the N terminus encompassed by amino acids 1–149. This region is homologous to the region of yPcf11 recently shown to dismantle a yeast EC (Zhang, 2005; Zhang, 2006).

The disappearance of the Pol IIA complex is not simply the consequence of binding protein to the CTD. CstF50 is a subunit of the polyadenylation factor CstF and has been shown to associate with the CTD. In contrast to dPcf11, treatment of the ECs with dCstF50 shift the mobility of the Pol IIA EC (Zhang, 2006).

To further analyze the activity of dPcf11, the effects of dPcf11 were examined on a mixture of Pol IIA and IIB ECs that were assembled on immobilized DNA templates. dPcf11 derivatives encompassing amino acids 1–149, 1–283, and 1–574 (full length) cause the immobilized complexes to release half of the transcript. No transcripts are released when Pol IIB ECs were challenged with dPcf11, so the portion of transcripts released from the Pol IIA/IIB mixture must be derived from Pol IIA ECs. In accord with this, Western blotting revealed that dPcf11 1–283 dissociate Pol IIA, but not Pol IIB, from the immobilized template. In agreement with the gel shift data, dCstF50 failed to dissociate either form of Pol II from the template. Taken together, the results demonstrate that the N-terminal region of dPcf11 completely dismantles the Pol IIA EC by a mechanism that requires the CTD but no nucleotides (Zhang, 2006).

Recently, it was discovered that yPcf11 can form a bridge between the CTD and RNA and that this bridge appears to be important for the dismantling reaction (Zhang, 2005). To determine if dPcf11 dismantles the EC by a similar mechanism, various derivatives of dPcf11 were tested for CTD binding activity. All the derivatives that were found to dismantle the EC were also found to associate with the CTD (Zhang, 2006).

dPcf11 1–283 was tested for RNA binding activity (this derivative was used because its yield from E. coli was significantly higher than the others). UV crosslinking was used to detect protein-RNA interactions. Pcf11, in the presence or absence of GST-CTD, was incubated with radiolabeled RNA and then UV irradiated to induce crosslinks between RNA and associated proteins. The RNA was degraded with RNase, and proteins crosslinked to the residual RNA tags were identified after SDS-PAGE. dPcf11 crosslinked to RNA. mPcf11, a mutant form of dPcf11 found to completely inhibit the dismantling activity of dPcf11, also failed to crosslink to RNA. This mutation has amino acids 75, 76, and 77 mutated to alanines and corresponds to a mutation in yPcf11 that impairs termination (Sadowski, 2003; Zhang, 2005; Zhang, 2006).

Because dPcf11 associates with both the CTD and RNA, tests were performed to see if dPcf11 bridges the CTD to RNA. GST-CTD was immobilized and then incubated with radiolabeled RNA in the presence or absence of dPcf11. In the absence of dPcf11 or the presence of mPcf11 1–238, no RNA associates with the CTD. In the presence of dPcf11, ~20% of the input RNA was observed to bind to the immobilized CTD. These results indicate that dPcf11 can bridge the CTD to RNA (Zhang, 2006).

A monoclonal antibody that binds specifically to the CTD of Pol II dismantles the EC. If the formation of a bridge between the CTD and RNA was essential to the dismantling reaction, it is predicted that the CTD antibody should associate with RNA. UV crosslinking showed that this was indeed the case. RNA crosslinks specifically to the light chain of the antibody with an efficiency similar to that of dPcf11. The absence of crosslinking to the heavy chain indicates that the method detects a specific association with the light chain rather than simply the random collision of molecules (Zhang, 2006).

CTD antibody bridges RNA to the immobilized form of the CTD. Because dPcf11 and the CTD antibody can both bridge the CTD to RNA, yet their structures are unrelated, it is concluded that a crucial step in the dismantling reaction is formation of a bridge between the CTD and RNA. This conclusion is further supported by the observation that another monoclonal antibody that binds RNA, but associates with the body of Pol II, fails to dismantle the EC (Zhang, 2004; Zhang, 2006).

Unlike the CTDs of yeast and human Pol II in which 19 of 26 and 21 of 52 of the respective heptad repeats exactly match the consensus YSPTSPS, only two of the 42 repeats in the Drosophila CTD exactly match the consensus. Hence, it was of interest in determining where dPcf11 associates with the CTD, since this might provide insight into the dismantling mechanism. Pull-down assays with immobilized derivatives of the Drosophila CTD fused to GST reveal that dPcf11 preferentially associates with the region encompassing the two consensus heptads. dPcf11 was observed to bind equally well to an intact version of the CTD and to a truncated version of the CTD containing the two conserved heptads but bind markedly less well to parts of the CTD that flanked the conserved heptads (Zhang, 2006).

The CTD antibody inhibits transcription at low, but not high, nucleotide concentrations (Zhang, 2004). Low nucleotide concentrations promote pausing by the EC. As pausing appears to precede termination in vivo, tests were performed to see if there were any differences in the way dPcf11 affects elongation reactions at two nucleotide concentrations. dPcf11 inhibits elongation by Pol IIA at low, but not high, nucleotide concentrations. Fifty percent inhibition of the Pol IIA/IIB mixture occurs when transcription is done at a low nucleotide concentration in the presence of dPcf11 1–238, whereas no inhibition occurs at a high nucleotide concentration. Neither mPcf11 nor dPcf11 151–574 affect transcription at either nucleotide concentration. Also, no inhibition of transcription is observed at either nucleotide concentration when transcription is performed with purified Pol IIB. Together, these results indicate that dPcf11 inhibits transcription elongation by Pol IIA under conditions that promote pausing (Zhang, 2006).