Drosophila RNA polymerase II mutants that affect transcription elongation.

We have examined the properties of two Drosophila RNA polymerase II mutants, C4 and S1, during elongation, pyrophosphorolysis, and DmS-II-stimulated transcript cleavage. The C4 and S1 mutants contain a single amino acid substitution in the largest and second largest subunits, respectively. Compared with wild type, C4 had a lower elongation rate and was less efficient at reading through intrinsic elongation blocks. S1 had a higher elongation rate than wild type and was more efficient at reading through the same blocks. During elongation, C4 and wild type responded similarly to DmS-II and NH4+ whereas the S1 mutant was less responsive to both. Differences between the two mutants also appeared during DmS-II-mediated transcript cleavage and pyrophosphorolysis. During extended pyrophosphorolysis, S1 polymerase was fastest and C4 polymerase was slowest at generating the final pattern of shortened transcripts. S1 and wild type were equal in the rate of extended DmS-II mediated transcript cleavage, and C4 was slower. Our results suggest that the S1 mutation increases the time spent by the polymerase in elongation competent mode and that the C4 mutation may affect the movement of the polymerase.

We have examined the properties of two Drosophila RNA polymerase II mutants, C4 and S1, during elongation, pyrophosphorolysis, and DmS-II-stimulated transcript cleavage. The C4 and S1 mutants contain a single amino acid substitution in the largest and second largest subunits, respectively. Compared with wild type, C4 had a lower elongation rate and was less efficient at reading through intrinsic elongation blocks. S1 had a higher elongation rate than wild type and was more efficient at reading through the same blocks. During elongation, C4 and wild type responded similarly to DmS-II and NH 4 ϩ , whereas the S1 mutant was less responsive to both. Differences between the two mutants also appeared during DmS-II-mediated transcript cleavage and pyrophosphorolysis. During extended pyrophosphorolysis, S1 polymerase was fastest and C4 polymerase was slowest at generating the final pattern of shortened transcripts. S1 and wild type were equal in the rate of extended DmS-II-mediated transcript cleavage, and C4 was slower. Our results suggest that the S1 mutation increases the time spent by the polymerase in elongation competent mode and that the C4 mutation may affect the movement of the polymerase.
Recent studies have revealed that the control of transcription elongation is an important target for the regulation of eukaryotic gene expression (1,2). As the central component of the mRNA transcription machinery, RNA polymerase II itself is the final receptor of various kinds of interactions that activate or inactivate transcription elongation, and it carries out the fundamental catalysis of RNA chain elongation. However, functional roles played by its individual subunits during the process of elongation have not been thoroughly illuminated.
A number of mutations in RpII215 and RpII140, the genes encoding the two large subunits of Drosophila RNA polymerase II have been mapped at the DNA sequence level. Among them, the C4 mutation in RpII215 changes amino acid 741 from Arg to His in the largest subunit (3), and the S1 mutation in RpII140 changes amino acid 728 from Ser to Cys in the second largest subunit (4). In vivo, the C4 mutation induces ␣-amanitin resistance and the Ubx effect (5,6), whereas the S1 mutation suppresses the temperature-sensitive mutant phenotype caused by another mutation (WJK1) in the largest subunit (7). Little is known about what functions of RNA polymerase II are altered and how transcription is affected by these and other mapped mutations, except that it was shown previously that the C4 mutant enzyme is resistant to ␣-amanitin and is slower in elongation in vitro (8,9). Having previously mapped muta-tions in both RpII215 and RpII140, we next initiated biochemical studies of a subset of the mutant enzymes that are amenable to purification (10). In our studies, we found that as for the C4 mutant enzyme, the S1 mutant enzyme is also different from wild type in elongation. Because the C4 mutation alters the largest subunit, whereas the S1 mutation alters the second largest subunit of Drosophila RNA polymerase II, the availability of C4 and S1 RNA polymerase II mutant enzymes provides an opportunity possibly to identify functional roles for the two large subunits during elongation and to dissect the molecular mechanism by which RNA polymerase II achieves efficient elongation.
Using cell-free transcription systems in which elongation complexes can be assembled, a number of factors have been identified that affect elongation properties of RNA polymerase II. For example, the Drosophila transcription factor, DmS-II, which was initially purified from Drosophila Kc cell nuclear extracts, has been shown to have a stimulatory effect on elongation in vitro (11,12). DmS-II is a 36-kDa protein that is the Drosophila counterpart for mammalian S-II (TFIIS); the gene encoding DmS-II has been cloned and sequenced (13). DmS-II has been shown to reduce the time spent by RNA polymerase II at a subset of the numerous pause sites encountered on a dC-tailed 1 template, but it does not stably bind to the elongation complex (11). DmS-II has also been shown to be necessary and sufficient to activate nascent transcript cleavage by Drosophila RNA polymerase II during transcription of a dC-tailed template, and the C-terminal half of DmS-II is required for its cleavage-activating function. A mechanism for pause suppression by DmS-II has been proposed. DmS-II binds to the paused polymerase, causes one cleavage event, and is then released from the complex. Elongation by the polymerase then allows a second encounter with the pause site and a second chance of passing the site. Complete pause suppression may require multiple transcript shortening events for some polymerase molecules (14).
In the study presented here, we investigate several elongation-related properties of wild type and mutant RNA polymerase II, namely recognition of intrinsic blocks to elongation, read-through in response to DmS-II, DmS-II-stimulated cleavage of nascent transcripts, and pyrophosphorolysis. We demonstrate that the C4 and S1 mutations affect different functional processes during elongation, and we discuss the implications of our observations.

EXPERIMENTAL PROCEDURES
Materials-Polymin P was purchased from ICN Biochemicals. Lyophilized ribonucleoside triphosphates were purchased from Pharmacia Biotech Inc. ␣-Amanitin, terminal deoxynucleotidyl transferase, and restriction enzymes were purchased from Boehringer Mannheim. [5Ј,6Ј-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 3 H]UTP, [␥-32 P]ATP, and [␣-32 P]CTP were purchased from ICN Radiochemicals. Miracloth was purchased from Chicopee Mills. DNA size standards 50 -2000-base pair ladder, 5.0-ml Econo-Pac heparin column, and Bio-Spin 30 columns were purchased from Bio-Rad. DEAE cellulose (DE52 pre-swollen cellulose) was from Whatman. 1.0-ml Mono-Q fast protein liquid chromatography column was from Pharmacia. All other materials were reagent grade.
All purification steps were carried out at 0 -4°C. Frozen dechorionated embryos (50 g) were suspended in 100 ml of embryo buffer and homogenized using a Waring blender for 45 s at low speed and 30 s at high speed. Following the addition of 100 ml of HGE, homogenization was continued for 45 s each at low and high speed. The extract was centrifuged at 10,000 rpm (16,000 ϫ g) for 30 min in a GSA rotor (Sorvall), and the resulting supernatant was collected after being filtered through one layer of Miracloth. One hundredth volume of a 10% (w/v) solution of Polymin P was stirred in dropwise. After 20 min of gentle stirring, the precipitated material was collected by centrifugation in the GSA rotor at 8,500 rpm (12,000 ϫ g) for 20 min. The polymin P precipitate was washed in 50 ml of 0.1 M HGKE using a glass-glass Dounce homogenizer to obtain uniform resuspension, followed by recentrifugation at 8,000 rpm (10,000 ϫ g) for 20 min in the GSA rotor. RNA polymerase II activity was then extracted from the pellet into 50 ml of 0.3 M HGAE using a Dounce homogenizer, followed by centrifugation at 13,000 rpm (25,000 ϫ g) for 20 min in the SA-600 rotor (Sorvall). The resulting polymin P eluate was filtered through one layer of Miracloth. Solid (NH 4 ) 2 SO 4 (0.3 g/ml of polymin P eluate) was added and dissolved in by gentle stirring over 30 -40 min. The precipitated protein was collected by centrifugation at 40,000 rpm (140,000 ϫ g) for 30 min in a T-865 ultracentrifuge rotor (Sorvall). The pellet was resuspended in 10 -15 ml of HGE with a Dounce homogenizer to yield ammonium sulfate enzyme, which was frozen in liquid nitrogen and stored at Ϫ80°C. The thawed ammonium sulfate enzyme was loaded onto a 30-ml DE-52 column equilibrated in 0.12 M HGAE and was eluted with a 0.3 M HGAE step. The peak of RNA polymerase II activity was pooled and was loaded onto a 5.0 ml Econo-Pac heparin column equilibrated in 0.18 M HGAE, and was eluted with a 0.5 M HGAE step. The RNA polymerase II activity peak was pooled and loaded onto a 1.0 ml Mono Q fast protein liquid chromatography column equilibrated in 0.15 M HGKE. After extensive washing of the column with 0.15 M HGKE, the enzyme was eluted with a 22.5 ml 0.15 M to 0.6 M HGKE gradient. RNA polymerase II activity eluted around 0.45 M KCl and was collected, frozen in liquid nitrogen and stored at Ϫ80°C. RNA polymerase II activity (in units) was determined for each enzyme preparation under standard assay conditions (16). Note that under these conditions wild type and S1 enzymes have almost identical specific activities and that the C4 enzyme's specific activity is approximately 40% lower (e.g. Ref. 9 and "Results"). Full-length DmS-II was expressed and purified as described in Guo and Price (14).
dC-tailed Templates-The plasmid pPCP was constructed by subcloning a EcoRI-BamHI fragment containing the Drosophila actin 5C promoter into the pSP73 vector (Promega). The 3.6-kilobase pair pPCP was linearized by PstI, and dC tails were added to the 3Ј ends with terminal deoxynucleotidyl transferase, 1 mM CoCl 2 , and 1 mM dCTP at 37°C for 45 min. The plasmid pGEMTerm (obtained from Dr. Caroline Kane, University of California, Berkeley) contains a TaqI fragment of human histone H3.3 gene (17) subcloned in the AccI site of the pGEM-2 vector. pGEMTerm was linearized with SmaI, dC-tailed at 37°C for 2 h, and then digested by PstI and EcoRI. Immobilized dC-3025 template was prepared as described (18).
Elongation Assay-A pulse-chase protocol was applied to study the elongation of RNA polymerase II from dC-pPCP. A 30-l pulse reaction contained 20 mM HEPES (pH 7.6), 1 mM DTT, 5 mM MgCl 2 , 600 uM GTP, UTP, and ATP, 30 uM CTP, 1.5 uM [␣-32 P]CTP, 200 g/ml bovine serum albumin, 10 g/ml dC-pPCP, KCl, or (NH 4 ) 2 SO 4 , and RNA polymerase II. RNA polymerase II was first preincubated with dC-pPCP for 5 min in chase buffer minus UTP. UTP was then added to allow elongation and incorporation of labeled CTP for 1 or 3 min as specified. Following the pulse period, 10 l of chase buffer (20 mM HEPES, pH 7.6, 1 mM DTT, 5 mM MgCl 2 , 600 uM GTP, UTP, and ATP, and 4.8 mM CTP) was added to raise the cold CTP concentration. The chase reaction was continued for the indicated time and stopped by removing aliquots into 100 l of ice-cold sarkosyl solution (1% sarkosyl, 100 mM NaCl, 100 mM Tris, pH 8.0, 10 mM EDTA, and 250 g/ml tRNA). The solution was phenol extracted, and RNAs were ethanol precipitated.
Read-through Assay-A pulse-chase protocol was also applied to study the ability of RNA polymerase II to read through elongation blocks from dC-pGEMTerm. A 30-l pulse reaction contained 20 mM HEPES, pH 7.6, 1 mM DTT, 5 mM MgCl 2 , 600 uM GTP, UTP, and ATP, 1.0 uM [␣-32 P]CTP, 200 g/ml bovine serum albumin, 20 g/ml dC-pGEMTerm, 100 mM KCl, and 180 units of RNA polymerase II. RNA polymerase II was first preincubated in this mixture minus UTP and CTP for 5 min. [␣-32 P]CTP and UTP were then added together, and labeling was allowed to proceed for 2 min. Finally, 10 l of chase buffer (as described above for the elongation assay) was added. The chase reactions were continued in the presence of indicated reagents and stopped at the indicated times, and RNAs were recovered as described above.
Cleavage Assay and Pyrophosphorolysis-Two kinds of elongation complexes were employed. Elongation complexes formed on dC-pGEM-Term were isolated by Bio-Spin 30 columns. The columns were equilibrated by overlaying each column with 200 l of 1 ϫ reaction buffer (20 mM HEPES, pH 7.6, 1 mM DTT, and 5 mM MgCl 2 ) and spinning for 2 min in an International Equipment Corp. tabletop/clinical centrifuge with a swinging bucket rotor at a setting of 4 (approximately 1000 ϫ g). This procedure was repeated four times. Reactions were set up as described for the read-through assay. After 10 min of chase, the reaction mixtures were directly applied to the equilibrated columns and centrifuged for 4 min at a setting of 4. This was repeated with a second column.
Reactions containing immobilized dC-3025 were carried out in the following manner. In order to generate comparable amounts of the 13-mers before the elongation complexes were isolated, different unit amounts of P2, S1, and C4 enzymes were used. In a typical 15-l preincubation, 178 units of P2, 238 units of S1, and 400 units of Reactions were stopped at indicated times, and RNAs were recovered as described above.

Mutant RNA Polymerases II Show Abnormal Elongation
Rates in Vitro-To compare the elongation rates of mutant enzymes with wild type enzyme, we used a dC-tailed template that allows the study of elongation in the absence of a physiological promoter and initiation factors (19). The size distribution of radioactively labeled transcripts produced during a specific time period gives a measure of the elongation rate. Fig. 1A shows the RNA profiles for P2 and S1 enzymes when elongation assays were carried out with 100 mM KCl, Fig. 1B shows the RNA profiles when elongation assays were carried out with 50 mM KCl plus 80 mM (NH 4 ) 2 SO 4 , and Fig. 1C plots the maximum lengths of RNA produced by P2 or S1 at each time point under the two different salt conditions. We found that the S1 mutant enzyme was 2-fold faster in elongation than the P2 wild type enzyme when no NH 4 ϩ was present. Under the same assay conditions, however, the C4 mutant enzyme was 50% slower in elongation (data not shown). In the presence of NH 4 ϩ , the elongation rate of P2 was increased by about 2-fold, but the elongation rate of S1 was almost not changed. On the other hand, the elongation rate of C4 was also increased by about 2-fold in the presence of NH 4 ϩ (data not shown), although it remained 50% slower than P2 (9).
Mutant RNA Polymerases II Show Abnormal Ability to Read through Intrinsic Elongation Blocks-During in vitro transcription from various dC-tailed templates, RNA polymerase II pauses or is blocked at numerous sites along the template, as revealed by discrete bands in the RNA profile. Reines et al. (17) characterized three sequence specific sites (TIa, TIb, and TII) in the first intron of the human histone H3.3 gene that efficiently block elongation by mammalian RNA polymerase II. Later, Christie et al. (20) showed that these three sites also block the elongation of purified yeast RNA polymerase II. Because C4 and S1 mutant enzymes showed abnormal elongation rates when transcribing a dC-tailed template (Fig. 1), we used dC-pGEMTerm, which contains TIa, TIb, and TII, to investigate if the two mutations would affect RNA polymerase II in recognizing and reading through the characterized intrinsic elongation blocks.
We found that both wild type and mutant Drosophila RNA polymerases II could be blocked at TIa, TIb, and TII in vitro. As shown in Fig. 2A, after a 5-min chase, besides the run-off transcripts (RO), transcripts with 3Ј ends at TII, TIb, and TIa also accumulated for both wild type and mutant enzymes. However, in the absence of any elongation factors the mutant and wild type polymerases displayed differences in the relative amounts of enzyme reaching the end of the template (RO) versus the amount held up at the intrinsic elongation blockage sites (principally TIa and TIb). Thus, the ratios of RO/(RO ϩ TIa) and (RO ϩ TIa)/(RO ϩ TIa ϩ TIb) were different for wild type and mutant enzymes. As compared in Fig. 2B, after 5 min of chase, the ratios of RO/(RO ϩ TIa) were 34, 11, and 54% for P2, C4, and S1, respectively; after 60 min of chase, the ratios increased to 58, 49, and 62%, respectively. After 5 min of chase, the ratios of (RO ϩ TIa)/(RO ϩ TIa ϩ TIb) were 64, 53, and 71% for P2, C4, and S1, respectively; after 60 min of chase, the ratios increased to 82, 78, and 82%, respectively. In these experiments, the ratios of RO/(RO ϩ TIa) and (RO ϩ TIa)/(RO ϩ TIa ϩ TIb) reflect the half-life of pausing at TIa and TIb, respectively, but they are also influenced by the timing of the arrival of polymerases from other earlier pause sites and by the number of polymerases that become arrested (unable to proceed) at TIa and TIb. The different ratios clearly indicate a functional difference between the polymerases. Compared with P2 wild type, the C4 mutant enzyme reads less efficiently through TIa and TIb, whereas the S1 mutant enzyme reads more efficiently through the same blocks.
Mutant RNA Polymerases II Show Different Response to DmS-II-Mammalian TFIIS or the yeast TFIIS analog, P37, has been shown to stimulate mammalian or yeast RNA polymerase II to read through TIa, TIb, and TII (20,21). Because C4 and S1 mutants were different from wild type in reading through TIa and TIb when no elongation factors were present, we further investigated how they would respond to DmS-II. We found that DmS-II was able to promote Drosophila RNA polymerase II to read through intrinsic elongation blocks, but although P2 and C4 were stimulated by DmS-II to a similar degree, S1 was much less responsive to the action of DmS-II.
A time course of elongation is shown in Fig. 3A and the quantitation is shown in Fig. 3B. After 5 min of chase in the presence of 2.2 nM DmS-II (note that before DmS-II was added there was a 5-min initial chase in the absence of DmS-II), the ratio of RO/(RO ϩ TIa) for P2 was 79%, increasing from 45% after 10 min of chase in the absence of DmS-II ( Fig. 2A, 10-min point); this represents a 1.8-fold stimulation. Within the period of 60 min, the presence of DmS-II resulted in an average of 1.7-fold increase in read-through of TIa by P2 (n-min points in Fig. 3 versus n ϩ 5-min points in Fig. 2). Similarly, DmS-II resulted in an average of 1.7-fold increase in read-through of TIa by C4 , although compared with P2, the C4 mutant remained less efficient at reading through TIa. In contrast, the same amount of DmS-II only resulted in an average of 1.2-fold increase in read-through of TIa by S1 mutant enzyme (compare Figs. 3 and 2).
The experiment shown in Fig. 4 also examines the effect of DmS-II by comparing the RNA profile generated in its absence or presence. For P2 wild type, although the maximum elongation rate (determined from the leading edge of the transcript distribution at each time point) was not stimulated very much, the quantity of pulse-labeled transcripts was greater in the presence of DmS-II, indicating that DmS-II increased the number of P2 polymerase molecules that passed through early pause sites (compare lanes 5, 6, 7, and 8 with lanes 1, 2, 3, and  4). Although there was generally much more labeled RNA after an 8-min chase in the presence of DmS-II than in the absence of DmS-II, an RNA of about 140 nucleotides (Fig. 4, indicated by the arrow) diminished when DmS-II was present (compare lane 8 with lane 4), indicating pausing of P2 at this specific site was reduced by DmS-II. However, for the S1 mutant, there was not much increase in the amount of labeled RNA when DmS-II was present (compare lanes 13, 14, 15, and 16 with lanes 9, 10,  11, and 12), and there was not much decrease in the 140nucleotide RNA (compare lane 16 with lane 12), confirming that S1 was less responsive to the action of DmS-II.
Mutant RNA Polymerases II Act Differently in DmS-II-medi-ated Transcript Cleavage and Pyrophosphorolysis-Mammalian or yeast TFIIS has been shown to stimulate nascent transcript cleavage in elongation complexes stalled at the TIa site (20,22). This cleavage precedes and is necessary for efficient read-through. In order to determine how the two mutations would affect RNA polymerase II in DmS-II-mediated transcript cleavage, we initiated transcription from dC-pGEMTerm and isolated elongation complexes stalled at TIa, TIb, or TII using gel filtration spin columns. As shown in Fig. 5, for both wild type and mutant polymerases, when isolated elongation complexes were incubated with DmS-II, transcript shortening was observed and the patterns of cleaved transcripts were similar. When NTPs were added back to the elongation complexes that had been incubated with DmS-II for 60 min and had undergone extensive transcript cleavage, elongation of the shortened transcripts back to the TIa, TIb, or TII site was also observed (lanes 6, 12, and 18). These results indicate that the ternary elongation complex formed with either C4 or S1 mutant RNA polymerase II is still able to carry out DmS-II-mediated transcript cleavage and suggest that the shortened transcripts remain stably associated with the complex. Quantitation of the remaining transcripts with 3Ј ends at TIa revealed that after 10 min, 90% of the TIa transcripts in the P2 or C4 containing elongation complexes were cleaved, whereas only 57% were cleaved in the S1 containing complexes, consistent with S1 being different in interacting with DmS-II.
In order to look in more detail at DmS-II-mediated transcript cleavage and compare it with pyrophosphorolysis, we isolated elongation complexes stalled at a major elongation block after the incorporation of 13 nucleotides with immobilized dC-3025 template. When isolated elongation complexes were incubated with 1 mM sodium pyrophosphate or 1 nM DmS-II, transcript shortening from the 13-mer was observed for all three polymerases (Fig. 6). The pattern of the transcripts shortened by DmS-II mediated cleavage was similar to that shortened by pyrophosphorolysis, with the 13-mer decreasing and transcripts of 11, 9, 7, and 5 nucleotides appearing (14). The amount of run-off transcripts found with the three polymerases (compare lanes 1, 6, and 11 in Fig. 6A and lanes 1, 7, and 13 in Fig. 6B) indicates that S1 passes the pause site after incorporation of 13 nucleotides more easily than wild type, whereas C4 has more difficulty, consistent with findings obtained with the dC-pGEMTerm template.
In the presence of 1 mM pyrophosphate (Fig. 6A), 70 -75% of the 13-mers in the P2 or S1 complexes were shortened after 90 min of pyrophosphorolysis, whereas 90% were shortened in the C4 complexes. In reaching the final pattern of 9-mer, 7-mer, or 5-mer, S1 was fastest, and C4 was slowest (compare lanes 4, 9,  and 14 and compare lanes 5, 10, and 15). In the presence of 1 nM DmS-II (Fig. 6B), 90% of the 13-mers were shortened in the P2 or C4 complexes in 10 min, whereas 55% were shortened in the S1 complexes. In reaching the 9-mer, 7-mer, or 5-mer, S1 was similar to P2, but C4 was significantly slower (compare lanes 4,  10, and 16). These results confirm that relative to wild type and C4, S1 is different in its interaction with DmS-II. The results also suggest that C4 is altered in its ability to move backward along the template. DISCUSSION We have examined the properties of two Drosophila RNA polymerase II mutants in elongation, recognition of intrinsic elongation blocks, read-through in response to DmS-II, transcript cleavage in response to DmS-II, and pyrophosphorolysis. Our results show that the two mutations affect distinct functions associated with the two large subunits of RNA polymerase II. The C4 mutation in the largest subunit probably affects the translocation of the polymerase such that the forward and backward movement of the mutant elongation complex is slowed down. The S1 mutation in the second largest subunit probably affects the conformation of the polymerase and increases the time spent by the polymerase in the elongation competent mode such that the mutant enzyme is less responsive to the action of DmS-II and NH 4 ϩ . It has been shown previously that the C4 mutant polymerase is resistant to ␣-amanitin in RNA polymerization due to decreased binding affinity to the toxin (8). We found that in the presence of 1 g/ml of ␣-amanitin, the DmS-II-stimulated cleavage of nascent transcripts by wild type polymerase was not observed within 60 min, confirming the inhibitory effect of ␣-amanitin on this process (14,(23)(24)(25). In contrast, the cleavage by C4 mutant was still observed (data not shown), consistent with C4 not binding ␣-amanitin normally. Although the detailed inhibitory mechanism of ␣-amanitin is yet to be elucidated, recent studies suggest that ␣-amanitin inhibits pyrophosphorolysis and DmS-II-mediated transcript cleavage differently. For example, the toxin allows (slowed) pyrophosphorolysis within a paused elongation complex but completely blocks DmS-II action (26).
Because the response of C4 to DmS-II and NH 4 ϩ parallels that of the wild type, the structural features involved in interacting with DmS-II and NH 4 ϩ are probably not affected by the C4 mutation. Furthermore, because C4 carries out initial DmS-II-mediated transcript cleavage and pyrophosphorolysis as well as the wild type enzyme, the catalytic steps during these two processes may not be affected, either. It is possible that although the C4 mutation reduces the affinity of the enzyme for ␣-amanitin, it also introduces certain functional changes similar to those that would be caused by ␣-amanitin binding. One possibility is that the forward and backward translocation is slowed. This possibility is consistent with our observations that C4 is slower in reading through elongation blocks and is also slower in extended transcript shortening.
Three FIG. 5. Comparison of DmS-II-stimulated transcript cleavage from TIa by P2 wild type, C4 mutant, and S1 mutant RNA polymerase II. Elongation complexes were formed on dC-pGEMTerm and isolated by spin columns as described under "Experimental Procedures." Isolated complexes were then incubated with 4.5 nM DmS-II. Aliquots were removed after 0, 2, 5, 10, and 60 min of incubation. Finally, 600 uM GTP, 600 uM ATP, 100 uM CTP, and 450 uM UTP were added to the reaction, which had been incubated with DmS-II for 60 min, and the incubation was continued for another 10 min. RNA products were isolated and resolved by electrophoresis on a 5% polyacrylamide-6 M urea-TBE gel. The run-off transcripts and transcripts with 3Ј ends at TIa, TIb, and TII are indicated by RO, TIa, TIb, and TII. The major cleavage products are indicated by the arrows. PII 0 , RNAPII A , and RNAPII B , enzyme whose C-terminal repeated domain (CTD) is phosphorylated, not phosphorylated, or proteolyzed, respectively. With glycerol gradients and electrophoresis under native conditions, they found that TFIIS binds preferentially to II 0 and II A , suggesting the CTD is involved in binding TFIIS. However, TFIIS can stimulate elongation by RNA polymerase II in which the entire CTD has been proteolyzed during purification. Furthermore, Christie et al. (20) showed that for Saccharomyces cerevisiae RNA polymerase II the CTD and subunits four and seven, which are essential in vivo, are not required in vitro for read-through of intrinsic elongation blocks and nascent transcript cleavage in response to TFIIS. Rappaport et al. (28) showed that a fusion protein containing a portion of the largest subunit of human RNA polymerase II is able to inhibit the stimulatory effect of TFIIS. A monoclonal antibody against this fusion protein can also inhibit the stimulation by TFIIS. However, TFIIS binds only weakly to the fusion protein they used, suggesting the protein domain around conserved region B in the largest subunit may be only part of the TFIIS-binding site. Archambault et al. (29) isolated seven mutations in the gene encoding the largest subunit of S. cerevisiae RNA polymerase II (rpo21) that confer increased growth inhibition by the uracil analog, 6-azauracil, which is also a mutant phenotype associated with yeast mutants lacking TFIIS. The 6-azauracil-sensitive phenotype of RNA polymerase II mutants can be suppressed by overexpression of TFIIS, suggesting that the region identified by the 6-azauracil-sensitive rpo21 mutations, which is located between conserved regions G and H, may be involved in interacting with TFIIS.
Our results with S1 mutant enzyme demonstrate the involvement of the second largest subunit of RNA polymerase II in responding to the action of DmS-II. One simple explanation for the S1 mutant biochemical phenotypes would be that the S1 mutation, which results in changing Ser-728 to Cys, directly affects the binding of DmS-II. However, the observations that S1 mutant enzyme has a higher elongation rate and is more efficient at reading through elongation blocks in the absence of DmS-II suggest that if DmS-II binding is affected, it is not the only property altered. The stimulatory effect of NH 4 ϩ on tran-scription has been reported in several studies (10,30), and it was speculated that NH 4 ϩ stimulates transcription through an effect on the conformation of the polymerase. The observation that NH 4 ϩ has a much weaker effect on S1 mutant enzyme suggests that the polymerase conformation is probably altered by the S1 mutation.
Price et al. (12) proposed that during the elongation cycle the polymerase can be in a paused conformation or an elongationcompetent conformation, and the conversion from a paused to an elongation-competent conformation can be promoted by the action of an elongation factor, such as Drosophila Factor 5 (TFIIF). From a TFIIS mutant that was able to stimulate cleavage but that failed to promote read-through, Cipres-Palacin and Kane (31) suggested that conformational changes in the polymerase, in addition to transcript cleavage, are probably necessary for efficient read-through. It is possible that conformational changes can be induced by different mechanisms depending on the interacting factors and that several steps are involved, some of which may be common for different pathways. The S1 mutation may slow one of the intermediate steps, so that even though the ternary complex can physically interact with DmS-II (or NH 4 ϩ ) and carry out transcript cleavage, the final conformational change can not be reached normally. At the same time, the S1 mutation may alter the basal conformation of the polymerase in a way that it is already more competent for elongation in the absence of elongation factors. Comparative physical studies of elongation complexes containing wild type or mutant polymerase should provide additional insights into conformation changes involved in transcript elongation and may reveal how the regions of the subunits affected by the C4 and S1 mutations are involved in these processes.
FIG. 6. Comparison of pyrophosphorolysis and DmS-II-mediated transcript cleavage from the 13-mer by P2 wild type, C4 mutant, and S1 mutant RNA polymerase II. Elongation complexes were formed on immobilized dC-3025 as described under "Experimental Procedures" and as indicated by the diagram in the lower portion of this figure. The black circle with W symbol indicates that the elongation complexes were washed at this step. Elongation complexes were incubated for the indicated times with 1 mM pyrophosphate (A) or 1 nM DmS-II (B). RNAs were analyzed on an 18% polyacrylamide-6 M urea-TBE gel. Transcripts sizes are indicated in nucleotides.