Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Nov 10, 2009; 106(45): 18942–18947.
Published online Oct 23, 2009. doi:  10.1073/pnas.0905402106
PMCID: PMC2776430

Allosteric control of catalysis by the F loop of RNA polymerase


Bacterial RNA polymerases (RNAPs) undergo coordinated conformational changes during catalysis. In particular, concerted folding of the trigger loop and rearrangements of the bridge helix at the RNAP active center have been implicated in nucleotide addition and RNAP translocation. At moderate temperatures, the rate of catalysis by RNAP from thermophilic Thermus aquaticus is dramatically reduced compared with its closest mesophilic relative, Deinococcus radiodurans. Here, we show that a part of this difference is conferred by a third element, the F loop, which is adjacent to the N terminus of the bridge helix and directly contacts the folded trigger loop. Substitutions of amino acid residues in the F loop and in an adjacent segment of the bridge helix in T. aquaticus RNAP for their D. radiodurans counterparts significantly increased the rate of catalysis (up to 40-fold at 20 °C). A deletion in the F loop dramatically impaired the rate of nucleotide addition and pyrophosphorolysis, but it had only a moderate effect on intrinsic RNA cleavage. Streptolydigin, an antibiotic that blocks folding of the trigger loop, did not inhibit nucleotide addition by the mutant enzyme. The resistance to streptolydigin likely results from the loss of its functional target, the folding of the trigger loop, which is already impaired by the F-loop deletion. Our results demonstrate that the F loop is essential for proper folding of the trigger loop during nucleotide addition and governs the temperature adaptivity of RNAPs in different bacteria.

Keywords: nucleotide addition, RNA cleavage, streptolydigin, temperature adaptation, transcription

Cellular multisubunit RNA polymerases (RNAPs) transcribe genes of up to tens of thousands nucleotides long with high precision and processivity. The catalytic cycle of RNAP is driven by complex conformational changes that accompany NTP binding, catalysis, and RNAP translocation. Recent studies identified two elements in the largest RNAP subunit (β′ in bacteria, Rpb1 in eukaryotic RNAP), the trigger loop (TL, or G loop), and the bridge helix (BH, or F-bridge helix), which appear to play the key role during the nucleotide addition cycle (Fig. 1). In the absence of a nucleotide substrate, the TL [amino acids 1234–1254 in Thermus aquaticus (Taq) and Thermus thermophilus RNAPs; amino acids 1077–1097 in Saccharomyces cerevisiae (Sce) RNAPII] adopts an open conformation in which its central part is unstructured (1, 2). Binding of an NTP substrate induces folding of the TL, resulting in extension of the two α-helixes at the base of the TL and creating a closed, catalytically competent conformation of the active center in which the NTP is properly aligned with the 3′-OH of the nascent RNA for facile catalysis (Fig. 1 A–C) (3, 4). The folded TL and the BH form a three-helix bundle that interacts with the substrate NTP and the template DNA base. In particular, the TL residues M1238 and H1242 (Taq numbering is used throughout the manuscript unless otherwise indicated) contact the base and the phosphate moieties of the incoming NTP, respectively, whereas the central part of the BH contacts the template base (Fig. 1 A–C); importantly, many additional β and β′ residues make essential interactions with the NTP substrate (3). The tip of the folded TL is accommodated in a pocket (Fig. 1) that is jointly formed by the N terminus of the BH (1068–1074 in Taq RNAP; 811–817 in Sce RNAPII); a loop adjacent to the N terminus of the BH (amino acids 1039–1067; 762–810 in Sce RNAPII), herein called the F loop; a short α-helix preceding the F loop (amino acids 1033–1038; 756–761 in Sce RNAPII); and several conserved amino acids of the β-subunit. The role of these elements in RNAP function has not been studied.

Fig. 1.
Structure of the RNAP active site and location of substitutions in the F loop. (A–C) Three different views of the active center in the EC of T. thermophilus RNAP with the incoming NTP substrate (black) in the insertion site (3). The whole EC structure ...

After NMP incorporation and PPi release, the enzyme must move one nucleotide downstream the DNA to allow for addition of the next nucleotide. This process is likely accompanied by (i) a helix→loop transition of the TL (35) and (ii) local unfolding or shifting of the BH, as observed in different RNAP and elongation complex (EC) structures (58). Both elements have been proposed to play a role in translocation, and many mutations in the TL and the BH impair the catalytic cycle (3, 912). However, certain substitutions in the TL (G1136S in Escherichia coli RNAP; ref. 10) or in the BH (F773V in E. coli RNAP; ref. 13), and many substitutions in Methanocaldococcus jannaschii (14), lead to moderate increase in the rate of RNA synthesis.

The catalytic cycle of RNAP is interrupted by inhibitors that restrict mobility of the TL and the BH (Fig. S1). In bacterial RNAP, antibiotic streptolydigin (Stl) binds in the middle of the BH and blocks folding of the TL (3, 15). In eukaryotic RNAPII, cyclopeptide inhibitor α-amanitin binds in a pocket formed by the F loop, contacts the TL, and presumably prevents its folding during the catalytic cycle (5, 12, 16). α-Amanitin was also proposed to trap an EC intermediate in which the “wedged” TL stabilizes a shifted conformation of the BH, which in turn occludes the site that the template DNA base would occupy after translocation (5). Thus, coordinated structural transitions of the TL and the BH may control the nucleotide addition cycle. RNAP can also carry out pyrophosphorolysis, the reversal of the nucleotide addition reaction, and the reactions of exonucleolytic and endonucleolytic RNA cleavage (17). These reactions are catalyzed by the same pair of catalytic Mg2+ ions bound in the RNAP active center (18), but the role of other RNAP elements involved in nucleotide addition in these reactions remains unclear.

RNAPs from thermophilic bacteria display much lower catalytic rates at low and moderate temperatures than the enzymes from mesophilic bacteria (1921). In particular, we demonstrated that Taq RNAP transcribed much slower than its closest mesophilic relative, Deinococcus radiodurans (Dra) RNAP, at temperatures below 37 °C (21). This “cold sensitivity” of thermophilic enzymes is thought to result from their reduced conformational mobility at low temperatures (22)—thus, the analysis of “cold sensitivity” could reveal key contributors to the conformational dynamics during catalysis. In this work, we sought to identify structural elements of RNAP responsible for the observed differences between RNAPs from Taq and Dra. We show that this difference can be largely accounted for by the differences in the F loop and an adjacent segment of the BH. We also demonstrate that the F loop is necessary for the efficient RNA synthesis but is less important for endonucleolytic RNA cleavage. We propose that the F loop promotes the proper folding of the TL during nucleotide addition.


Elongation by Taq RNAP and Its Mosaic Variants.

Even single amino acid substitutions in the structural elements involved in catalysis dramatically alter the catalytic properties of RNAP (3, 9, 12, 14, 17, 18). We thus wondered whether the differences in the cold sensitivity of the Taq and Dra enzymes could be conferred by changes of just a few residues. Sequence alignments of the “catalytic” β′-subunits revealed several amino acid substitutions that could play a key role in defining the rate of RNA synthesis. In particular, Dra RNAP contained four substitutions in the TL (V1246I, V1248G, T1250G, and Q1254M), one substitution (S1074T) in the BH, and eight substitutions in the F loop (with no substitutions in the adjacent N-terminal α-helix and in the β-subunit residues that contact the folded TL; Fig. 1D). Any of these substitutions (shown in Fig. 1 A and C) could potentially affect the rate of catalysis by altering the concerted transitions of the TL and the BH. Whereas changes in these two elements may directly alter their mobility, the F-loop substitutions could act indirectly, through altering the F-loop contacts with the TL and the BH. For example, Q1046A substitution changes an amino acid that is in close proximity to the tip of the folded TL, whereas S1049D, F1053I, and S1059A substitutions are near the upper part of the BH (Fig. 1 A and C).

To assess the role of these substitutions, we created two mosaic derivatives of Taq RNAP in which the selected regions were replaced with the corresponding sequences from Dra. In the first mosaic enzyme (G-loop-Dra), we replaced the TL (residues 1236–1255). In the second mosaic RNAP (F-Dra), we replaced the region including the F loop and the N-terminal part of the BH (residues 1039–1074, nine substitutions, including the single substitution in the BH). To facilitate subsequent structural analysis, all experiments were performed with RNAP variants lacking the nonconserved β′-N-terminal domain (amino acids 159–453). This deletion does not affect the rate of RNA synthesis but enables efficient crystallization of the recombinant RNAP (23).

The elongation properties of RNAP variants were analyzed on a linear DNA template containing the λPR promoter followed by the his terminator. On this template, RNAP is halted after addition of AMP at position 26 if transcription is carried out in the absence of CTP (the first C residue is encoded by the 27th template position). Stable, radiolabeled A26 ECs were chased by the addition of all four NTPs to 100 μM, and the aliquots were withdrawn at the desired times and quenched (Fig. 2A). The experiment was performed at 40 °C, the temperature optimum for the Dra RNAP (21). In agreement with the published data, there was a large difference in the average elongation rates of the Taq and Dra RNAPs; the half-times of the synthesis of the terminated RNA transcript were ≈80 s and ≈10 s, respectively. Mosaic G-loop-Dra RNAP did not differ significantly from the Taq enzyme; in fact, it had a somewhat reduced rate (Fig. 2A). In contrast, F-Dra RNAP displayed a much higher elongation rate (the half-time of ≈15 s; Fig. 2A).

Fig. 2.
Elongation rates of Taq, Dra, G-loop-Dra, and F-Dra RNAPs. The elongation rates were measured as described in Materials and Methods at 40 °C (A) and 20 °C (B). Positions of the halted 26-nt RNA and 157-nt terminated RNA product are shown ...

We then tested elongation properties of the F-Dra mosaic RNAP at 20 °C because the difference in elongation rates between Taq and Dra RNAPs is exacerbated at lower temperatures (21). Indeed, at 20 °C, Dra RNAP transcribed dramatically faster than Taq RNAP (≈45 s and [dbl greater-than sign]1,800 s, respectively; Fig. 2B). The mosaic RNAP was also significantly faster than the Taq enzyme (the half-time of ≈400 s).

Substitutions in the F Loop and the BH Affect the Rate of Single-Nucleotide Addition.

The observed differences in the aggregate elongation rate could be due to the differences in the NTP substrate binding, different rates of catalysis or RNAP translocation, or the altered pausing propensities of variant enzymes. To focus on the differences in the catalytic rate, we measured the rates of single-nucleotide addition in artificially assembled ECs. These experiments were performed by using a previously characterized minimal scaffold template (Fig. 3A) (24). The complex of RNAP with this template adopts a posttranslocation conformation in which the nascent RNA is resistant to pyrophosphorolysis and intrinsic cleavage (Fig. S2). The catalytic rates were measured at the saturating substrate concentration (1 mM UTP, specified by the template), which is several-fold higher than the apparent KM value for UTP for wild-type RNAP on this scaffold (≈200 μM; Fig. S3). The measurements were performed at 40 °C and 20 °C, allowing for a direct comparison with the elongation-rate measurements performed under the same conditions; the results of representative experiments performed at 20 °C are shown in Fig. 3B. The rate of nucleotide addition by the mesophilic Dra RNAP (kobs ≈ 500 s−1 at 40 °C) was similar to the rate reported previously for E. coli RNAP on another scaffold that was shown to support rapid nucleotide addition (3), indicative of the active conformation of the scaffold used in our experiments. The observed rate of UMP incorporation by Taq RNAP was ≈10-fold slower than the rate of Dra RNAP at 40 °C, and it was ≈320-fold slower at 20 °C (Fig. 3C, Table 1 and Table S1). These differences correlate well with the elongation-rate measurements. Thus, the differences in the average elongation rates of Taq and Dra RNAPs can be explained largely by their different rates of catalysis.

Fig. 3.
The rates of single-nucleotide addition measured for different RNAPs. (A) The structure of the minimal nucleic acid scaffold. (B) Electrophoretic separation of reaction products from a quench-flow assay of wild-type Taq and Dra RNAPs. The reaction was ...
Table 1.
Relative rates of catalysis by Taq RNAP and its mutant variants at 20 °C

The rate of catalysis by the mosaic G-loop-Dra RNAP was similar to that of the Taq enzyme, whereas the mosaic F-Dra RNAP was ≈6-fold faster at 40 °C and was ≈40-fold faster at 20 °C (Fig. 3C, Table 1, and Table S1). In fact, the rates of catalysis by the F-Dra RNAP were closer to those by the Dra RNAP. Control experiments demonstrated that the mosaic RNAP had about the same apparent KM value for UTP as the wild-type Taq RNAP (≈200 μM; Fig. S3); therefore, the observed differences are likely due to differences in catalysis and not in the substrate binding. These data show that nine amino acid residues in the F loop and the BH apparently account for the large part of the catalytic rate difference between Taq and Dra RNAPs.

To determine which of these residues play a key role in catalysis, we constructed additional variants of Taq RNAP with substitutions in this region: (i) S1074T substitution, the only variant residue in the BH; (ii) Q1046A substitution of a residue in close contact with the TL in the insertion complex; (iii) substitution of the central part of the F loop (F-loop-Dra; five substitutions in total: Q1046A, K1047R, S1049D, E1051S and F1053I); and (iv) deletion of the central part of the F loop (ΔF-loop; amino acids 1044–1056 replaced with Gly; Fig. 1). The rates of nucleotide addition by RNAPs with single substitutions Q1046A and S1074T were significantly faster than the rate of wild-type RNAP (1.5- to 2-fold at 40 °C and 3- to 3.5-fold at 20 °C; Table 1 and Table S1). The F-loop-Dra RNAP containing five substitutions in the F loop, including Q1046A, had the same rate of catalysis as the RNAP with the single Q1046A substitution, suggesting that the other four substitutions either do not affect or compensate each other's effects on the rate of catalysis. These results demonstrate that the observed difference in the catalytic rates of Taq and Dra RNAPs may result from the combination of just a few amino acid substitutions, including the two β′-residues, 1046 in the F-loop and 1074 in the BH. Further experiments are required to establish whether these two substitutions have additive effects on catalysis and to identify additional residues that affect the rate of RNA synthesis.

Deletion of the central part of the F loop had the most dramatic effect on the rate of NMP addition, which was reduced ≈700- and ≈70-fold at 40 °C and 20 °C, respectively, relative to the wild-type Taq RNAP (Table 1 and Table S1), as well as on pyrophosphorolysis (Fig. S4). On the other hand, the ΔF-loop RNAP had the same apparent KM value for UTP as the wild-type enzyme (Fig. S3). Thus, the F loop plays an essential role in nucleotide addition but is dispensable for the substrate binding.

Effects of Stl on Nucleotide Addition by T. aquaticus RNAP with Mutations in the F Loop.

Stl binds near the center of the BH in the bacterial RNAP and prevents folding of the TL to stabilize the EC with the NTP trapped in the preinsertion state (3). Stl strongly inhibits nucleotide addition, pyrophosphorolysis, and intrinsic RNA cleavage (15); it has been proposed that these diverse effects of Stl are mediated by freezing of a nonproductive RNAP conformation, but the detailed mechanism of inhibition has not been studied.

We tested effects of Stl on catalysis by Taq RNAP and its mosaic variants at 20 °C, the temperature at which the analyzed RNAPs have pronounced differences in the catalytic rates (Table 1). Stl (at 100 μg/mL) slowed nucleotide addition by the wild-type Taq RNAP about 80-fold (Fig. 3C and Table S1). The effect of Stl was much higher in the case of the F-Dra RNAP, which was inhibited about 3,300-fold at the same conditions (Fig. 3C and Table S1). In contrast, ΔF-loop RNAP was essentially resistant to Stl (less than 2-fold inhibition; Fig. 3C and Table S1). At the same time, the mutation did not impair the antibiotic binding (see below, the last section of Results). Remarkably, all three enzymes had similar rates of catalysis in the presence of Stl, comparable to the rate of the ΔF-loop RNAP in the absence of the antibiotic (Fig. 3C). Together, these data suggest that Stl and the deletion of the β′-F-loop may affect the nucleotide addition reaction in a similar fashion.

Effects of Mutations in the F Loop on Substrate Discrimination.

Structural modeling suggests that the F-loop may control catalysis through direct interactions with the TL (see Discussion). Mutations in the TL of Sce RNAPII impair discrimination of correct nucleotides against noncomplementary and deoxyribonucleotide substrates (12). We therefore tested whether mutations in the F loop confer similar effects. We analyzed incorporation of complementary dTMP and noncomplementary CMP on the minimal nucleic acid scaffold template (Fig. 3A). Both substrates were incorporated by Taq RNAP with a very low efficiency; the rates of dTMP and CMP addition (measured at 40 °C at 1 mM concentration of the substrates) were ≈17,000- and 3,500-fold slower, respectively, than the UMP addition rate (Table S1). F-Dra RNAP displayed similar differences in the rates for correct and incorrect nucleotides. In contrast, the ΔF-loop RNAP was unable to discriminate against the noncognate substrates with the same efficiency as the wild-type RNAP: the rates for dTMP and CMP incorporation were only ≈120- and 700-fold slower than the rate for UMP incorporation (Table S1). These results demonstrate that the F loop is required for substrate selection by RNAP, likely by affecting TL/BH dynamics (see Discussion), and is particularly important for rNTP/dNTP discrimination.

Role of the F Loop in RNA Cleavage.

Nascent RNA cleavage in backtracked ECs is mediated by the RNAP active site and requires the two catalytic Mg2+ ions (17, 18). This reaction is stimulated by the Gre cleavage factors (2528) and the 3′ end of the nascent RNA (29), and it is abolished by substitutions of the catalytic Asp residues (18). However, the role of other RNAP elements involved in catalysis, including the TL and the BH, in this reaction has not been studied. It also remained unknown whether RNAPs that differ in the rates of RNA synthesis (e.g., Taq and Dra RNAPs) have similar differences in the rates of RNA cleavage.

We measured the rates of intrinsic RNA cleavage by the wild-type Taq and Dra RNAPs and by the mosaic Taq RNAP variants in ECs assembled on a nucleic acid scaffold shown in Fig. 4A; this scaffold has been shown to support robust RNA cleavage (18). Incubation of ECs in the presence of Mg2+ (10 mM) resulted in cleavage of the two 3′-terminal nucleotides from the 13-nt-long RNA to yield an 11-nt RNA product (Fig. 4B). At 20 °C, the difference in the rates of RNA cleavage by the Taq and Dra RNAPs was only about 6-fold, in a sharp contrast to the 320-fold difference in the nucleotide addition rates (Fig. 4B, Table 1, and Table S1). The rate of RNA cleavage by the mosaic F-Dra RNAP was about 2-fold faster than the rate of the wild-type Taq RNAP. Remarkably, the RNA cleavage rate by the ΔF-loop enzyme was only ≈3-fold slower than the rate of the wild-type Taq RNAP.

Fig. 4.
Kinetics of RNA cleavage by RNAP from Taq and its mutant variants. (A) The scaffold used in the experiments. The position of RNA cleavage is shown by an arrowhead. (B) Kinetics of RNA cleavage at 20 °C. Positions of the starting 13-nt RNA and ...

Stl efficiently inhibited RNA cleavage by all three Taq RNAP variants, reducing it to undetectable levels (Fig. 4C). The Stl IC50 values in this reaction were comparable for all three RNAPs (Fig. S5) indicating that they do not differ significantly in the antibiotic binding, and therefore the observed differences in the effects of Stl on nucleotide addition by these RNAPs (see above and Fig. 3C) cannot be explained by differences in the efficiencies of Stl binding. Although it remains unknown whether Stl inhibits the cleavage reaction itself or backtracking of RNAP that precedes the cleavage, it can be concluded that the inhibition does not depend on the F loop. Together, these results indicate that the F loop is not required for the RNA hydrolysis by RNAP (see Discussion).


In this work, we identify the β′-subunit F-loop as an additional structural element near the active center which, along with the TL and the BH, plays an essential role in catalysis by bacterial RNAP. The F loop is located near the upper part of the BH and directly contacts the tip of the folded TL in the EC with the substrate bound in the RNAP active center (Fig. 1). Together with the BH, the F loop forms a gateway that accommodates the folded TL during nucleotide addition. The F loop may therefore modulate catalysis by affecting the conformation of the TL and/or the BH. Several lines of evidence obtained in our work suggest that the F loop may be required for the proper folding of the TL and may stabilize the closed conformation of the active center during catalysis.

First, a deletion in the F loop in Taq RNAP led to marked defects in the catalytic properties of RNAP, which were strikingly similar to those conferred by substitutions in the TL. In particular, deletions of the TL in Taq and E. coli RNAPs (9, 15) and amino acid substitutions in the TL in E. coli RNAP (3, 9) and in Sce RNAPII (12) also resulted in a dramatic decrease in the rate of nucleotide addition.

Second, substitutions and deletions in the TL in bacterial and eukaryotic RNAPs (3, 9, 12, 15) and the F-loop deletion did not change the apparent KM value for NTP, suggesting that their principal defect was in catalysis rather than in substrate binding. Thus, efficient NTP binding can occur in the open active center conformation independently of the TL folding.

Third, Stl does not inhibit RNA elongation by RNAPs with the deletion in the F loop (Fig. 3C) or in the TL (15). Stl binds to the unfolded TL and blocks its folding, freezing the EC in an inactive state (Fig. S1A) (3). Resistance to the antibiotic can be conferred by an unusual folded state of the TL, such as in the hairpin-stabilized paused EC, which is thought to block Stl binding (9). By contrast, the deletions of the F loop and the TL apparently do not impair the antibiotic binding (Fig. S5 and ref. 15); in this case, the resistance to Stl likely results from the loss of its functional target, the folding of the TL, which is already impaired by these deletions.

Finally, similarly to the TL substitutions in Sce RNAPII (12) and in E. coli RNAP (10), the F-loop deletion impaired substrate selection. It has been proposed that the TL acts as a critical component of kinetic selection of correctly matched NTP substrates by promoting their faster incorporation in comparison with incorrect substrates (12, 30). In support of this hypothesis, substitution of a residue in the α-helix at the base of the TL that affected the equilibrium between the closed and open TL conformations decreased fidelity of RNA synthesis by Sce RNAPII (30). Our results suggest that the F loop also contributes to substrate selection by promoting proper folding of the TL.

Crystal structures of T. thermophilus ECs with or without bound NTP reveal local conformational changes of the F loop that may relate to the TL folding (Fig. 5). Although the conformation of the polypeptide backbone of the F loop does not change significantly upon NTP binding, the positions of several amino acid side chains in this region change by 1.6–4.5 Å. In particular, these changes involve β′-residues Q1033, Q1037, R1042, L1044, and Q1046, which contact the tip of the folded TL, and E1051, which is located at the tip of the F loop (Fig. 5). Although the significance of the observed changes for the TL folding remains unknown, this suggests a certain degree of conformational mobility of the F loop that may play an important role in catalysis. Interestingly, the F loop also appears to be a target for the CBR compounds that were proposed to act by freezing an unproductive conformation of RNAP (31). E. coli mutations conferring resistance to CBR have been mapped to the F loop and to an adjacent segment of the BH; P750L substitution in the central part of the F loop and F773V (correspond to positions P1048 and F1071 in Taq; Fig. 5) resulted in an increased rate of transcript elongation in vitro and CBR dependence in vivo (31); the latter result indicates that the F loop is not a binding site, but rather a functional target of CBRs.

Fig. 5.
Possible conformational changes in the F loop during catalysis. A close-up view of the TL–F-loop interface in the NTP-bound EC (3). The RNAP orientation and the color scheme are the same as in Fig. 1A. Residues Q1033, Q1037, R1042, L1044, Q1046, ...

Another inhibitor of multisubunit RNAPs, α-amanitin, binds at the F loop of RNAPII and inhibits both RNAP translocation and catalysis (5, 12, 16). α-Amanitin directly interacts with the TL residue H1085 and fixes the TL in a wedged conformation, which likely stabilizes the shifted state of the BH and blocks translocation (Fig. S1B) (5). α-Amanitin also makes contacts with the F-loop residues that interact with the tip of the folded TL (Fig. S1 B and C), leading to a model in which α-amanitin acts by inhibiting the TL folding (5, 12). Indeed, similarly to substitutions in the TL, α-amanitin impairs substrate selection by RNAPII (12). The F loop of RNAP is therefore an attractive target for design of novel RNAP inhibitors.

In contrast to RNA synthesis, the cleavage reaction apparently does not depend on the F loop, because RNAP with the deletion in the F loop can efficiently catalyze RNA cleavage. Furthermore, RNA cleavage, but not RNA synthesis, by the ΔF-loop RNAP was efficiently inhibited by Stl. Therefore, the hypothetical F-loop-promoted TL folding may not be a rate-limiting step during RNA cleavage. In recently published structures of backtracked ECs of Sce RNAPII, the TL adopts open or partially open conformations and contacts backtracked RNA (32). These observations suggest that the nucleotide addition and cleavage reactions may depend on different TL conformations, only the first of which depends on the F loop. Such a model would explain the inhibitory effect of Stl on RNA cleavage by the ΔF-loop RNAP; for example, through blocking of a cleavage-competent conformation of the TL. It remains to be established whether the TL plays a key role in RNA cleavage, or whether other structural element(s) in the RNAP active center are sufficient for this reaction.

We demonstrated that substitutions in the F loop and in the adjacent segment of the BH may serve as a mechanism of adaptive regulation of the catalytic rates of RNAPs from mesophilic and thermophilic bacteria. Substitution of this region in Taq RNAP with the corresponding region from Dra RNAP resulted in a significant increase in the rate of nucleotide addition, especially at low and moderate temperatures. Remarkably, the observed increase was much more dramatic than reported previously for substitutions in TL and BH that stimulated the rate of elongation 2- to 3-fold (10, 14). We also showed that substitutions of two individual residues from this region in Taq RNAP, Q1046A and S1074A, increased the rate of catalysis about 3-fold. The first substitution changes amino acid residue that is in close contact with the TL in the substrate-bound EC and may therefore directly affect the TL folding during catalysis. The second substitution, S1074A, is located in the upper part of the BH. Its effect on catalysis may be explained by altered contacts between the BH and the F loop which, in its turn, interacts with the TL. Substitutions that affect the rate of RNA synthesis by E. coli RNAP were mapped to adjacent positions in the BH (13, 31). One of these substitutions, F773V (corresponds to Taq F1071; Fig. 5) was suppressed by R744C substitution (R1042 in Taq; Fig. 5) (31) in the F loop. It is therefore possible that certain substitutions in the F loop may also affect catalysis through changes in the conformation of the adjacent BH. The functional interplay between substitutions in the F loop and the BH suggests that the conformational changes of the BH and the F loop are coordinated in the catalytic cycle. These changes may be a target for transcription regulation and may underlie adaptive differences in catalytic rates of RNAPs from different organisms. Interestingly, the differences in the RNA cleavage rates by Taq and Dra RNAPs were much less dramatic than the differences in rates of RNA synthesis, implying that RNA cleavage is less dependent on thermal adaptation and is likely determined by distinct elements of the active center.

Sequence alignments reveal insertions at the tip of the F loop (varying in size from 5 to 20 residues) in several bacterial lineages, as well as in archaeal and eukaryotic species (Fig. 5 and Fig. S6). Although the effects of these insertions on RNAP catalysis have not been studied, it is possible that they alter catalytic properties of RNAP in a species-specific way. Indeed, previous observations pinpointed another insertion, the 188-residue-long SI3 segment in the TL in E. coli RNAP (see Fig. 5 for its position), to play an essential role in catalysis, and its deletion had complex effects on transcription elongation, pausing, and RNA cleavage (3, 13). Thus, further structural analysis of the conformational dynamics of the tripartite F-loop/TL/BH unit will provide important insights into the mechanisms of catalysis by multisubunit RNAP.

Materials and Methods

Wild-type core RNAPs and σA subunits from Taq and Dra were purified as described previously (20, 21, 23). Mosaic Taq RNAPs were obtained by site-directed mutagenesis and purified from E. coli cells similarly to the wild-type enzyme. The detailed description of RNAP purification and transcription assays is provided in the SI Methods.

Supplementary Material

Supporting Information:


We thank K. Severinov and K. Kuznedelov (Waksman Institute, Rutgers University, Piscataway, NJ) for plasmids and discussions, R. Landick for fruitful discussions, D. G. Vassylyev for help with structural modeling, A. Mustaev for support, and N. Zenkin for helpful comments. This work was supported in part by a Russian Academy of Sciences Presidium Program in Molecular and Cellular Biology grant (to A.K.), Russian Foundation for Basic Research Grant 07-04-00247, and National Institutes of Health Grants GM67153 (to I.A.), GM074840 (to D.G. Vassylyev), and GM30717. N.M.'s work at the Institute of Biotechnology in Vilnius was supported by a Federation of European Biochemical Societies Collaborative Experimental Scholarship.


The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

This article contains supporting information online at www.pnas.org/cgi/content/full/0905402106/DCSupplemental.


1. Vassylyev DG, Vassylyeva MN, Perederina A, Tahirov TH, Artsimovitch I. Structural basis for transcription elongation by bacterial RNA polymerase. Nature. 2007;448:157–162. [PubMed]
2. Kettenberger H, Armache KJ, Cramer P. Complete RNA polymerase II elongation complex structure and its interactions with NTP and TFIIS. Mol Cell. 2004;16:955–965. [PubMed]
3. Vassylyev DG, et al. Structural basis for substrate loading in bacterial RNA polymerase. Nature. 2007;448:163–168. [PubMed]
4. Wang D, Bushnell DA, Westover KD, Kaplan CD, Kornberg RD. Structural basis of transcription: Role of the trigger loop in substrate specificity and catalysis. Cell. 2006;127:941–954. [PMC free article] [PubMed]
5. Brueckner F, Cramer P. Structural basis of transcription inhibition by alpha-amanitin and implications for RNA polymerase II translocation. Nat Struct Mol Biol. 2008;15:811–818. [PubMed]
6. Zhang G, et al. Crystal structure of Thermus aquaticus core RNA polymerase at 3.3 A resolution. Cell. 1999;98:811–824. [PubMed]
7. Cramer P, Bushnell DA, Kornberg RD. Structural basis of transcription: RNA polymerase II at 2.8 angstrom resolution. Science. 2001;292:1863–1876. [PubMed]
8. Vassylyev DG, et al. Crystal structure of a bacterial RNA polymerase holoenzyme at 2.6 A resolution. Nature. 2002;417:712–719. [PubMed]
9. Toulokhonov I, Zhang J, Palangat M, Landick R. A central role of the RNA polymerase trigger loop in active-site rearrangement during transcriptional pausing. Mol Cell. 2007;27:406–419. [PubMed]
10. Bar-Nahum G, et al. A ratchet mechanism of transcription elongation and its control. Cell. 2005;120:183–193. [PubMed]
11. Epshtein V, et al. Swing-gate model of nucleotide entry into the RNA polymerase active center. Mol Cell. 2002;10:623–634. [PubMed]
12. Kaplan CD, Larsson KM, Kornberg RD. The RNA polymerase II trigger loop functions in substrate selection and is directly targeted by alpha-amanitin. Mol Cell. 2008;30:547–556. [PMC free article] [PubMed]
13. Svetlov V, Belogurov GA, Shabrova E, Vassylyev DG, Artsimovitch I. Allosteric control of the RNA polymerase by the elongation factor RfaH. Nucleic Acids Res. 2007;35:5694–5705. [PMC free article] [PubMed]
14. Tan L, Wiesler S, Trzaska D, Carney HC, Weinzierl RO. Bridge helix and trigger loop perturbations generate superactive RNA polymerases. J Biol. 2008;7:40. [PMC free article] [PubMed]
15. Temiakov D, et al. Structural basis of transcription inhibition by antibiotic streptolydigin. Mol Cell. 2005;19:655–666. [PubMed]
16. Bushnell DA, Cramer P, Kornberg RD. Structural basis of transcription: alpha-amanitin-RNA polymerase II cocrystal at 2.8 A resolution. Proc Natl Acad Sci USA. 2002;99:1218–1222. [PMC free article] [PubMed]
17. Sosunov V, et al. Unified two-metal mechanism of RNA synthesis and degradation by RNA polymerase. EMBO J. 2003;22:2234–2244. [PMC free article] [PubMed]
18. Sosunov V, et al. The involvement of the aspartate triad of the active center in all catalytic activities of multisubunit RNA polymerase. Nucleic Acids Res. 2005;33:4202–4211. [PMC free article] [PubMed]
19. Xue Y, Hogan BP, Erie DA. Purification and initial characterization of RNA polymerase from Thermus thermophilus strain HB8. Biochemistry. 2000;39:14356–14362. [PubMed]
20. Minakhin L, Nechaev S, Campbell EA, Severinov K. Recombinant Thermus aquaticus RNA polymerase, a new tool for structure- based analysis of transcription. J Bacteriol. 2001;183:71–76. [PMC free article] [PubMed]
21. Kulbachinskiy A, Bass I, Bogdanova E, Goldfarb A, Nikiforov V. Cold sensitivity of thermophilic and mesophilic RNA polymerases. J Bacteriol. 2004;186:7818–7820. [PMC free article] [PubMed]
22. Fields PA. Review: Protein function at thermal extremes: Balancing stability and flexibility. Comp Biochem Physiol A Mol Integr Physiol. 2001;129:417–431. [PubMed]
23. Kuznedelov K, et al. Recombinant Thermus aquaticus RNA polymerase for structural studies. J Mol Biol. 2006;359:110–121. [PubMed]
24. Korzheva N, et al. A structural model of transcription elongation. Science. 2000;289:619–625. [PubMed]
25. Laptenko O, Lee J, Lomakin I, Borukhov S. Transcript cleavage factors GreA and GreB act as transient catalytic components of RNA polymerase. EMBO J. 2003;22:6322–6334. [PMC free article] [PubMed]
26. Sosunova E, et al. Donation of catalytic residues to RNA polymerase active center by transcription factor Gre. Proc Natl Acad Sci USA. 2003;100:15469–15474. [PMC free article] [PubMed]
27. Opalka N, et al. Structure and function of the transcription elongation factor GreB bound to bacterial RNA polymerase. Cell. 2003;114:335–345. [PubMed]
28. Vassylyeva MN, et al. The carboxy-terminal coiled-coil of the RNA polymerase beta'-subunit is the main binding site for Gre factors. EMBO Rep. 2007;8:1038–1043. [PMC free article] [PubMed]
29. Zenkin N, Yuzenkova Y, Severinov K. Transcript-assisted transcriptional proofreading. Science. 2006;313:518–520. [PubMed]
30. Kireeva ML, et al. Transient reversal of RNA polymerase II active site closing controls fidelity of transcription elongation. Mol Cell. 2008;30:557–566. [PubMed]
31. Artsimovitch I, Chu C, Lynch AS, Landick R. A new class of bacterial RNA polymerase inhibitor affects nucleotide addition. Science. 2003;302:650–654. [PubMed]
32. Wang D, et al. Structural basis of transcription: Backtracked RNA polymerase II at 3.4 angstrom resolution. Science. 2009;324:1203–1206. [PMC free article] [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • MedGen
    Related information in MedGen
  • PubMed
    PubMed citations for these articles
  • Substance
    PubChem Substance links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...