GreB requires SI3, but not TL folding, to stimulate transcript hydrolysis. (a) Transcript cleavage rates with 50 nM to 20 µM GreB by wild-type (black circles), TH^PP (red triangles), and ΔSI3 (blue triangles) RNAPs. Hyperbolic fits were applied to wild-type and TH^PP curves. (b) Estimated apparent V_max of GreB-stimulated RNA cleavage and apparent K_d for GreB to backtracked complexes. Kinetic parameters were derived from curve-fitting in (a) and Fig. 2d.

TL folding is similarly rate-limiting for rapid nucleotide addition, his pause escape, and pyrophosphorolysis. (a) Sequence alignment of TL sequences from representative RNAP or RNAPII enzymes. Residues are colored as in (b) and (c). (b) Active-site configurations for pyrophosphorolysis (left), pause escape (middle), and rapid nucleotide addition (non-pause; right). Location of selected TL residues that contact substrate NTP (M932, R933, and H936) or stabilize the THB (T931, T934, and I1134) are shown6. (c) Effects of single amino-acid substitutions of selected TL residues on pyrophosphorolysis (solid bars), pause escape (hatched bars), and non-pause nucleotide addition (open bars). Rates were normalized to those observed for the wild-type enzyme; numbers above the bars indicate fold difference with wild-type RNAP. Pause escape data for T928A, T931A, M932A, R933A, T934A, F935A, H936A, I1134V, and G1136S are from Toulokhonov et al. 200722.

Significant discrimination against 2′ dNTPs is retained in ΔTL RNAP. (a) Nucleic-acid scaffold used to measure discrimination between ATP and 2′dATP and reaction scheme. Red asterisk, location of [α-³²P]GMP. The gel panel illustrates the distinct mobility of RNA^A10 (incorporated AMP, lane 3) and RNA^2′dA10 (incorporated 2′ dAMP, lane 2). (b) Selection of ATP over 2′ dATP by ΔTL RNAP. Reaction rates are plotted versus the concentration of ATP (filled squares; left y axis) or 2′ dATP (hollow squares; right y axis; Supplementary Methods). Error bars represent standard deviations from 3 or more independent measurements. (c) k_pol^app, K_d^app, and selectivity (k_pol^app/K_d^app)2 for ATP and 2′dATP by ΔTL RNAP were derived by hyperbolic curve-fitting weighted by standard deviation as depicted in (b), with errors propagated from standard error of the hyperbolic curve-fitting parameters (R²=0.967 and 0.959 for ATP and 2′dATP curves, respectively; Supplementary Methods).

The TL is dispensable for intrinsic RNA hydrolysis on a locked scaffold. (a) Nucleic-acid scaffold. DNA and RNA are shown in black and red, respectively. Red asterisk, ³²P label. (b) Reaction scheme and electrophoretic separation of RNA products from a typical reaction at pH 9.0 and 20 mM MgCl₂ by wild-type and TH^PP RNAPs in the absence or presence of 1 µM GreB (Supplementary Methods). (c) Plot of the fraction of remaining RNA^U-2 against time by wild-type (black circles), TH^PP (purple triangles), and ΔTL (red squares) RNAPs in the absence of GreB. (d) Rates of intrinsic RNA^U-2 cleavage in the presence of 1–80 mM Mg²⁺ and no GreB by wild-type (black circles) and ΔTL (red squares) RNAPs. Error bars in all figures represent standard deviations from three or more replicates. (e) Estimated apparent V_max of intrinsic RNA cleavage and apparent K_d for Mg²⁺II derived from the curve-fitting in (d).

TL dynamics in nucleotidyl transfer, transcriptional pausing, and transcript cleavage. Different TL conformations occur during nucleotidyl transfer (top row) and RNA hydrolysis (bottom row) reactions. During nucleotidyl transfer, a substrate (NTP or pyrophosphate) enters the active site while the TL is relaxed (green, PDB 1IW7)7, possibly inducing partial TL folding (blue, PDB 2O5I)40. Correct base-pairing of NTP to DNA template allows TH formation (orange, PDB 2O5J)6. After nucleotide addition, PPi exits the active site and TH unfolds into TL (top row). At non-backtrack pauses, the active site is transiently inhibited by a structural rearrangement (red) that allows or is caused in part by fraying of the RNA 3′ nt and that inhibits TL→TH folding22. The inhibited active site, nonetheless, appears able to translocate and to bind NTP22,33,34. To escape the pause, the TL must release from the restricted conformation (red, middle row). In contrast, backtracked complexes do not require TH formation, and the TL remains unfolded (light blue, PDB 1Y1V)12 during GreB- or TFIIS-stimulated RNA hydrolysis (bottom row).

Nucleotidyl transfer and RNA hydrolysis reactions catalyzed by multisubunit RNAPs. (a) Nucleotidyl transfer (left) and RNA hydrolysis (right) reactions occur by S_N2 nucleophilic attack with electron and proton transfer via a two-metal-mediated trigonal bipyramidal transition state, with reactants positioned in the P and A sites. Blue, electron path. Green, proton path. A:, general acid. B:, general base. For nucleotidyl transfer (left), apparent NTP K_m^app and V_max^app (forward reaction) and rates of reverse reaction at 1 mM PPi are shown for wild-type RNAP (black) and a mutant RNAP blocked in TL folding (TH^PP, red)6,8,40. For hydrolysis (right), apparent Mg²⁺II K_d^app and V_max^app (min⁻¹) for RNA cleavage in a 2-nt-backtracked scaffold (Figs. 2 and 3) in the absence (black) or presence (purple) of GreB are shown.
(b) Structure of RNAP catalytic center and proposed TH/TL transitions. DNA template, nascent RNA, incoming NTP, Mg²⁺ ions, and the bridge helix are shown in black, red, green, yellow, and cyan, respectively. TH (orange) PDB ID 2O5J6; unfolded TL (green) PDB ID 1IW77. E. coli SI3 (gray) is depicted in an approximate location24 with flexible connection to TL/TH (dotted lines). (c) The extensive network of electrostatic (red dotted lines) and van der Waals and hydrogen bonding (blue dotted lines) interactions that stabilize the THB (view rotated 70 ° relative to b). Key contacting residues outside the THB also are shown (light gray).