Trigger loop of RNA polymerase is a positional, not acid-base, catalyst for both transcription and proofreading

Abstract

The active site of multisubunit RNA polymerases (RNAPs) is highly conserved from humans to bacteria. This single site catalyzes both nucleotide addition required for RNA transcript synthesis and excision of incorrect nucleotides after misincorporation as a proofreading mechanism. Phosphoryl transfer and proofreading hydrolysis are controlled in part by a dynamic RNAP component called the trigger loop (TL), which cycles between an unfolded loop and an α-helical hairpin [trigger helices (TH)] required for rapid nucleotide addition. The precise roles of the TL/TH in RNA synthesis and hydrolysis remain unclear. An invariant histidine residue has been proposed to function in the TH form as a general acid in RNA synthesis and as a general base in RNA hydrolysis. The effects of conservative, nonionizable substitutions of the TL histidine (or a neighboring TL arginine conserved in bacteria) have not yet been rigorously tested. Here, we report that glutamine substitutions of these residues, which preserve polar interactions but are incapable of acid-base chemistry, had little effect on either phosphoryl transfer or proofreading hydrolysis by Escherichia coli RNAP. The TL substitutions did, however, affect the backtracking of RNAP necessary for proofreading and potentially the reactivity of the backtracked nucleotide. We describe a unifying model for the function of the RNAP TL, which reconciles available data and our results for representative RNAPs. This model explains diverse effects of the TL basic residues on catalysis through their effects on positioning reactants for phosphoryl transfer and easing barriers to transcript backtracking, rather than as acid-base catalysts.

Keywords: RNA hydrolysis; acid–base catalysis; proofreading; transcription; trigger loop.

Fig. 1.

Activities of RNA polymerase. (A) Nucleotide addition, pyrophosphorolysis, and backtracking events in RNAP active site. RNA synthesis (sequence of events at the top) is accomplished in a NAC consisting of four steps. Step 1: translocation of DNA and RNA to position the 3′-OH of RNA in i site on RNAP for the reaction with an incoming NTP; step 2: NTP binding to i+1 site; step 3: formation of a new phosphodiester bond; and step 4: pyrophosphate release. Template DNA is shown in black, nontemplate DNA in green, RNA in red, incoming NTP in blue, and catalytic magnesium ions in yellow. One position on template/nontemplate DNA is highlighted in purple, to illustrate translocation. Chemical reactions of the reversible phosphoryl transfer (step 3) and hydrolysis of a backtracked RNA are shown in purple and blue boxes, respectively. The flow of electrons in the reactions is marked with arrows. Protons are supplied by the putative general acid (“A”) and base (“B”). (B) Basic residues of the trigger helix as candidates for a general acid–base catalyst (PDB ID code 2O5J). The electron flow in the nucleotide addition reaction is indicated with arrows. (C) Structure of a 1-nt backtracked elongation complex (PDB ID code 4WQS), with the trigger loop in a partially unfolded, “bent” conformation (shown in pink). The folded trigger helix (from PDB ID code 2O5J) is shown for reference in orange.

Fig. 2.

Nucleotide addition by TL His mutants. (A) Multiround nucleotide addition assay. RNA extension by the RNAPs was monitored using scaffold 1 ligated to a 2-kb double-stranded DNA (dsDNA) fragment of rpoB gene via a phosphorylated overhang (in green on scaffold 1) complementary to a StyI-generated sticky end on the dsDNA (Materials and Methods and Fig. S1). Transcription progress along the ligated scaffold over time was visualized on a denaturing 8% polyacrylamide gel (Upper, time course at pH 8 is shown). Average transcription speeds are reported in nucleotides per second. The wild-type and H936Q RNAPs elongated at approximately the same speed across the tested pH range (Lower, tested pH values are shown above the data bars). (B) Addition of two consecutive nucleotides to an RNA primer. RNA in the starting EC was radiolabeled by incorporation of [α-³²P]GMP (G19) and further extended in the reaction with 1 mM ATP, to form A20 and A21 RNAs (Top; see Fig. S1 and Table S2 for complete oligonucleotide sequences). The reactions were quenched with HCl at various time points, and RNA products separated by gel electrophoresis (Middle). Quantified AMP addition data from at least three replicates per time point for the wild-type (closed circles) and the H936Q RNAP (open circles) closely matched. Data from all replicates were globally fitted to a single-intermediate reaction model (solid and dashed curves, respectively), to obtain individual rate constants shown (Bottom).

Fig. S1.

Ligated-scaffold transcription assay. (A) Nucleic-acid scaffold 1 used in elongation rate measurements. The starting 17mer RNA is italicized and bolded, with the sequence complementary to the template DNA shown in uppercase letters. Positions of halted 26-nt RNA and 66-nt run-off transcript are marked as A26 and C66, respectively. The 4-nt overhang and 3′ phosphate of the template DNA, necessary for ligation, are boxed. (B) Schematic of the experimental set-up for EC reconstitution, labeling, and ligation to a StyI-treated dsDNA fragment of rpoB (see Materials and Methods for details). RNAP and nucleic acids at various stages are represented as a cartoon. (C) Elongation of ligated A26 complexes by the wild-type and mutant RNAPs at pH 8.0. Indicated are the positions of halted A26 complexes, run-off transcript of unligated ECs (C66), and elongation products of successfully ligated ECs. Average transcription speeds were determined as previously described (41).

Fig. 3.

Pyrophosphorolysis by TL mutants. (A) RNA sequence of the nucleic-acid scaffold 2 was designed to promote reconstitution of RNAP in the pretranslocated register from which pyrophosphorolysis of the terminal UMP is possible. (B) Pyrophosphorolysis by the TL H936Q mutant. RNA cleavage data were fit to a single-exponential function. The wild-type and H936Q RNAPs catalyzed pyrophosphorolysis with similar kinetics. Error bars are smaller than data points. (C) Pyrophosphorolysis by the TL Arg mutants. All pyrophosphorolysis reactions were performed in the presence of apyrase, to degrade the released UTP, thereby favoring forward reaction (i.e., cleavage of the starting RNA). Refer to Table S2 for complete nucleic-acid sequences.

Fig. 4.

Intrinsic RNA cleavage by TL His mutants. (A) RNA hydrolysis in ECs backtracked by 4–6 nt. The starting RNA in the scaffold 3 was radiolabeled at the 3′-end by incorporation of [α-³²P]CMP and extended to the halt position (C21), previously demonstrated to cause RNAP to backtrack by multiple nucleotides (30) (Upper; refer to Table S2 and Fig. S2 for complete sequence of the nucleic acids and Fig. 1A for translocation states). Hydrolysis of the backtracked ECs was initiated with high concentration of Mg²⁺ in pH 9.0 buffer. Single-exponential kinetics of the C21 RNA disappearance is illustrated in the Lower panel. The TL His substitutions affected hydrolysis only slightly. (B) Exonuclease activity and backtracking. When reconstituted on the shown nucleic-acid scaffold 4 (Upper; see Table S2 for sequences) and walked one position forward to C17, wild-type RNAP partitions between pretranslocated and 1-nt backtracked states, as evidenced by the presence of both the 1- and 2-nt cleavage products in the hydrolysis reactions (Lower and Fig. S3). The TL His mutants preserve the ability to cleave 1 nt off of the RNA, and do so with kinetics similar to the wild-type RNAP, but fail to backtrack, as suggested by the lack of the 2-nt product (open blue and red circles, Lower, and Fig. S3). Thin black lines separate gel portions containing C17 reactant and short cleavage products. See Fig. S3B for an image of the full gel.

Fig. S2.

Intrinsic RNA cleavage by the TL mutants in ECs backtracked by multiple nucleotides. (A) Nucleic-acid scaffold 3 used in cleavage experiments. The starting 16mer RNA was radiolabeled at 3′-end by incorporation of [α-³²P]CMP and extended to the halt position (C21), previously demonstrated to backtrack by multiple nucleotides (30). (B) Cleavage of the C21 RNA by TL His mutants. The backtrack positions giving rise to the cleavage products observed on the gel (5–7 nt in length) are numbered on the cartoon of the EC. (Inset) Distribution of major cleavage products in end-point hydrolysis reactions of the wild-type and TL His mutant polymerases. (C) Cleavage of the C21 RNA by TL Arg mutants. (Inset) Populations of hydrolysis products in end-point reactions of the wild-type and TL Arg variants.

Fig. S3.

Endo- and exonuclease activities of the TL variants. (A) Nucleic-acid scaffold 4 used in cleavage experiments. The scaffold is based on the published consensus pause sequence (31), on which a portion of ECs was shown to backtrack by one nucleotide (32). The 16mer RNA was extended to the pause position by incorporation of [α-³²P]CMP, and the resulting C17 ECs allowed to equilibrate between the pretranslocated and 1-nt backtracked registers. RNA hydrolysis was commenced with high pH and Mg²⁺ concentration. (B) Cleavage of the C17 RNA by TL His mutants. The 1- and 2-nt cleavage products are marked on the gel. (C) Cleavage of the C17 RNA by TL Arg mutants. The translocation state of an EC that gives rise to each cleavage product is shown as a cartoon.

Fig. 5.

Nucleotide addition by TL Arg variants. (A) Multiround nucleotide addition by a point R933Q mutant and a double mutant carrying glutamine substitutions at both the TL Arg and His. Average transcription speeds are reported in nucleotides per second. The pH activity profiles are shown (Lower). (B) Addition of two consecutive nucleotides to an RNA primer in pH 8.0 buffer. A 1.5- and 1.8-fold decrease in the rates of the first and second AMP addition, respectively, was observed relative to the wild-type RNAP (Fig. 2B). The experiments and data analysis were performed in a manner identical to those with the TL His mutants (Fig. 2).

Fig. 6.

Intrinsic RNA cleavage by TL Arg mutants. (A) RNA hydrolysis in ECs backtracked by multiple nucleotides was not reduced by the R933Q substitution. The double mutant hydrolyzed RNA ∼3.4-fold slower. (B) Exonuclease activity and backtracking of RNAP are not compromised by the R933Q mutation. The experiments and data analysis were performed in a manner identical to those with the TL His mutants (Fig. 4). Thin black lines separate gel portions containing C17 reactant and short cleavage products. See Fig. S3C for an image of the full gel.

Fig. 7.

Positional catalyst model of TL/TH function. The TL/TH serves as a positional, not an acid–base, catalyst in transcription and proofreading by RNAP. During the bond formation step of the nucleotide addition cycle (A), the TL is folded into the active site, thus allowing the His and Arg to orient the NTP phosphates for the substitution reaction through hydrogen-bonding contacts (Inset, PDB ID code 2O5J). Upon incorporation of an incorrect nucleotide, RNAP reverse threads the nucleic-acid template (backtracks) and the TL assumes a partially folded conformation, such that its invariant His can form ionic and hydrogen-bonding contacts with the backtracked RNA nucleotide (B and C). The TL/TH His could then contribute to proofreading RNA hydrolysis by directly positioning the backtracked ribobase for excision, thereby lowering the activation barrier for the reaction (B, green dashed line). Alternatively, the interactions with the His could stabilize 3′ nucleotide in the backtracked state (C, blue dashed line) or reduce the energy of the transition states/intermediates leading to that state (C, purple dashed line), consequently promoting RNAP passage into the 1-nt backtracked register. The stabilization of the 1-nt backtracked state in turn eases the path of RNAP to subsequent, multinucleotide backtracked registers.