Quantifying the impact of initial RNA primer length on nucleotide addition by RNA polymerase I

Abstract

Transient state kinetic studies of eukaryotic DNA-dependent RNA polymerases (Pols) in vitro provide quantitative characterization of enzyme activity at the level of individual nucleotide addition events. Previous work revealed heterogeneity in the rate constants governing nucleotide addition by yeast RNA polymerase I (Pol I) for each position on a template DNA. In contrast, the rate constants that described nucleotide addition by yeast RNA polymerase II (Pol II) were more homogeneous. This observation led to the question, what drives the variability of rate constants governing RNA synthesis by Pol I? Are the kinetics of nucleotide addition dictated by the position of the nascent RNA within the polymerase or by the identity of the next encoded nucleotide? In this study, we examine the impact of nucleotide position (i.e. nascent RNA primer length) on the rate constants governing nine sequential nucleotide addition events catalyzed by Pol I. The results reveal a conserved trend in the observed rate constants at each position for all primer lengths used, and highlight that the 9-nucleotide, or 9-mer, RNA primer provides the fastest observed rate constants. These findings suggest that the observed heterogeneity of rate constants for RNA synthesis by Pol I in vitro is driven primarily by the template sequence.

1. Introduction

Regulation of gene expression is essential for survival of organisms from all domains of life. A key target for gene regulation is synthesis of RNA from a DNA template. This transcription of the DNA is catalyzed by DNA-dependent RNA polymerases (Pols) [1]. Transcription occurs in three phases: 1) an initiation phase, during which the polymerase is recruited to a target gene locus, binds to the DNA, locally unwinds the DNA, and synthesizes a short RNA primer de novo; 2) an elongation phase, during which the polymerase escapes the promoter region and synthesizes a nascent RNA species complementary to the template DNA sequence; and 3) a termination phase, during which the polymerase dissociates from the DNA and the synthesized RNA transcript is released from the Pol. The multitude of steps in each phase can each be targeted for regulation, resulting in finely tuned control of RNA synthesis [1–3].

The elongation phase of transcription minimally requires: ribonucleoside triphosphate binding, chemical bond formation, pyrophosphate release, and translocation [4–7]. To define the kinetics of transcription elongation, our lab and others have utilized a 9-nucleotide, or 9-mer, RNA primer hybridized to a DNA template (RNA:DNA) as the scaffold for promoter-independent in vitro transcription experiments [2,3,8–16]. The 9-mer RNA primer has been shown to be the optimum length to achieve robust elongation complex (EC) assembly for E. coli RNAP and yeast Pol II in scaffold-based in vitro transcription experiments [2,3,14]. For the experiments presented in this study, the 9-mer is annealed to a complementary DNA template strand, incubated with Pol I, then addition of the fully complementary DNA non-template strand is added to complete formation of an active EC. The first nucleotide we provide is α-³²P-CTP, which forms the visualizable radiolabeled 10-mer RNA. We then provide ATP and GTP in solution, this allows for extension of the 10-mer RNA by nine sequential nucleotide addition events to form the final 19-mer RNA product.

We collect discrete reactions using a chemical quench-flow. This instrument rapidly mixes two solutions, incubates the reaction for a user-determined period of time, and rapidly quenches the reaction with a third solution. In this study, we mix the radiolabeled ECs with a solution of NTPs, magnesium, and heparin. Heparin serves as a trap to sequester Pol I enzymes unbound to the RNA:DNA hybrid, ensuring single turnover reaction conditions [8,13]. The reactions are quenched with 1 M hydrochloric acid (HCl), resulting in termination of the elongation reaction [13]. The EC and NTP solutions are rapidly mixed in <2 milliseconds, and time points are collected as Δt ≥ 5 milliseconds. The final 19-mer RNA product has the sequence 5′-AUCGAGAGGCAGGAGGGAA-3′ [8,11]. The next encoded nucleotide is U, and UTP is omitted from the reaction preventing further extension.

For transcription elongation experiments reporting on nine sequential nucleotide addition events, it was observed that the rate constants governing each nucleotide addition varied widely for Pol I, but not Pol II [8,16]. This observation led to the question: is the rate of RNA synthesis by Pol I influenced by the identity of the next encoded nucleotide or by the position of the 5′ end of the nascent RNA within the Pol, potentially through interactions between the RNA and the exit channel of the Pol?

To test whether the length of the initial RNA primer influences the rate constants governing downstream nucleotide addition events, we altered the length of the RNA primer by increasing or decreasing the sequence by one nucleotide. This resulted in the 8-mer, referred to as (n−1), 9-mer, referred to as (n), and 10-mer, referred to as (n + 1), RNA primers. The downstream template and non-template DNA sequences were unaltered. We evaluated the rate of RNA synthesis and the stability of the elongation complexes for each RNA primer. If interactions between the end of the RNA and the polymerase were driving the observed heterogeneity, one would observe shifts in the patterns of fast versus slow incorporation events in response to primer length. On the contrary, these results demonstrate a conserved trend in the rate constants governing nine sequential nucleotide addition events and reveal that the 9-mer (n) RNA primer results in the fastest observed rate constants. Differences in elongation complex stability were also observed; ECs formed with the shortest initial RNA primer (n−1) were less stable over time as compared to ECs formed with (n) and (n + 1). These results lead to the conclusion that heterogeneous nucleotide addition by Pol I is influenced slightly by the position of the RNA primer within the polymerase, but primarily results from the identity of the next encoded nucleotide.

2. Materials and methods

2.1. Buffers

In vitro time course studies were performed in reaction Buffer A: (40 mM KCl, 20 mM Tris-Acetate (OAc) pH 8.39 at 25 °C, 2 mM dithiothreitol (DTT), 0.2 mg·mL⁻¹ bovine serum albumin (BSA)); prepared immediately before use from stocks filter sterilized through 0.22 μm Millipore vacuum-driven filters (EMD Millipore, Billerica, MD).

2.2. Protein purification

Pol I was purified from Saccharomyces cerevisiae as previously described [8,13].

2.3. Oligonucleotides

All template and non-template oligonucleotides and primer oligoribonucleotides were purchased from Integrated DNA Technologies (Coralville, IA), (Supplemental Data Table 1). Oligonucleotides were gel purified, both oligonucleotides and oligoribonucleotides were dialyzed against Buffer A [10,13].

2.4. Rapid mixing time courses

Purified Pol I was incubated with pre-annealed RNA:DNA hybrids before adding the non-template DNA to form active ECs in Buffer A. The ECs were then radiolabeled for visualization by the addition of α-³²P-CTP and Mg²⁺ at ambient temperature. The radiolabeling reaction was halted with the addition of ethylenediaminetetraacetic acid (EDTA). Labeled ECs were loaded opposite a solution containing 1 mM ATP, 1 mM GTP, magnesium, and heparin, in syringes on the left and right ports of the KinTek chemical quench-flow [KinTek Corporation, Snow Shoe, PA]. The two solutions were rapidly mixed and the reaction proceeded for pre-determined time points before being quenched with 1 M HCl. The resultant sample was neutralized then mixed with formamide dye. RNA species were separated via polyacrylamide gel electrophoresis (Supplemental Data Fig. 1), then quantified using phosphorimage analysis [13].

2.5. EC stability time courses

Pol I EC stability assays for each RNA hybrid primer were performed as described previously [8]. In brief, labeled ECs were subjected to a solution of 750 mM KCl and RNase A. Formation of cleavage product over time was quantified to determine EC collapse over time.

2.6. Data analysis

At least three time courses of each RNA primer length were collected and quantified using ImageQuant (Cytiva, Marlborough, MA). The normalized signal intensities for each time point were first plotted using KaleidaGraph (Synergy Software, Reading, PA). Parameter value determinations and calculations of observed rate constants were determined for each time course using the custom-built MatLab toolbox MENOTR (Multi-start Evolutionary Nonlinear OpTimizeR) a hybrid nonlinear least squares (NLLS) and genetic algorithm analysis tool as described in detail previously [8,12,13].

3. Results and discussion

Transient state kinetic analyses of Pol I transcription elongation have been previously described by our group utilizing a 9-nucleotide, or 9-mer, RNA primer with the sequence 5’–AUCGAGAGG–3′ [8–13]. From these studies, heterogeneity of the rate constants of nucleotide addition by Pol I was observed [8,11]. Each intermediate, from the formation of the 11-mer to the final 19-mer RNA product are described by rate constants ranging from (26 ± 8) s⁻¹ to (200 ± 20) s⁻¹ [8]. Using the same experimental design, this heterogeneity was not as pronounced for the closely related enzyme Pol II [8].

We hypothesized that the observed heterogeneity of rate constants was due to either interactions between the 5′ end of the RNA and the exit channel of the Pol, or intrinsic sensitivity of Pol I to the template sequence. To differentiate between these possibilities, we altered the initial RNA primer length but left the downstream template DNA sequence unchanged. This resulted in three initial RNA primer lengths: 8-mer, referred to as (n−1), the original 9-mer, referred to as (n), and 10-mer, referred to as (n + 1). The downstream template and non-template DNA sequences were identical in each scaffold, only the length of the initial RNA primer was varied (Fig. 1A).

Fig. 1.

Alterations to initial RNA primer length for in vitro transcription catalyzed by Pol I. A) Illustration of in vitro RNA primer templates for RNA synthesis catalyzed by Pol I. B) Representative sequencing gel of nine nucleotide addition events. C) Overlay of each RNA intermediate from the +1-mer to the +9-mer. Plotted points represent the average of at least three experimental replicates, the error bars represent the calculated standard deviation.

During the elongation reaction, each nucleotide addition forms an intermediate RNA species of a defined length. These intermediate RNA species, as well as the final RNA product, are visualized using sequencing gels and phosphorimaging of the α-³²P-CTP-labeled RNA products (Fig. 1B). The intensity of each RNA species is quantified and normalized to the initial starting material to determine the normalized fraction of RNA comprising each RNA species [13]. The appearance and disappearance of the RNA intermediates and appearance of the final RNA product are initially plotted for qualitative evaluation (Fig. 1C).

To determine the rate constants governing the formation and disappearance of each RNA intermediate, and formation of the final RNA product, we use the MATLAB tool MENOTR to simultaneously describe all nine time courses by a given reaction mechanism [8,11–13]. These model-dependent analyses are performed using the kinetic mechanisms described by Schemes 1–3 [8,11] (Fig. 2A). Schemes 1–3 are the same kinetic mechanism but define the reactions for (n−1), (n), and (n + 1), respectively (Fig. 2A).

Fig. 2.

Identity of each intermediate RNA species by nucleotide position. A) Kinetic mechanisms describing the formation of each intermediate RNA species following incorporation of the radiolabeled nucleotide. Included is k_obs,10 describing the intrinsic nuclease activity of Pol I resulting in the formation of a dimer species and the +7-mer RNA intermediate. B) Plotted increase and decrease of each RNA intermediate by nucleotide position overlaid for three initial RNA primer lengths. C) Plot of the observed rate constants (s⁻¹) vs. nucleotide position of three initial RNA primer lengths. Plotted points represent the average of three experimental replicates, error bars represent the calculated standard deviation.

We found that the rate constants for all three primer lengths exhibited similar trends for each RNA intermediate (+1-mer through +9-mer) (Fig. 2B, C). Interestingly, the second incorporation is always faster than the remaining addition events. Moreover, the second incorporation for the (n) primer exhibits the fasted observed rate constant (Fig. 2C, Table 1). We also observe intrinsic nuclease activity catalyzed by Pol I, resulting in cleavage of a dimer species from the 3′ end of the RNA [8,10] (Fig. 2A). The inclusion of the rate constant k_{obs,10,cleavage} is required to describe the dinucleotide cleavage step [8] (Table 1). Evidence of dinucleotide cleavage from the 3’end of the nascent RNA is revealed by the remaining signal intensity of the +7-mer and + 8-mer RNA intermediates (Fig. 2B). When Pol I halts at +9 due to omission of UTP from the reaction, the enzyme backtracks and cleaves a dinucleotide from the 3′ end of the nascent RNA. This results in +7-mer nascent RNA which is then extended again, resulting in steady-state abundance of +7-mer and + 8-mer (Fig. 2B).

Table 1.

Calculated rate constants for k_obs,1 through k_obs,10.

Observed Rate Constant (s−1)	(n−1)	(n)	(n + 1)
kobs,1	75 ± 13	133 ± 46	92 ± 24
kobs,2	172 ± 22	292 ± 20	141 ± 3
kobs,3	83 ± 6	83 ± 2	105 ± 21
kobs,4	32 ± 8	49 ± 2	44 ± 8
kobs,5	82 ± 9	142 ± 11	70 ± 6
kobs,6	72 ± 12	137 ± 11	80 ± 11
kobs,7	37 ± 12	53 ± 3	29 ± 9
kobs,8	137 ± 42	126 ± 30	105 ± 36
kobs,9	87 ± 22	89 ± 19	80 ± 21
kobs,10,cleavage	13 ±5	26 ± 8	27 ± 13

Average calculated rate constants governing each nucleotide incorporation for at least three experimental replicates.

While all three initial RNA primer lengths exhibited similar trends based upon the intermediate RNA species formed, the amplitude of signal intensity and maximum observed rate constants differed at some positions for each primer length. Under these single turnover conditions, the amplitude is directly proportional to the fraction of polymerase bound in an active elongation complex at time zero. Thus, the trend in the amplitude can be interpreted as proportional to the strength of the binding interaction between the RNA:DNA hybrid and Pol I. This observation implies that Pol I binds with the highest affinity to the RNA: DNA hybrid containing the (n) primer, followed by (n−1) and (n + 1) (Fig. 2B, C, Supplemental Data Fig. 1).

Based upon the observation that the amplitudes decrease as (n) > (n−1) > (n + 1), which is consistent with K_d (n) < K_d (n−1) < K_d (n + 1), we hypothesized that altering the initial RNA primer length may also result in decreased stability of (n−1) and (n + 1) ECs, relative to ECs formed with the (n) primer. To test if initial RNA primer length impacts the stability of ECs formed by Pol I, we utilized an RNase A protection experiment [8]. This experiment begins with the same EC formation steps as previously described in this study, including incorporation of α-³²P-CTP to form visualizable RNA products. We then subject the ECs to a destabilizing salt solution and RNase A [8]. Under these conditions, if the elongation complex is intact when the sample is collected, RNA is protected from RNase activity and a full length species is observed. However, if the complex has collapsed by the time the sample is collected, RNase A will cleave the RNA at a specific sequence and shorter ³²P-RNA species are observed. Based upon the RNA primer sequence, (n−1) and (n) primers have a single cleavage site. The (n + 1) primer has two cleavage sites. Representative sequencing gels and illustrations of RNase A cleavage sites are presented in Fig. 3A.

Fig. 3.

Stability of Pol I ECs formed with three initial RNA primer lengths. A) Representative gel images of EC stability time courses. The position of RNase A cleavage is indicated by vertical bars. B) Plot of normalized EC stability for each initial RNA primer length over time. Plotted points represent that average of three experimental replicates, error bars represent calculated standard deviation.

Consistent with our hypothesis, the Pol I (n−1) ECs demonstrate a decreased stability over time (Fig. 3B). This agrees with previous work detailing the stability of ECs formed by yeast Pol II and E. coli RNAP [3,14,15]. We observed that Pol I ECs formed with (n) and (n + 1) primers demonstrate nearly identical stability over time (Fig. 3B). These results suggest (n−1) may be the minimal primer length for efficient Pol I EC assembly in vitro. This finding agrees with the minimal primer length required for in vitro assembly of yeast Pol II and E. coli RNAP ECs [3,14,15]. The results also suggest that (n) and (n + 1) are more effective for stable Pol I EC assembly than (n−1), but nucleotide addition is faster when Pol I is complexed with the (n) primer.

In these studies, we pre-bind Pol I to the RNA:DNA hybrid and do not directly measure the binding equilibrium. However, the dissociation equilibrium constant for Pol I binding to the RNA:DNA hybrid is the ratio of the dissociation rate constant, k₂, and the bimolecular association rate constant, k₁. Thus, K_d = k₂/k₁. The trend in the observed amplitudes of the time courses implies that the K_d is increased as K_d (n) < K_d (n−1) < K_d (n + 1). One might anticipate the stability measurement to exhibit the same trend. However, the (n) and (n + 1) primers exhibit similar stability and (n−1) exhibits reduced stability. As designed, the stability measurement is proportional to the dissociation rate constant k₂. Therefore, the lack of an observed effect implies that the association rate constant, k₁, is reduced for Pol I binding the (n−1) primer.

4. Conclusions

Using the same experimental design, our previous in vitro transcription elongation studies revealed that Pol I demonstrates heterogeneity in the rate constants governing multiple nucleotide addition events, while nucleotide addition by Pol II exhibited less heterogeneity [8]. We hypothesized that this observation was a result of either the DNA template sequence or the position of the nascent RNA as it approaches the exit channel of the polymerase. To test if the position (i.e. length) of the nascent RNA altered the rate constants governing nine nucleotide addition events catalyzed by Pol I, we increased, (n + 1), and decreased, (n−1), the length of the initial RNA primer by one nucleotide.

Interestingly, we observed a conserved trend in the heterogeneity of rate constants for each initial RNA primer length. We recapitulated the previously observed trend for the (n) primer; a fast first addition of AMP followed by a 1.5 to 2.2-fold faster addition of GMP [8]. The rate constants of all subsequent nucleotide additions were slower than the second incorporation. We observe this same trend for both the (n−1) and (n + 1) primers (Fig. 2C, Table 1). For (n−1), there was a 2.3-fold faster second nucleotide addition event, while for (n + 1), we observed a 1.5-fold faster second nucleotide addition event as compared to the maximum observed rate constant for the first nucleotide addition (Table 1). All subsequent nucleotide addition events (+3 through +9) were slower than the (+2) position (Fig. 2C, Table 1).

This conserved trend of fast first addition of AMP, even faster second addition of GMP, followed by additions of AMP and GMP which are slower than the second addition, suggest that the heterogeneous rate constants governing nucleotide addition by Pol I are dictated by the sequence of the DNA template. Future studies are required to determine the sequence features that govern fast versus slow incorporation during processive transcription elongation.

While the eukaryotic enzymes Pol I and Pol II are closely related, the chromatin environments each enzyme encounters in vivo are significantly different. Pol I solely synthesizes ribosomal RNA (rRNA) from the largely nucleosome-free ribosomal DNA (rDNA) [17–21]. In contrast, Pol II synthesizes RNA from nucleosome-bound genes, encountering different chromatin environments across the genome [17–21]. The different chromatin environments, transcriptional demands, and associations with different trans-acting factors may have contributed to Pol I and Pol II responding differently to the same template DNA and RNA primer sequences, as we have previously observed in vitro [8].

Structural studies have revealed that 6 to 10 bases of nascent RNA form interactions with subunits in the active site of multi-subunit RNAPs [7,18,19,22,23]. Previous biochemical studies revealed the impact of nucleotide concentration and chemical environment on the kinetics of RNA synthesis catalyzed by yeast Pol I and E. coli RNAP [8–11,17,24,25]. With consideration of the conserved structural features but unique functional features of Pol I, we evaluated the impact of nucleotide position within the polymerase on the rate constants governing elongation. We determined that the length of the initial RNA primer does impact the rate constants governing nucleotide addition, with the 9-mer, (n), RNA primer resulting in the fastest observed rate constants. We also observed a similar trend in relative rate constants based upon nucleotide position for each RNA primer length. This observation suggests that the DNA template sequence drives the heterogeneity of rate constants, rather than interactions between the end of the RNA and subunits of Pol I.

Previous in vitro analyses using E. coli RNAP and yeast Pol II have shown the minimum stable RNA primer length to be 8 nucleotides, and the most stable primer to be 9 nucleotides [3,14,15]. Our results with Pol I (n) ECs agree with these previous findings, however, we also observed similarly stable ECs of Pol I bound to the 10-mer, (n + 1), RNA primer. We attempted to form an EC of Pol I bound to a 14-mer RNA primer but were unable to observe processive transcription elongation, likely due to Pol I lacking the ability to unwind the RNA:DNA hybrid after exceeding a certain length. We also observed a decrease in the stability of Pol I 8-mer, (n−1), ECs. It has been previously shown that to form transcriptionally active E. coli RNAP ECs in vitro, the minimal length of RNA in the RNA:DNA hybrid is 8 nucleotides [3,14]. Our data suggest that the same minimal length requirement may apply to Pol I as well.

We hypothesized that the variability observed for Pol I transcription elongation kinetics could be due to interactions of the nascent RNA approaching the exit channel of the polymerase. While previous work by our lab determined that altering the 5’end of the RNA to be either 5′-triphosphate or 5′-hydroxyl had no observable effect on nucleotide addition rate, the impact of RNA primer length had yet to be evaluated for in vitro transcription catalyzed by Pol I [26]. The results of this study suggest that Pol I elongation kinetics are impacted by the position of the nascent RNA within the polymerase but the heterogeneity in the observed rate constants is a result of the identity of the nucleotide being incorporated. The sequence-specific influence on nucleotide addition by Pol I suggests that the rDNA sequence may impact the kinetics of rRNA synthesis and processing events [27]. We hypothesize that these processes have co-evolved to be tightly regulated and highly efficient. Interestingly, nucleotide addition by Pol II does not demonstrate such sequence-specific regulation using the (n) RNA primer [8]. The impact of the (n−1) and (n + 1) RNA primers on nucleotide addition by Pol II remains to be measured.

Synthesis of rRNA by Pol I is the first and rate-limiting step in ribosome biosynthesis, a process that is dysregulated in cancer cells and implicated in tumorigenesis [27–32]. In recent years, Pol I has emerged as a putative target for anti-cancer therapeutics, while Pol II remains an “undruggable” target [28–32]. Characterization of the enzymatic properties of Pol I, and the closely related enzymes Pol II and Pol III, is critical to inform rational drug design. This work has revealed that Pol I is responsive to both the length of the initial RNA primer as well as sequence elements. Perhaps, because of this, Pol I may be more vulnerable to substrate analogues. The scaffold-based in vitro transcription experiment described in this study has recently been used to determine the specificity of a novel small molecule inhibitor, BMH-21, for effectively targeting Pol I, having an inhibitory effect on Pol III, and no detectable effect on RNA synthesis by Pol II [28]. Additional targeting strategies may emerge in the future, exploiting the fundamental properties revealed in studies such as these.