Maturation of 6S regulatory RNA to a highly elongated structure

Abstract

As bacterial populations leave the exponential growth phase and enter the stationary phase, their patterns of gene expression undergo marked changes. A key effector of this change is 6S RNA, which is a highly conserved regulatory RNA that impedes the transcription of genes associated with exponential growth by forming an inactivating ternary complex with RNA polymerase and sigma factor σ⁷⁰ (σ⁷⁰–RNAP). In Escherichia coli, the endoribonuclease RNase E generates 6S RNA by specific cleavage of a precursor that is nearly twice the size of the 58 kDa mature form. We have explored recognition of the precursor by RNase E, and observed that processing is inhibited under conditions of excess substrate. This finding supports a largely distributive mechanism, meaning that each round of catalysis results in enzyme dissociation and re-binding to the substrate. We show that the precursor molecule and the mature 6S share a structural core dominated by an A-type helix, indicating that processing is not accompanied by extensive refolding. Both precursor and mature forms of 6S have a highly stable secondary structure, adopt an elongated shape, and show the potential to form dimers under specific conditions; nonetheless, 6S has a high structural plasticity that probably enables it to be structurally adapted upon binding to its cognate protein partners. Analysis of the 6S-σ⁷⁰–RNAP complex by native mass spectrometry reveals a stable association with a stoichiometry of 1 : 1 : 1. A theoretical 3D model of mature 6S is presented, which is consistent with the experimental data and supports a previously proposed structure with a small stem-loop inside the central bubble.

Introduction

Non-coding regulatory RNAs play diverse roles in the control of gene expression in all domains of life, and considerably boost the potential complexity of genetic regulatory networks. Bacterial non-coding RNAs may be broadly categorized into three groups according to activity and mechanism. One group comprises trans-encoded small regulatory RNAs (sRNAs), which modulate the translation or stability of specific transcripts with which they share imperfect sequence complementarity [1,2]. In Escherichia coli and related proteobacterial species, close to 100 such sRNAs have been identified to date, and those RNAs are typically 50–300 nt long. A second group of bacterial RNAs are cis-encoded, and act as antisense regulators of their target transcripts, with which they share perfect sequence complementarity. Important examples of cis-regulating RNA elements in bacteria are the RNA thermometers and riboswitches [3–6].

While many of the regulatory RNAs characterized to date act post-transcriptionally, the non-coding 6S RNA represents the third group of bacterial non-coding RNAs, which function by impeding transcription [7,8]. The 183 nt 6S RNA was identified more than four decades ago as a highly abundant RNA in E. coli [9], and was subsequently found to be a component of a cellular nucleoprotein complex [10,11] with RNA polymerase (RNAP) [12]. Further studies have revealed that 6S specifically targets the holoenzyme assembly comprising RNA polymerase and sigma factor σ⁷⁰ (σ⁷⁰–RNAP). By forming a stable complex with σ⁷⁰–RNAP, 6S RNA impedes transcription from σ⁷⁰-responsive promoters, and enables the alternative σ^S-dependent transcription of stationary phase-expressed genes [13–16]. Recent findings indicate that 6S also influences the levels of the stringent response signalling molecule guanosine tetraphosphate (ppGpp), with effects on the networks of stress responses and growth adaptation [17,18].

In E. coli and other proteobacteria, 6S is transcribed from the ssrS gene [19]. There are two alternative promoters for this gene: a proximal site dependent on σ⁷⁰ (P1) and a distal site dependent on σ^S and σ⁷⁰ (P2). Thus, two primary transcripts are produced: a long precursor molecule of 404 nt and a short precursor molecule of 194 nt [20,21]. Both precursors are processed at their 5′ end to generate the 183 nt mature functional form of 6S RNA (hereafter referred to as ‘mature 6S’ or ‘mature form’). Ribonuclease RNase E is responsible for maturation of the long precursor, while both RNase E and its paralogue RNase G are able to process the shorter transcript [19,20,22].

The predicted secondary structure of the long 6S precursor (hereafter referred to as ‘precursor 6S’ or ‘precursor form’) is shown in Fig. 1A, which also indicates the site where RNase E cleaves to generate the mature 6S (nucleotide A222). The secondary structure of mature 6S has been experimentally determined to be principally duplex with an extended hairpin and three main domains: a closing stem, a central bubble with an extended single-stranded region, and a terminal loop [23,24]. Formation of a small stem-loop inside the central bubble was demonstrated [24].

Fig. 1. Secondary structure and quality of 6S preparations.

(A) The predicted secondary structure of precursor 6S RNA. The molecule is processed by endoribonuclease RNase E at A222 (indicated by an arrow) to produce the mature 6S molecule. (B) Urea/polyacrylamide gel of mature and precursor 6S samples prepared by in vitro transcription and gel purification. Lane M, molecular weight markers (RiboRuler low- range RNA ladder, Fermentas, Waltham, MA, USA). (C) Size distribution for mature (red) and precursor (green) 6S RNA determined by dynamic light scattering. The data were obtained at room temperature in a buffer comprising 10 mm Tris/HCl, pH 8.0, 5 mm MgCl₂, 100 mm KCl. (D) Nanoflow electrospray ionization mass spectra of the 6S-σ⁷⁰–RNAP complex, suggesting a binding stoichiometry of 1 : 1 : 1 for 6S RNA : σ⁷⁰ : RNAP. The RNAP complex without added 6S RNA (lower trace) has a mass of 457 292 ± 41 Da (measured using charge state peak fronts), in good agreement with a calculated mass of 459 297 Da (based on the sum of the measured individual subunit masses for a stoichiometry of α₂ßß’ωσ⁷⁰), while the complex in the presence of 6S RNA (upper trace) yields a measured mass of 517 649 ± 86 Da. The mass difference between the two complexes, 60 357 Da, is consistent with binding of 6S RNA (measured mass from charge state peak fronts: 59 667 ± 6 Da, spectrum not shown). The samples are dissolved in 250 mm ammonium acetate.

The role of 6S in transcription regulation depends on its ability to specifically recognize and interact with the σ⁷⁰–RNAP complex with high affinity, so as to suppress transcription initiation from σ⁷⁰-dependent promoters. The single-stranded region of the molecule is proposed to mimic a transcription bubble in the quaternary assembly of the 6S-σ⁷⁰–RNAP complex, accounting for the inhibitory effect of 6S on the polymerase holoenzyme [23,24]. Moreover, 6S not only blocks the active site of RNAP [25], but also adopts an active conformation that enables it to be used as template for de novo transcription when the cellular growth conditions become favorable [14,26]. When this occurs, the polymerase reads through the single-stranded region in the central bubble to destabilize the 6S fold and cause its release from the holoenzyme [25,27–29]. At least for E. coli, it has been proposed that release of 6S from RNAP is orchestrated by a hairpin inside the central bubble [29,30]. The segments of 6S that bind RNAP and the portion of 6S that is transcribed for de novo RNA synthesis have been identified by several studies [12,27,31]. Moreover, obtaining the crystal structure of the E. coli σ⁷⁰– RNAP complex [32] has enabled modelling of the hypothetical structure of its ternary complex with 6S RNA [31].

Although much work has focused on determination of the secondary structure of 6S using biochemical assays, the biophysical parameters of the overall 3D structure of this RNA remain unknown. Crystallization attempts with Aquifex aeolicus 6S RNA have only resulted in the crystal structure of a small 12 nt nucle-olytic fragment that is part of the closing stem [33]. Here, we have studied the maturation of E. coli 6S RNA, and characterize both the mature and precursor molecules in biophysical terms. We show that processing of the precursor by RNase E is inhibited under conditions of excess substrate, which supports the existence of a distributive versus a processive mechanism. We confirm that the highly conserved secondary structure of 6S is also highly stable, and we show that the overall structure of both forms demonstrate high plasticity in response to moderate changes in temperature and solvent constituents. Both molecules are highly elongated, have an A-type helix in common, and have the potential to form dimers under our experimental conditions. Our sample efficiently binds σ⁷⁰–RNAP with a stoichiometry of 1:1:1 (6S:σ⁷⁰:RNAP). Moreover, we present a theoretical 3D model of mature 6S that is consistent with the experimental data. This model favours the presence of a small stem-loop inside the central bubble in a dimeric association of 6S molecules. Further studies are required to fully explore whether the dimerization observed in this work is relevant to the biological activity of 6S RNA.

Results

Quality of RNA preparations

The preparation of mature and precursor 6S RNAs is described in Experimental procedures. The purity of the samples is shown in Fig. 1B. Structural heterogeneities due to the annealing/refolding procedure were checked and found to be minimal (Fig. 2A). The polydispersity index was estimated by dynamic light scattering (DLS) (Fig. 1C) to be 19 ± 3% for mature 6S and 18 ± 1% for the precursor molecule at 20 °C, indicating good homogeneity for both RNA preparations.

Fig. 2. Native gel electrophoresis of 6S samples.

(A) Native gel electrophoresis of 6S and comparison with size markers. Various amounts (25, 50, 100, 200 and 400 ng, lanes 1-5 respectively) from a mature 6S sample (100 μg·mL⁻¹) were loaded onto a 5% polyacrylamide 0.5x TBE (50 mm Tris, 50 mm boric acid, 1 mm EDTA), 5% glycerol native gel. Lanes M1 and M2 contain molecular weight markers: RiboRuler low-range RNA ladder (Thermo Fisher Scientific, Waltham, MA, USA) and 100 bp DNA ladder (New England BioLabs, Hitchin, UK), respectively. The 6S sample ran as a sharp, well-defined band with only minor smearing, which indicates a well-folded, structured sample and high structural homogeneity. Comparison of the 6S band with the markers shows that no safe conclusions regarding for the size and oligomerization of the molecule may be drawn on the basis of native electrophoresis. This usually occurs because the secondary structure of RNA alters its migration pattern in native gels such that it does not migrate according to its true size. (B) Gel mobility shift assay of 6S RNA with core and holo RNAP. A constant amount of mature 6S RNA sample was incubated with various amounts of protein under conditions appropriate for binding, i.e. 50 mM HEPES, pH 8.0, 150 mM KCl, 5 mM MgCl₂, 5% glycerol, 1 mM dithiothreitol. After 30 min incubation at room temperature, the samples were loaded onto a 0.8% native agarose gel, and electrophoresis was performed in the cold room in 0.5x TAE (20 mM Tris, 10 mM acetic acid, 0.5 mM EDTA). The gel was stained with SYBR Gold. 6S RNA (0.5 μM) was incubated in the absence of protein (lane 1) or titrated with 0.25, 0.5 and 1 μM core RNAP (lanes 2, 4 and 6, respectively) or 0.25, 0.5 and 1 μM σ⁷⁰–RNAP (lanes 3, 5 and 7, respectively).

Confirmation that the mature 6S has been properly folded in its biologically relevant form came from its ability to bind the RNA polymerase holoenzyme (σ⁷⁰–RNAP). The association was monitored and confirmed by gel-shift assays (Fig. 2B), and further analysed by soft ionization electrospray mass spectrometry under conditions that preserve macromolecular complexes. The mass estimates are consistent with a 1 : 1 : 1 complex of 6S RNA:σ⁷⁰:RNAP (Fig. 1D). Moreover, RNase E was active on the precursor 6S, confirming that the RNA folds to a biologically relevant conformation.

Precursor 6S processing by RNase E

It has been shown that 5′ maturation of precursor 6S by the RNase E to generate mature 6S is achieved by predominant digestion at position A222 (Fig. 1A) and several minor digestions at the 5^’ end of the molecule [20]. It has also been suggested that cleavage at the minor positions occurs only after the initial digestion at position 222. To further explore the characteristics of this processing, we performed time-course cleavage assays of the recombinant precursor 6S RNA by the N-terminal catalytic domain of RNase E (NTD/RNase E) [34,35]. As shown in Fig. 3A, the recombinant precursor 6S is recognized and accurately processed by NTD/RNase E, and two main bands appear during the early stages of the reaction. One of them runs at the size of mature 6S species (indicated by an asterisk) and accumulates over the time course, while the other migrates slightly higher than the 200 nt marker band (indicated by a cross), and is likely to be the 5′ 221 nt fragment remaining from cleavage at A222. In contrast to the mature 6S fragment, the second fragment does not accumulate over the course of the digestion, but instead appears to be degraded. As processing progresses, many smaller bands appear, and their intensity grows stronger as the intensity of the 5′ 221 nt fragment fades, indicating that these smaller fragments are the products of degradation of the liberated 5′ 221 nt fragment. These data suggest the course of events comprises a single initial digestion that produces the mature 6S, followed by multi-site degradation of the remaining 5′ 221 nt fragment.

Fig. 3. Recognition and processing of 6S precursor by the RNase E catalytic domain.

(A) Representative time-course experiment for processing of precursor 6S by NTD/RNase E. The products were resolved on a denaturing gel. Lane M, molecular weight markers (RiboRuler low-range RNA ladder; Fermentas). A control sample of RNA, mixed with all components of the reaction except NTD/RNase E and incubated for 120 min at 37 °C, was also resolved (lane I). (B) Dependence of initial rate on the substrate concentration as revealed by fixed-time assays. These experiments were performed varying the substrate concentration, keeping constant the enzyme concentration (1 μM), and measuring the burst rate during the first minute of the reaction. Values are mean ± SD of at least four independent measurements. (C) Quantification of the time-course experiments, showing the percentage of precursor 6S reduction as a function of the time of reaction. Non-linear least-squared data fitting was performed as described by Kemmer and Keller [67]. Red symbols represent 0.1 μM substrate; orange symbols represent 0.3 μM substrate; yellow symbols represent 0.5 μM substrate; green symbols represent 1.0 μM substrate; light blue symbols represent 1.5 μM substrate; blue symbols represent 2 μM substrate; purple symbols represent 3 μM substrate.

To evaluate how this multi-processing affects the rate of precursor 6S decay, we performed time-course experiments for cleavage of precursor 6S RNA at various substrate concentrations, using a fixed amount of the RNase E catalytic domain. The initial rate at which precursor 6S is consumed shows a strong dependence on substrate concentration, and the enzyme was inhibited at high substrate concentrations (Fig. 3B). To further characterize this behaviour, we performed fixed-time assays, varying the substrate concentration while keeping the enzyme concentration constant and measuring the burst rate during the first minute of the reaction (Fig. 3C). The Lineweaver-Burke plot for these data (not shown) confirmed substrate inhibition of the reaction.

To discover the reason for this inhibition, we next assessed whether the precursor or mature 6S (substrate and product) formed an inhibitory complex with the RNase E. To explore this, binding of the mature and precursor RNA molecules by the catalytic domain of RNase E was investigated using a catalytically inactive double mutant of NTD/RNase E (D303R/D346R). Gel electrophoretic mobility shift assays indicated binding of both precursor and mature 6S to the RNase E catalytic domain only at high concentrations of the enzyme. Figure 4A shows a titration experiment for mature 6S with increasing concentrations of mutated NTD/RNase E (D303R/D346R). A substantial shift of the RNA band is only evident in the micromolar affinity range when the ratio of the protomers to RNA is as high as 40:1. Similar results were obtained when a sample of precursor 6S was titrated with the protein (Fig. 4B). Furthermore, we observe 6S RNA retardation shifts upon increasing protein concentrations, indicating that multiple NTD/RNase E molecules bind to the 6S molecules with varied affinities. This finding supports previous suggestions that RNase E, which is a homotetramer, binds to large, structured RNAs with two or more protomers which could belong to the same or different tetramers [35,36]. Quantification of the gel mobility shift analysis based on disappearance of the unbound 6S molecules is shown in Fig. 4C. Sedimentation velocity experiments performed using samples containing a fixed amount of 6S RNA and increasing concentrations of protein protomers indicated that only a small fraction of the RNA population forms a complex with the protein, and only when the protein is present in great excess to RNA (data not shown). Nanoflow ESI mass spectrometry did not identify a stable complex of NTD/RNase E with the mature or precursor 6S (data not shown). Our experiments indicate that 6S forms a weak or transient complex with RNase E.

Fig. 4. Binding of the 6S RNA by an inactivated RNase E catalytic domain.

(A, B) Gel mobility shift assays of various 6S complexes. A constant amount of RNA sample was incubated with various amounts of protein. After 30 min incubation at room temperature, the samples were loaded on a native agarose gel and electrophoresis was performed in the cold room. The gels were stained with SYBR Gold, which binds specifically to nucleic acids. (A) mature 6S (228 nm) was titrated with 1.7, 2.6, 3.4, 6.8, 14, 20 and 27 μM of mutated NTD/RNase E (D303R/ D346R). (B) Precursor 6S (80 nm) was titrated with 1.7, 2.6, 3.4, 6.8, 14, 16 and 20 μM of mutated NTD/RNase E (D303R/D346R). (C) The fraction of mature 6S (red circles) or precursor 6S (green diamonds) that is shifted upon titration with inactive NTD/RNase E (D303R/D346R). Non-linear least squared data fitting was performed as described by Kemmer and Keller [67]. Values are mean ± SD of two experiments.

Thus, the inhibition suggested by the results shown in Fig. 3 cannot be attributed to formation of a stable complex between either the precursor 6S or the mature 6S with RNase E. We propose that this inhibition may be the result of a distributive mechanism of action of RNase E on the precursor 6S. At each cleavage site, for each round of catalysis, such a mechanism requires substrate binding, product(s) release and substrate rebinding for processing the next cleavage site. Thus, after each single RNA modification, the competition for processing by RNase E gradually increases as a growing number of potential substrates are derived. At relatively low concentrations of precursor 6S, a sufficient number of active sites are available to process all of the precursor and the generated by-products. At high concentrations of precursor 6S, the majority of active sites are occupied by the by-products, and the effective number of active sites available for processing of 6S is reduced, resulting in the apparent substrate inhibition.

Stability of 6S precursor and mature forms

The functional significance of a conserved secondary structure of mature 6S has been shown in previous studies [23,24] and discussed in recent reviews [11,37]. Here we attempt to biophysically characterize the 3D entity of the molecule in its precursor and mature forms, to elucidate features of the 3D structures and to probe possible changes upon transition from one form to the other.

Thermal stabilities of the purified, recombinant RNAs were evaluated by determining hyperchromic denaturation profiles [38]. The melting curve of mature 6S (Fig. 5A) is dominated by a sharp sigmoidal transition, indicative of cooperative unfolding, with a midtransition temperature (T_m) of 78 °C. Initiation of melting at a high temperature indicates a highly stable duplex and a lack of extensive tertiary interactions. In contrast to the mature form, denaturation of the precursor molecule (Fig. 5A) starts at a lower temperature, and the melting curve shows an almost monotonic increase of hyperchromicity. Given the homogeneity of the precursor preparation, much of the breadth of the melting curve may be assigned to sequential non-cooperative transitions of substructures. This is also supported by the extent of hyperchromicity (24%). The first derivative of the melting curve (Fig. 5A) confirms a peak at 79 °C, which coincides with the main peak of mature 6S and probably represents melting of their common domain.

Fig. 5. Folding stabilities and structure comparison of mature and precursor 6S.

(A) Thermal denaturation profiles of 6S RNA. The absorbance at 260 nm is recorded as a function of temperature for mature 6S (red circles) and precursor 6S (green circles). Inset panels: first derivatives of the thermal UV (upper graph) and thermal fluorescence (lower graph) melting curves for mature 6S (red line) and precursor 6S (green line), respectively. (B) Degradation of mature and precursor 6S. The degraded mixtures were analysed using denaturing urea/ polyacrylamide gels. Lanes 1 and 3 and lanes 2 and 4 are the mature and precursor 6S mixtures, respectively. A calibration curve was prepared using the markers on the gel (lanes M1 and M2) to calculate the lengths of the fragments. The dominant bands are indicated by numbers.

The thermal stability of 6S RNA was also explored using the dye SYBR Green, which intercalates between stacked bases and fluoresces in the RNA-bound state.

The spectra were recorded at 522 nm, which is the emission maximum of the dye complexed with doublestranded RNA [39], and first derivatives are presented in Fig. 5A. The melting profiles are noisy for all samples measured up to approximately 60 °C, and intercalating may destabilize the samples, giving rise to a peak at 64-65 °C. However, the shoulder at 74-75 °C is consistent with the UV melting data presented above, and the presence of a highly stable duplex in both molecules is confirmed. Our melting dissociation curves suggest that a highly stable, extensive double helix is the predominant structural feature of both forms of 6S. In the case of the precursor, this regular double helix is decorated by a complex net of lower-energy tertiary interactions.

Furthermore, during thermal denaturation, the samples undergo CD profile changes (data not shown) indicative of a helix-to-coil transition [40]. As the temperature increases, the positive band at 268 nm is red-shifted by 5 nm, while the negative band at 210 nm fades. The transition shows one clear isodichroic point at 285 nm, indicative of a simple two-state equilibrium without intermediates [41,42]; this is consistent with the UV thermal melting data of mature 6S, which also indicate a co-operative two-state transition of the duplex. The main difference in the melting profiles is seen from 220 to 250 nm. In this area, the precursor’s spectra show a pseudo-isodichroic point at approximately 247 nm, while those of mature molecule almost coincide in the full range from 225 to 240 nm. We conclude that the spectra of both molecules lack evidence for a complex higher-level architecture. Small but noticeable spectral deviations of the CD of the precursor at room temperature, as well as differences in the heating profiles, support the presence of extra structural features which may interact with the common double helix.

Differences with regard to the susceptibility to degradation/transesterification [43] of mature and precursor 6S molecules are shown in Fig. 5B. The most obvious feature is that bands 1 and 2, although common to both samples, are of much greater relative intensity in the mature 6S lane. These bands have an estimated size of approximately 140 and 125 nt, respectively, and, as they are common in both molecules, are most probably 3′ fragments produced by cleavage at the central bubble. This result implies that the central bubble is better protected in the precursor’s tertiary arrangement, and that the processing exposes this site and makes it available for interactions in the mature form.

6S is folded in an elongated dynamic conformation

The shape and size of the precursor and mature 6S RNA were evaluated by measurement of the sedimentation velocity using analytical ultracentrifugation. The mature molecule has a sedimentation coefficient of 5.8 ± 0.4 and a calculated frictional ratio (f/f ₀) of 2.0, implying an elongated molecular shape (Fig. 6A). A larger sedimentation coefficient (8.4 ± 0.4) was estimated for the precursor, with a frictional ratio of 2.2 (Fig. 6A). These results were reproducible using buffers of various ionic strengths.

Fig. 6. Analytical ultracentrifugation of mature and precursor 6S.

(A) Normalized sedimentation velocity profiles for mature 6S (red line) and precursor 6S (green line) in 25 mM NaCl, 25 mM KCl, 10 mM CaCl₂, 25 mM HEPES, pH 7.5. (B) s₂₀,_w as a function of MgCl₂ concentration (circles) and KCl concentration (diamonds) for mature 6S (red) and precursor 6S (green). The horizontal lines indicate the s₂₀,_w values for the fully folded and compact molecules of mature 6S (red line) and precursor 6S (green line). The error bars show the 95% confidence limits of each determination. Even a low concentration of MgCl₂ (2 mM) promotes a dramatic increase in the sedimentation coefficient of the mature molecule from 4.9 to 5.6, reflecting the transition to a more compact structure. An apparent plateau is reached at 10 mM, with the sedimentation coefficient reaching 5.7. Analogous results were obtained for the precursor, even though a higher concentration of Mg²⁺ (20 mM) is required in order to compensate for the lack of monovalent ions. The effect of K⁺ is much weaker. S, Svedberg unit.

We next examined the effect of the two most common in vivo ions, K⁺ and Mg²⁺, on the sedimentation parameters of 6S RNA. The sedimentation behaviour of both the mature and the precursor molecule was analysed using titrations of MgCl₂ and KCl. Sedimentation of both precursor and mature 6S forms exhibited pronounced dependence on Mg²⁺, reflecting the transition to more compact structures upon increase of ion concentration (Fig. 6B). The effect of K⁺ is less prominent, and, even at 200 mm, the native compactness of the molecule is not achieved. Based on the graphs in Fig. 6B, the transition from the ‘low-salt’ to ‘high-salt’ conformation of mature 6S has an estimated K_d of 1.1 mm for the MgCl₂-driven transition and 26 mm for the KCl-driven transition. These values are 3.5 and 18 mm, respectively, for the precursor molecule. These results are consistent with expectation for poly-electrolyte interactions of cations with poly-anionic RNA, which increase in proportion to the cation charge.

The R _h of 6S was investigated over a temperature range using dynamic light scattering. The values presented in Table 1 show a steady increase of the hydrodynamic radius for the mature 6S with rising temperature. The corresponding loss of compactness upon heating to 50 °C is reflected by the 25% increase of the hydrodynamic radius at 20 °C. The process is reversible, and the molecule regains its initial hydrodynamic radius when the temperature returns to 20 °C. Interestingly, the hydrodynamic radius of the precursor molecule decreases at elevated temperatures, which may be due to shrinking of the hydration/ion sphere or to more complicated changes of the structure.

Table 1.

6S parameters derived by dynamic light scattering. %PDI indicates the polydispersity index (%).

Temperature (°C)	Mature 6S				Precursor 6S
	R h (Å)		%PDI		R h (Å)		%PDI
	Warming	Cooling	Warming	Cooling	Warming	Cooling	Warming	Cooling
20	54 ± 2	53 ± 2	19 ± 3	17 ± 3	99 ± 5	98 ± 3	18 ± 1	18 ± 1
30	57 ± 2	54 ± 3	16 ± 1	14 ± 2	97 ± 2	94 ± 2	17 ± 1	17 ± 1
40	58 ± 2	57 ± 2	15 ± 2	14 ± 2	92 ± 1	93 ± 1	12 ± 1	16 ± 1
50	59 ± 0		13 ± 1		92 ± 2		12 ± 2

Thus, analytical ultracentrifugation measurements suggest an elongated shape even in the native fully folded state, and confirm the significance of monovalent ions for neutralizing phosphate backbone repulsions and divalent Mg²⁺ ions for promoting collapse to native folding for the 6S RNA. DLS measurements show that the molecules have high plasticity and are folded into highly flexible conformations.

6S solution studies imply formation of dimers

The structure of mature 6S was further investigated by small angle X-ray scattering (SAXS) experiments performed at a range of RNA and Mg²⁺ concentrations (Fig. 7). The SAXS data collection and derived parameters are shown in Tables 2 and 3 following established guidelines [44]. Despite the non-ideal signal-to-noise ratio due to the low sample concentration, the 6S RNA gives good quality scattering data at the different sample concentrations measured (Fig. 7 and Table 3). Consistent with the analytical ultracentrifugation data, the P(r) functions (Fig. 7E) suggest an extended shape and display a series of maxima. The first peak is normally attributed to intra-particle distances, which may be an indicator of the thickness of the molecule in our case, whereas extra maxima are characteristic of elongated particles with domain arrangements [45]. In low salt (100 mm K ⁺, absence of Mg²⁺), the profile has two maxima, characteristic of a conformation extended by electrostatic repulsion and indicating a particle with separated subunits. Thus, the first peak at approximately 25 Å fits well with the diameter of the A-form duplex [46], while the other at approximately 95 Å could indicate the subunits separation. As the ionic strength increases by increasing Mg²⁺, the profiles adopt a shape consistent with a random ensemble between the extended and the compact state [47]. The lack of a peak in the (normalized) Kratky plot (Fig. 7C) under all conditions measured (a typical indicator of the unfolded state in the case of proteins) further supports the concept of a highly dynamic molecule. The R _g, computed either in reciprocal space using the Guinier approximation [48] or in real space from the P(r) function, and the D _max estimated from the P(r) function (Fig. 7D and Table 2), indicate increasing compactness of 6S upon increasing Mg²⁺ concentration.

Fig. 7. Small angle X-ray scattering for mature 6S at a range of Mg²⁺ concentrations.

(A) The linear Guinier regions for 6S at various Mg²⁺ concentrations are indicative of monodisperse solutions. (B, C) Experimental scattering profiles (B) and dimensionless Kratky plot (C) for 6S at various Mg²⁺ concentrations. (D) Overall size changes as a function of Mg²⁺ concentration. (E) Experimentaldistribution functions of 6S solutions for various Mg²⁺ concentrations.

Table 2. SAXS-derived parameters for 6S RNA.

	Mg2+ concentration (mm)
	0	2	5	20
Rg (Å) [from P(r)]	79 ± 5	75 ± 5	76 ± 10	72 ± 5
Rg (Å) [from Guinier approximation]	78 ± 5	75 ± 5	73 ± 10	72 ± 5
Dmax (Å)	260 ± 15	245 ± 15	240 ± 20	240 ± 15
Porod volume estimate (Å3)	140 000 ± 10 000	125 000 ± 10 000	125 000 ± 10 000	95 000 ± 10 000
Mr (kDa) [from Porod invariant]	140 ± 10	125 ± 10	125 ± 10	95 ± 10

Table 3. SAXS data collection for 6S RNA.

Parameter	Value
Instrument (source & detector)	EMBL P12 beamline (PETRA III; Deutsches Elektronen Synchrotron, Hamburg, Germany) 2D photon-counting Pilatus 2M (DECTRIS)
Beam geometry (mm2)	0.2 × 0.2
Wavelength (Å)	1.2
s range (Å−1)	0.005–0.6
Exposure time (s)	1 (20 × 0.05)
Concentration range	0.45–0.90
(mg mL −1)
Temperature (K)	296.15

The molecular mass estimated from the Porod volume (approximately 125 kDa, Table 2) suggests the possible existence of dimeric 6S, given that the expected mass of the monomer is 58 kDa. DLS measurements further support the notion of oligomerization of 6S under specific conditions. Fitting of the size distribution data shown in Fig. 1C at 20 °C gives an estimated hydrodynamic radius (R _h) of 54 ± 2 Å for mature 6S and 99 ± 2 Å for the precursor molecule. The estimated R _h of the mature 6S corresponds to a molecular weight of 150 kDa. Similarly, the precursor 6S, with an expected mass of approximately 120 kDa for a monomer, shows an R _h corresponding to a molecular weight of 300 kDa. Thus, both SAXS and DLS measurements imply the potential of 6S to form dimers under specific conditions in solution.

Theoretical model of the 6S 3D structure

Consistent with the above data, which indicate a rather flexible and highly anisometric molecule, all the attempts at dummy bead reconstruction were problematic. Twelve independent model reconstructions were averaged using the program suite DAMAVER [49], generating a normalized spatial discrepancy of 0.956 ± 0.364. Therefore, the heterogeneity of the models indicated that ab initio reconstruction of 6S is not meaningful.

Next, the program RNAComposer [50] was used to fold into a 3D model each of the secondary structures described in the literature for 6S [23,24], i.e. with and without a small internal hairpin in the central bubble. As shown in Fig. 8A,B, a small difference in the secondary structure induced significant changes in the tertiary stucture. The biophysical parameters (size) of the resultant models are small compared to the experimentally derived parameters. Thus, the next step was to combine two monomers into dimeric assemblies as the SAXS and DLS measurements suggested. The crystal structure of the A. aeolicus 6S RNA 12 nt nucleolytic fragment has shown the potential for the closing stem to provide an interface for monomers interaction [33]. Thus, in our models, we used the closing stem as the dimerization interface. Two possible dimeric assemblies were generated by associating stem-to-stem monomers of model 1, i.e. with the internal hairpin (Fig. 8A), and model 2, i.e. without the hairpin (Fig. 8B). The biophysical parameters (size) of the dimer comprising molecules of model 1 were comparable with those experimentally obtained by SAXS. Therefore, the model 1 dimer was used for rigid body modelling against the experimental scattering curve of 6S obtained with 2 mm Mg²⁺, which represents the best dimer according to the estimated molecular mass (see Table 2). The size of this model is comparable to the D _max estimated from the P(r) function (approximately 265 Å versus 245 ± 15 Å), with χ = 0.945 (data not shown). However, it must be borne in mind that the putative model discussed here may be an average model among the multiple conformations that 6S may adopt in solution.

Fig. 8. Theoretical 3D models of mature 6S.

Three-dimensional theoretical models of mature 6S RNA calculated for two possible secondary structures, i.e. with a small stem-loop inside the central bubble (A, model 1) and without the stem-loop (B, model2). The slight difference in base pairing in the central bubble results in significant size and orientation changes in three dimensions.

Discussion

In this work, we characterized the processing of precursor 6S by RNase E, and confirmed previous observations that the 5′ 221 nt fragment, which is generated together with the mature 6S, is susceptible to continued degradation by RNase E [20]. We show for the first time that the NTD/RNase E activity on processing of the precursor 6S is inhibited at high concentrations of substrate. The results of gel shift assays and nanoflow ESI mass spectrometry indicated that the reason for this inhibition is not stable binding of any of the 6S precursor or mature molecules to NTD/RNase E.

Given that RNase E is part of the larger degrado-some assembly [51], to rule out the possibility that our observations are due to the truncated RNase E used in our experiments, we repeated some of them using the full degradosome. Thus, by using a reconstituted degradosome that comprises an inactive mutant of RNase E we observed the formation of a complex with 6S that persists in agarose gels and has much greater stability than the complex of 6S with NTD/RNase E alone, suggesting that the additional binding sites for RNA influence the association with the substrate (data not shown). Next, we performed processing experiments for precursor 6S using an active preparation of endogenous degradosome, and observed a similar degree of inhibition to that observed with the catalytic domain of RNase E (data not shown). We conclude that catalysis from the full complex is also substrate inhibited. Our observations may be interpreted as indicating a largely distributive mechanism of action of RNase E on the precursor 6S RNA and the 5′ fragmets derived after the initial cleavage, i.e. each round of catalysis involves substrate binding, product(s) release and substrate re-binding.

Our results also show that NTD/RNase E is not able to further process the mature 6S, and confirm the ability of the enzyme to selectively process rather than destroy structured RNA substrates. Moreover, we also demonstrated that the presence of a potential cleavage site, i.e. the single-stranded loop of the mature 6S molecule, or even multiple binding sites for one or more protomers of the same or different NTD/ RNase E tetramers, are not sufficient for further processing. This is probably because 6S is not able to fulfil all the requirements of either simultaneous binding at sensing and catalytic sites or coupled binding and processing by more than one protomer, as previously proposed for the catalytic mechanism of RNase E [35,36].

The most puzzling finding of this work is the dimerization of the 6S RNA, which was observed by both SAXS and DLS. Comparison of the experimental profiles of P(r) function with known examples [47] shows that 6S follows the behaviour of a two-domain molecule, with these domains present in either an extended conformation (low magnesium concentration) or as a random ensemble of intermediate conformations (high magnesium concentration). Taking into account the size of 6S (R_g, R_h and D _max) and its molecular weight, the conformation of the two domains may reflect a dimer organization. Moreover, 3D models constructed based on the known secondary structure of 6S, in combination with the experimental biophysical parameters, also suggest the existence of dimers. It is worth noting that, in order to eliminate inter-particle interactions, we used the snap-cooling/refolding procedure after annealing for preparation of the sample. The sample produced by this treatment was checked and proved to be highly homogeneous. Moreover, it appears that the dimeric association does not interfere with the functionality of the sample, as both mature and precursor forms are able to interact with their native partners, i.e. RNAP and RNase E, respectively. It is interesting that, in the complex with σ⁷⁰–RNAP, however, 6S participates as one monomer. In general, the denaturing purification protocols, which necessarily involve annealing/refolding procedures, carry a high risk for formation of oligomeric artefacts, as the large RNAs may not be able to fold back into a native state because of possible multiple folding pathways. For that reason, one of our future aims is to develop a protocol for 6S purification under native conditions. Thus, it remains to be demonstrated whether this dimerization is biologically relevant, or whether it reflects the oligomerization state of 6S under the specific solution conditions and purification protocols used for production and isolation of our samples.

Based on the two experimental secondary structures of 6S RNA, i.e. with and without a small stem-loop inside the central bubble [23,24], two theoretical models were constructed. These models have quite different biophysical parameters, and only that with a stem-loop inside the central bubble is compatible (in its dimeric form) with the SAXS data. Recent studies highlight the significance of formation of a hairpin in the 3′ strand of the central bubble in the complex of 6S with the RNAP. It is believed that this hairpin orchestrates release of 6S from the complex. Our model suggests that a small hairpin in this area exists even in the unbound state of the 6S molecule.

Understanding the structural dynamics of the 6S RNA under different conditions is an important area for future work to increase our knowledge of biophysical properties of RNAs, and help us to understand the molecular mechanisms of recognition and interaction with protein partners.

Experimental procedures

Synthesis of 6S RNA by in vitro transcription

Precursor and mature 6S DNA fragments were amplified from E. coli K12 genomic DNA by PCR, using primer pairs that incorporated a T7 promoter in the upstream region and EcoRI and HindIII restriction enzyme cleavage sites at the 5′ and 3′ ends. The following primers were used for amplification of precursor 6S RNA (ssrSP2/ssrSR) and the mature form (ssrSF/ssrSR): ssrSP2, 5′-AAAGGCGAATTC TAATACGACTCACTATAGG ACTGAACAGTTGGTCTTCAT TGCCGCA-3′; ssrSF, 5′-AAAGGCGAATTC TAATACGACTCACTATAGG ATTTCTCTGAGATGTTCGCAAGCGGGC-3′; ssrSR, 5′-AAAGGCAAGCTT GAATCTCCGAGAT GCCGCCGCAGGCTGTAA-3′. Bold letters indicate restriction enzyme sequences, italic letters indicate the priming sequence, and underlined letters indicate the T7 promoter sequence. The isolated DNAs were then cloned into pUC19 [52], and used as templates in PCR reactions to produce amplified DNA fragments of 6S precursor and mature forms for run-off in vitro transcription of the corresponding RNAs. The primers used for amplification were 6SF (5′- CCCCCTAATACGACTCACTATAGG-3′) and 6SR (5′-GAATCTCCGAGATGCCGCCGCAGGCTGTAA-3′).

In vitro transcription reactions using T7 RNA polymerase in the presence of an rNTP mixture were incubated for 3-4 h at 37 °C. The RNA was then precipitated using ethanol at —20 °C, and further purified using an 8% preparative urea/polyacrylamide denaturing gel. The target RNA band was excised and electroeluted overnight at 4 °C [53]. To obtain structured RNA, the sample was annealed by heating at 95 °C for 5 min, cooled on ice for another 5 min, and then incubated at room temperature for at least 30 min before addition of magnesium ions. The RNA was concentrated using a centrifugal concentrator (Vivaspin, Sartorius Stedim Biotech GmbH, Goettingen, Germany).

Dynamic light scattering

The sample concentration used was 0.6 mg·mL^—1 for the precursor and 1.6 mg·mL^—1 for the mature 6S in 20 mm HEPES, pH 7.9, 100 mm KCl, 5 mm MgCl₂. Samples were centrifuged at 13 000 g, for 5 min at room temperature, and the supernatant was filtered through a poly(vinylidene difluoride) membrane filter, pore size 0.45 μm (Pall Life Sciences, Portsmouth, UK). DLS measurements were performed using a Zetasizer NanoS particle analyser (Malvern Instruments Ltd, Malvern, UK) operating at 633 nm. Data were obtained at 20, 30, 40 and 50 °C. The samples were temperature-equilibrated for 3 min before the measurement, and 12-17 measurements were averaged for each intensity distribution. The data were analysed using software available from the manufacturers.

Nanoflow electrospray (ESI) mass spectrometry

Endogenous σ⁷⁰–RNAP holoenzyme was prepared from E. coli BL21 cells using a combination of the protocols described previously [54–56]. Mass spectrometry data were acquired using a modified QTOF Ultima instrument (Waters, Vernon Hills, IL, USA) configured for nanoflow ESI in positive ion mode essentially as described previously [57,58]. A stoichiometric excess of mature 6S was added to a σ⁷⁰–RNAP sample in 50 mm Tris/HCl, pH 8.0, 100 mm Mg-acetate, 0.5 mm dithiothreitol and 0.1 mm EDTA. The mixture was incubated for 30 min at room temperature, then flash frozen and kept at −80 °C until analysis. Complexes were prepared for nanoflow ESI by buffer exchange from their purification buffer to ammonium acetate solution using Amicon Ultra–0.5 centrifugal ultrafiltration units with Ultracel membranes (RNAP, 10 kDa molecular weight cut-off, 500 mm ammonium acetate; RNAP + 6S RNA, 100 kDa molecular weight cut-off, 250 mm ammonium acetate; pH 7.8). Final ammonium acetate concentrations for nanoflow ESI were modified as required either by dilution of the buffer -exchanged complexes with water or addition of 7.5 M ammonium acetate solution (Sigma-Aldrich, Gillingham, UK).

Individual subunit masses were determined by nano -LC-MS using an RSLC system (Dionex, Hemel Hempstead, UK) equipped with a UV detector set at 214 and 280 nm. A denatured RNAP sample was obtained by preparing a 1 : 5 v/v dilution of RNAP (6 μM in 10 mm Tris/HCl, 5% v/v glycerol, 0.1 mm EDTA, 1 mm dithio- threitol, 300 mm NaCl) with 0.05% trifluoroacetic acid, and applying 1 μL of the resulting sample to a nano PS- DVB reverse-phase monolithic column (100 μm internal diameter, length 25 cm; Dionex) equilibrated with 92% solvent A (0.05% trifluoroacetic acid) and 8% solvent B (0.04% trifluoroacetic acid, 90% acetonitrile). RNAP proteins were eluted using a linear gradient of 25–80% solvent B over 24 min at a flow rate of 600 nL·min^–1. The column effluent was delivered to a nanoflow ESI source, and analysed using a Q-TOF-type mass spectrometer (QSTAR, MDS-Sciex, Concord, Canada).

RNA processing by the catalytic domain of RNase E

The catalytic domain of RNase E was over-expressed and purified as described previously [34]. A reaction mixture of 20 μL was prepared by mixing RNase E to a final concentration of 1 μM with varying concentrations of RNA in a buffer comprising 25 mm Tris, pH 7.5, 25 mm NaCl, 25 mm KCl, 1 mm dithiothreitol, 10 mm MgCl₂, and the mixture was incubated at 37 °C. At the indicated time points, 5 μL aliquots were mixed with 5 μL proteinase K solution (0.25 mg·mL^–1), incubated at 50 °C for 20 min, and then loaded onto an 8% urea/polyacrylamide gel.

Electrophoretic mobility shift assays

A constant amount of RNA sample was incubated with varying concentrations of NTD/RNase E for 30-60 min at room temperature in a buffer containing 25 mm Tris/HCl, pH 7.5, 25 mm NaCl, 25 mm KCl, 1 mm dithiothreitol in the presence of 4 U·μL ^–1 of RNaseOUT ribonuclease inhibitor (Invitrogen, Carlsbad, CA, USA). The mixtures were loaded onto 6-8% native agarose gels (90 V, constant), stained with SYBR Gold (Thermo Fisher Scientific), and the bands were visualized and quantified by phosphoimager analysis.

RNA thermal melting assays

The hyperchromicity of mature and precursor 6S RNA was measured using a Cary UV-Vis spectrophotometer (Agilent Technologies, Santa Clara, CA, USA) equipped with a thermostat control. The samples had a concentration of 50 ng·μL⁻¹ in a buffer containing 20 mm HEPES 7.9, 5 mm MgCl₂, 100 mm KCl, 0.5 mm dithiothreitol, 0.5 mm phenylmethanesulfonyl fluoride. During the measurements, the samples were heated from 25 to 95 °C at a rate of 1 °C·min⁻¹.

Fluorescence-based thermal shift assays were performed using a Bio-Rad iQ5 thermal cycler (Hemel Hempstead, UK). RNA samples were diluted in the presence of the intercalating fluorophore SYBR Green (Thermo Fisher Scientific). Reactions were performed at RNA concentrations varying from 48 to 240 nm. The temperature gradient was 1 °C·min⁻¹ and fluorescence was recorded at 522 nm. The melting temperature was calculated using bio-rad iq5 version 2.0 standard edition optical system software as the maximum of the derivative of the resulting denaturation curves.

Circular dichroism

CD spectra were collected using an Aviv circular dichroism spectrometer, model 410 (Lakewood, NJ, USA). Scans were recorded in the spectral range 320-200 nm with a 1 nm wavelength step and a mean time of 1 s. Each spectrum was the mean of four scans and was background-corrected. The sample concentrations were 150 and 235 ng·μL⁻¹ for the mature and precursor molecules, respectively, in a buffer containing 20 mm HEPES, pH 7.9, 100 mm KCl, 0.5 mm dithiothreitol, 0.5 mm phenylmethanesulfonyl fluoride, 5 mm MgCl₂.

Analytical ultracentrifugation

Experiments were performed using an Optima XL-I analytical ultracentrifuge (Beckman Coulter, High Wycombe, UK) at 20 °C, with an An60Ti four-hole rotor and Epon double sector centrepieces of 12 mm optical path with quartz windows. Sedimentation velocity experiments were performed at 30 krpm, and absorbance scans were taken at 260 nm. Sedimentation coefficient distributions, c(s), were calculated by boundary modelling of sedimentation velocity data using the program SEDFIT [59]. Buffer viscosity and density were estimated using the program SEDNTERP [60]. A partial specific volume of 0.53 cm³·g ¹ for nucleic acids was used.

Small angle X-ray scattering

Data for mature 6S RNA were collected (Table 3) and recorded using a Pilatus 2M pixel detector (DECTRIS, Baden, Switzerland), and the results are shown in Table 2. Scattering profiles were recorded at RNA concentrations ranging from 0.45 to 0.90 mg·mL⁻¹ and MgCl₂ concentrations ranging from 0 to 20 mm in a buffer comprising 10 mm HEPES, pH 7.5, 5 mm KCl, 0.01 mm EDTA. A robot sample changer [61] was used to automatically load the samples, and 20 successive exposures (frames) of 6S RNA (0.05 s per frame) were collected in order to assess radation damage (data not shown). Overall structural parameters such as the radius of gyration (R _g) were calculated using the Guinier approximation [48] implemented in the program PRIMUS [62] and assuming that, at very small angles (s < 1.3/Rg), the intensity is represented as I (s) = I(0)·exp(-(s R_g)²/3). The pair-distance distribution function P(r) was computed using the program GNOM [63], and the maximum particle dimension (D _max) as well as the real-space R _g were estimated from it. The molecular mass was estimated from the Porod invariant [64] as it has been shown that, for RNA, the molecular mass in kDa corresponds to the Porod volume in nm³. A dimeric model of mature 6S RNA was constructed by rigid-body modelling using the program SASREF [65]. The software used for SAXS data analysis is part of the package ATSAS 2.5 [66].

Degradation

Samples of mature and precursor 6S RNAs at a concentration of 40 ng·μL ⁻¹ were incubated for 21–23 h at 30 °C in 50mm Tris/HCl, pH 8.0, 20 mm MgCl₂, 100 mm KCl in the presence of 4 U·μL ⁻¹ of RNaseOUT ribonuclease inhibitor (Invitrogen). The resulting degradation products were analysed using denaturing urea/polyacrylamide gels.

Abbreviations

DLS

dynamic light scattering

NTD

N-terminal domain

RNAP

RNA polymerase

SAXS

small angle X-ray scattering

σ⁷⁰–RNAP

RNA polymerase holoenzyme complex with the sigma factor σ⁷⁰.