RNase E endonuclease activity and its inhibition by pseudoridine

The conserved endoribonuclease RNase E dominates the dynamic landscape of RNA metabolism and underpins control mediated by small regulatory RNAs in diverse bacterial species. We explored the enzyme’s hydrolytic mechanism, allosteric activation, and interplay with partner proteins in the multi-component RNA degradosome assembly. RNase E cleaves single-stranded RNA with preference to attack the phosphate located at the 5□ nucleotide preceding uracil, and we corroborate key interactions that select that base. Unexpectedly, RNase E activity is impeded strongly when the recognised uracil is isomerised to 5-ribosyluracil (pseudouridine), from which we infer the detailed geometry of the hydrolytic attack process. Kinetics analyses support models for recognition of secondary structure in substrates by RNase E and for allosteric auto-regulation. The catalytic power of the enzyme is boosted when it is assembled into the multi-enzyme RNA degradosome, most likely as a consequence of substrate channeling. Our results rationalize the origins of substrate preferences of RNase E and illuminate its catalytic mechanism, supporting the roles of allosteric domain closure and cooperation with other components of the RNA degradosome complex.

Decades of analysis of RNase E activity indicate that there is no simple sequence code for its substrates per se, but instead a strong preference to cleave within single-stranded regions enriched in A or U (Chao et al. 2017;Del Campo et al. 2015;Kime et al. 2010;2014;Mackie 2013). Global RNA target analyses performed both in vivo and in vitro identify uracil positioned to the 3 side adjacent to the nucleotide of the scissile phosphate (the +2 position) as a strong signature for RNase E activity (Chao et al. 2017). For many substrates in both destructive and maturation pathways, the enzyme is activated by transformation of the 5 end of the substrate from a triphosphate normally found on nascent transcripts, to a monophosphate found on processed species (Mackie 2013). For other substrates, the status of Most of the Ψ-containing substrate resists cleavage by RNase E in the course of the experiment. The cleavage that does occur is partially shifted to the -1 position. These findings indicate that the recognition of uracil is not simply due to a hydrogen bonding interaction with the principal substituents of the base, but also depends on the detailed interactions that influence the phosphodiester geometry. Most likely these details affect the hydration pattern of the substrate and the energy required to achieve the conformation that enables development of the enzymatic transition state.
Substitution of K112 with Q changes the cutting pattern for the Ψ containing substrate. In general, substitution of lysine by the polar glutamine is expected to retain capacity for hydrogen bond formation. Despite the reduced sensitivity to RNase E, the preferred cleavage site was moved to the +2 or +3 position relative to the site cut by wildtype enzymes (Figure 1b, middle and bottom panels). The cutting pattern and shift effect are substantially weakened for the K112A substitution, suggesting that activation of hydrolysis requires a long polar side chain at position 112 ( Figure 1a). The K112Q substitution perhaps cause the substrate to align differently in the active site so that it is displaced by one or two nucleotides in the 3 direction compared to the corresponding wild-type complex.

RNase E catalytic power can be boosted by substitutions at DNase I and RNase H-like domains
Earlier studies showed that the catalytic activity of RNase E is boosted by mutations of conserved, non-catalytic residues in the RNase H-like domain (D26N and D28N) and DNase I domain (D338N) (Figure 2b) (Bandyra, Wandzik, and Luisi 2018). The substitutions are at a distance from the active site but involve areas where the conformation changes associated with apo to substrate-bound states and are likely to impact on allosteric switching of the enzyme (Bandyra, Wandzik, and Luisi 2018) (Figure 2b). We compared the catalytic activity of the wild type and hyperactive triple mutant D26N, D28N, D338N (NTD-3M). For substrates, we used the sRNA GlmZ, which is inactivated by RNase E cleavage, and 9S RNA which is a precursor of ribosomal 5S RNA. The enzyme cleaves the 9S mainly at three sites to form the p5S precursor ribosomal RNA product (Figure 2a) (Cormack and Mackie 1992;Christiansen 1988). The 5 phosphorylation state of 9S RNA can impact on the cleavage events, with the second event having the activating 5 P group present and anticipated to be intrinsically accelerated if the group is read by the enzyme.
Compared to the wild type, NTD-3M showed higher activity for 5 PPP-9S and a boost in catalytic power due to both increased catalytic rate and decreased K m (Figure 2c and 2d, Table 1). The results support the proposed role of allosteric autoregulation of enzyme activity, in which domain closure helps to pre-organize the active site so that the apparent affinity of the Michealis-Menten complex increases probably by decreasing the energy barrier to capture and engulf the substrate.

Metals in the catalytic mechanism: RNase E active site may recruit one metal in the apo form
The active site, which is present within a DNase I sub-domain of the NTD, bears two conserved aspartate residues (D303 and D346) that recruit magnesium ion to activate a water molecule for nucleophilic attack on the scissile phospho-diester bond (Thompson, Zong, and Mackie 2015;Callaghan et al. 2005) (Figure 1). The binding interactions between RNase E and metal cofactors was evaluated by isothermal calorimetry (ITC) using a variant of RNase E with residue D346 replaced with a cysteine residue, which was reported previously to be catalytically active in presence of Mn ++ , not Mg ++ (Figure 3a) (Thompson, Zong, and Mackie 2015). Testing the activity of the D346C NTD on two different RNAs, 9S and the small RNA RprA, confirms that the enzyme is active for cleavage only in the presence of Mn ++ ( Figure   3b and 3c). Using isothermal titration calorimetry (ITC) and titrating the mutant enzyme against Mn ++ yields a K D for metal binding in the absence of RNA at 17 µM, with associated

Probing RNase E mechanism with unnatural amino acids
To explore interactions between NTD substrates we used derivatives of the protein with the photocrosslinkable amino acid para-azido-phenylalanine (p-AzidoPhe) incorporated at specific positions in the 5 -sensing pocket and duplex-RNA binding site using the amber- On the other hand, the substitution at the duplex binding surface had little impact on activity ( Figure 4c). Exposing NTD p-AzidoPhe derivatives to light at 254 nm in the presence of 9S subdomains did not yield appreciable photocrosslinking directly to the RNA, as detected by mobility shifts in denaturing protein gels ( Figure 4d). However, the protein migrated differently in the denaturing gel when uv-illuminated in the absence of RNA, and this may arise from intra-molecular crosslinks (Figure 4d). These results support earlier findings that the NTD undergoes conformational change upon RNA binding (Koslover et al. 2008;Bandyra, Wandzik, and Luisi 2018;Callaghan et al. 2005).

Activities of the degradosome for multi-step processing of complex substrates
To explore how RNase E activity is impacted by the degradosome organization, we explored the activity of the isolated catalytic domain with the degradosome for several RNA substrates ( Figure 5a). The activity of purified degradosome and the isolated NTD for 9S RNA cleavage was measured using as substrate 9S RNA with 5 -triphosphate (5 PPP-9S) or 5monophosphate (5 P-9S). The reaction was first explored with the NTD, which gave the expected p5S product ( Figure 5b) and the apparent kinetic parameters for formation of p5S are summarized in figure 5d and Table 1. These are comparable to values reported earlier, and differences are likely due to reaction conditions (Redko et al. 2003;Thompson, Zong, and Mackie 2015;Hadjeras et al. 2019). Comparing the kinetics parameters, the assays revealed significant acceleration to cleave 5 P-9S versus 5 PPP-9S (Figure 5b-5d), corroborating earlier findings that 5 -sensing can contribute to the first cleavage event in 9S processing by RNase E (Mackie 2013; Cormack and Mackie 1992).
Next, we explored the activity of RNase E within the context of the RNA degradosome. We prepared purified recombinant degradosome (comprising RNase E 1-1061, RhlB, enolase, and PNPase) as well as a subassembly comprising RNase E 1-850, RhlB, and enolase (truncated degradosome, TD; Figure 5a). The processing of 5 PPP-9S showed relatively faster cleavage of 9S for TD and full degradosome assemblies compared to the isolated catalytic domain under the same experimental conditions (Figure 5b-5d) and greater catalytic power (kcat/K m ), mostly through changes to the apparent K m (Figure 5d, Table 1). The cleavage rates were also seen to be greater using 5 P-9S as substrate (Figure 5c, 5d). These observations suggest that the degradosome assembly facilitates RNase E activity, most likely through substrate capture that decrease the effective K m .

Domain decomposition of 9S and impact on RNase E and degradosome activities
RNase E activity can be directed by structural elements, and for the 9S substrate stem-loop II has previously been shown as the minimal structural requirement needed for RNase E to

Activity of the degradosome in single-step cleavage of the riboregulatory GlmZ
The regulatory RNA GlmZ contributes to homeostasis of glucosamine-6 phosphate (G6P), a precursor in the peptidoglycan synthesis pathway and other metabolic routes (Urban and Vogel 2008;Kalamorz et al. 2007;Göpel et al. 2013). As part of the control network, GlmZ is inactivated by RNase E with the help of the adaptor protein RapZ (Gonzalez et al. 2017;Durica-Mitic and Görke 2019;Kalamorz et al. 2007;Urban and Vogel 2008;Göpel et al. 2013). We explored the effects of RapZ on RNase E mediated cleavage of GlmZ (Figure 6a).
Analysis of the cleavage reaction revealed production of the GlmZ-Pro, the cleavage product of GlmZ lacking the glmS recognition site (Figure 6b). While GlmZ-Pro was formed in the presence of RapZ for the 5 -PPP-GlmZ form, the 5 P-GlmZ form was rapidly chased into a smaller fragment. In the presence of another chaperone protein Hfq, less GlmZ was processed by the NTD, but the RapZ-specific cleavage product GlmZ-Pro was still formed (Figure 6b, c). The hyperactive version of NTD (NTD-3M) was also found to efficiently generate GlmZ-Pro in the presence of RapZ (Figure 6d).
We determined the kinetics parameters for GlmZ cleavage using degradosome assemblies, revealing faster cleavage of GlmZ by the degradosome and truncated degradosome compared to NTD (Figure 6d and 6e, Table 1). Similar to the findings with 9S RNA (Figure 5), these results indicate that the RNase E mediated cleavage of RNAs can be facilitated in the context of the degradosome assembly.

Discussion
The half-lives of most E. RNAs but also to release the products quicker than wild type.
The results presented here corroborate the importance of uracil at position +2 with regard to the cleavage site as a key feature of a preferred cleavage site by RNase E and the role of residue K112 in recognising the +2 uracil. Unexpectedly, cleavage by RNase E is strongly impeded when the +2 uracil is substituted with pseudouridine, which is surprising given that this substitution presents only one new hydrogen bonding group on the pyrimidine.
The isomerisation of uracil to pseudouridine presents the N1 as a hydrogen bond donor and most likely affects the hydration pattern that will include the phosphate backbone. In most RNA structures, N1 is predicted to interact with the phosphate backbone of both the pseudouridine and the 5 residue (Charette and Gray 2000). In the context of RNase E catalytic site, this interaction could restrict the backbone conformation at position +2 and disfavour the geometry necessary for catalysis.  (Sulthana, Basturea, and Deutscher 2016). As part of the mechanism of quality control, RNase E could hypothetically sense whether the precursors have been properly modified with pseudouridine and destroy those that have not undergone the isomerisation. However, our tests of RNase E activity on tRNAs isolated from cells that are deficient in pseudouridine synthase show that these species, as well as the wild type controls, are resistant to digestion (data not shown). Recent studies suggest that pseudouridine is also prevalent in mRNAs and noncoding RNAs, and that pseudouridylation is regulated by environmental stresses and Similarly, the membrane associated E. coli RNA degradosome forms transient clusters over the membrane during RNA turnover (Moffitt et al. 2016;Strahl et al. 2015). The ribonucleoprotein bodies stimulate RNA decay of target RNAs and complete mRNA turnover, so preventing accumulation of potentially harmful degradation intermediates.
The results presented here show that the catalytic power of RNase E is boosted when the enzyme is assembled into the multi-enzyme RNA degradosome assembly, most likely through substrate channeling. Our observations suggest that the degradosome facilitates RNase E activity, most likely through substrate capture and allostery-mediated acceleration of catalytic rates. We anticipate that the clustering of degradosomes in bodies with liquid-like phase separation further concentrates the enzymatic activities of the machinery and changes the physicochemical conditions that impact on activity. Our results rationalize the origins of substrate preferences of RNase E and illuminate its catalytic mechanism, supporting the roles of allosteric domain closure and cooperation with other components of the RNA degradosome complex.
Cultures were grown in 2xTY media supplemented with 100 µg/mL carbenicillin at 37ºC, in an orbital shaker set at 220 rpm. The culture was induced between 0.5 to 0.6 OD 600nm by adding 1 mM isopropyl-β-thiogalactopyranoside (IPTG) and harvested after 3 hours of incubation by centrifugation at 4200 g and 4º C for 30 minutes. Cell pellets were stored as suspension in nickel-column buffer A (20 mM Tris pH 7.9, 500 mM NaCl, 5 mM imidazole, 1 mM MgCl 2 ) at -80 o C. Once thawed, the cell culture suspension was supplemented with DNase I and EDTA-free protease inhibitor cocktail tablet (Roche) and cells were lysed by passing through an EmulsiFlex-05 cell disruptor (Avestin) for 2-3 times at 10-15 kbar pressure. The lysate was clarified by centrifugation at 35000 g for 30 minutes at 4º C and the supernatant was loaded onto a pre-equilibrated HiTrap Chelating HP column charged with nickel ions (GE Healthcare). The column was washed extensively with wash buffer (20 mM Tris pH 7.9, 500 mM NaCl, 100 mM imidazole, 1 mM MgCl 2 ), followed by gradient elution of RNase E with elution buffer (20 mM Tris pH 7.9, 500 mM NaCl, 500 mM imidazole, 1 mM MgCl 2 ). Fractions containing RNase E were pooled and loaded on a butyl sepharose HP column (GE Healthcare) which previously was equilibrated in high-salt buffer (50 mM Tris pH 7.5, 50 mM NaCl, 25 mM KCl, 1 M (NH) 2 SO 4 ). A gradient of a low-salt buffer (50 mM Tris pH 7.5, 50 mM NaCl, 25 mM KCl, 5% glycerol) was used to elute protein. Fractions containing RNase E were pooled, concentrated and loaded onto a size-exclusion column (Superdex TM 200 Increase 10/300, GE Healthcare) equilibrated previously in storage buffer

RNA degradation assays with pseudouridine substrates
polyA 20-mer, A20U and A20Ψ were obtained from Dharmacon. Oligoribonucleotides were 5 labelled with 32 P using polynucleotide kinase (Fermentas), according to manufacturer instructions. Assays were carried out in reaction buffer (25 mM Tris-HCl pH 7.5, 50 mM NaCl, 50 mM KCl, 10 mM MgCl 2 , 1 mM DTT, 0.5 U/µL RNase OUT) at 37°C. 100 nM purified RNase E NTD was used for the reactions. Time course reactions were stopped at indicated time points by addition of STOP solution (20 mM EDTA, 2% w/v SDS). RNA loading dye (Thermo Fisher) was added to samples which were denatured (98°C, 2 min) and loaded onto polyacrylamide gels containing 7.5 M urea. Gels were dried and exposed to phosphor screens (GE Healthcare) and the signal analysed with TyphoonT 9400 (GE Healthcare).

Kinetic assay
Ribonuclease cleavage of RNAs by RNase E was carried out at 30º C in the reaction buffer as above (Bandyra, Wandzik, and Luisi 2018). In case of time-course assay, samples were quenched at a predetermined time points by adding proteinase K mix (proteinase K in proteinase K buffer of 100 mM Tris-HCl pH 7.5, 150 mM NaCl, 12.5 mM EDTA, 1% SDS), followed by incubation at 50º C for 30 minutes. In the case of kinetic assay, RNA degradation was monitored against 10, 15, 20, 25, 50, 100, 125, 150, 200, 250, 300, 350, 400, 500, 600, 700 nM of the RNA while reaction was quenched as before by proteinase K within the linear range of the time-course curve. RNA samples were thereafter mixed with loading dye (Thermo Fisher), heated at 95º C for 2 minutes and loaded onto 8% urea-PAGE gel. The degradation products were visualized under UV transilluminator (GeneSnap, Syngene). To quantify, intensity of the reaction products was calculated using GeneTools (Syngene) with respect to a known amount of reference sample. Each kinetic parameter represents an average of three individual experiment.