Escherichica coli strain MG1655 rnr::kan was obtained by P1 transduction using the Keio rnr::kan disruption strain (Baba et al., 2006) as a donor. A plasmid expressing the coding region of N-terminal domain of λ-cI protein (pλ-cI-NS), which encodes for an N-terminal His6 epitope but lacks in-frame stop codons, served as the source of the non-stop reporter mRNA. A control reporter plasmid (pλ-cI-S), which encodes for an N-terminal His6 epitope and has two tandem in-frame stop codons, served as the source of the “normal” or stop codon containing reporter mRNA. A pACYCDuet-1 plasmid (prnr) harboring the E. coli rnr gene under the control of the arabinose inducible P_BAD promoter was used as a template to synthesize RNase R variants by site directed mutagenesis. For purification of RNase R and its variants, the coding region of E. coli rnr gene was cloned into the MCS of pET15b under the control of isopropyl β-D-thiogalactoside (IPTG)-inducible T7 promoter. This plasmid was used as a template to synthesize RNase R variants by site directed mutagenesis. The plasmids were transformed into W3110 (DE3) rnr rnb strain for over-expression and purification of the corresponding RNase R variants, as previously described (Ge et al., 2010).

Purified RNase R samples were prepared in 50 mM Tris-HCl (pH 7.5), 250 mM KCl, and 1 mM DTT within 24 h of data acquisition and stored at 4^°C. Scattering data were collected at beamline X9 of the National Synchrotron Light Source (NSLS, Upton, NY, USA) using a Pilatus 300K located 3.4 m from the sample for small-angle X-ray data. Wide-angle X-ray scattering data were collected simultaneously with SAXS data using a Photonic Science CCD camera located 0.47 m from the sample. Twenty microliters of sample were continuously flowed through a 1 mm diameter capillary and exposed to a 400 μm × 200 μm X-ray beam with a wavelength of 0.9184 Å for 30 s of total measurement time. Scattering data were collected at concentrations of 2.2, 3.7, 4.2, 5.0, and 10 mg/mL for RNase R^WT and at concentrations of 0.7, 2.8, 3.0, 3.5, and 6.0 mg/mL for RNase R⁷²³. Each concentration was measured in triplicate. Normalization for beam intensity, buffer subtraction, and merging of data from both detectors were carried out using PRIMUS (Konarev et al., 2003). The radius of gyration (R_g) was calculated using a Guinier approximation, I(q)=I(0)exp(-q² R_g ² /3), where a plot of I(q) and q² is linear for q <1.3/R_g (Guinier, 1939). Three independent scattering trials were averaged. GNOM was used to determine the pair distribution function, P(r), and maximum particle dimension, D_max (Semenyuk and Svergun, 1991). The linearity of the Guinier region and the forward scattering intensity were used to validate that the samples were monodisperse in solution. The forward scattering intensity, I(0), is the theoretical scattering at a q value of 0 and is proportional to the molecular weight of the sample (Putnam et al., 2007). I(0)/c, where c is sample concentration, was identical for all RNase R measurements. For RNase R⁷²³, ten independent ab initio beads models that describe the experimental data were calculated using DAMMIN (Svergun, 1999). The resulting models, which had an average χ² of 0.65 ± 0.16, and a normalized spatial discrepancy (NSD) of 0.71 Å, were subsequently aligned, averaged, and filtered by occupancy using the DAMAVER suite of programs (Volkov and Svergun, 2003). An electron density map was calculated from the filtered average model using SASTBX (Liu et al., 2012). For RNase R^WT, the crystal structure of RNase II was modeled against the solution structure data by combined rigid body fitting and ab initio bead modeling, using BUNCH (Petoukhov and Svergun, 2005).

Escherichica coli strain MG1655 rnr::kan harboring the reporter plasmid pλ-cI-NS and the various prnr variants were grown at 37^°C in Luria-Bertani (LB) broth containing 0.01% arabinose, 100 μg/ml of Ampicillin, and 30 μg/ml of Chloramphenicol. At an optical density at 600 nm (OD₆₀₀) of ~0.6, expression of the λ-cI-NS reporter mRNA was induced with a final concentration of 1 mM IPTG for 1 h. Equal number of cells were harvested and total RNA was isolated using TRI Reagent (MRC Inc.). To test the activity of RNase R variants, in vitro RNA degradation assays were performed as previously described (Ge et al., 2010). The oligoribonucleotide substrate used was a single stranded RNA with an internal step-loop structure and a 3′ poly A overhang (trpAt-A₁₀: 5′-GCAGCCCGCCUAAUGAGCGGGCAAAAAAAAAA-3′). Reaction mixtures containing the 32P-labeled trpAt-A₁₀ substrate and the desired RNase R variants were assembled and incubated at 37^°C. Aliquots of 5 μl were taken at indicated time points and quenched by the addition of 5 μl of the formaldehyde gel-loading dye. Full-descriptions of the in vivo reporter mRNA stability and ribosome enrichment assays have been described elsewhere (Ge et al., 2010; Mehta et al., 2012). A brief description of the ribosome enrichment assay is provided here. Non-stop or stop reporter mRNAs were expressed in MG1655 rnr::kan strain complemented with designated plasmid borne RNase R variant, under a P_BAD promoter. Cells were grown in media containing 0.01% arabinose to OD₆₀₀ of 1.0 and the expression of reporter constructs was induced by the addition of IPTG to a final concentration of 1 mM. Cells were allowed to grow for another 45 min, harvested, resuspended in Buffer E (50 mM Tris-HCl pH 7.5, 300 mM NH₄Cl, 20 mM MgCl₂, 2 mM β-ME, and 10 mM Imidazole) and lysed by French Press. Ribosomes were pelleted from cleared cell lysates by centrifugation at 29,000 rpm for 16–18 h through a 32% sucrose cushion in Buffer E. Ribosomes translating the reporter mRNAs were purified from tight-coupled ribosomes by Ni²⁺-NTA affinity chromatography. The ribosome-associated RNase R was detected by Western blot analysis using polyclonal anti-RNase R antibody. Protein bands were quantified using ImageJ.

To gain insight into the overall 3D architecture and relative domain positioning of RNase R, we determined its solution structure at low resolution using SAXS techniques. As a preliminary step, and as a complement to the low resolution SAXS data, we used the I-TASSER platform (Zhang, 2008), employing the entire protein data bank (PDB) as the search space, to generate a homology model of full-length RNase R (RNase R^WT). Our results provide secondary structure predictions for the N- and C-terminal domains of RNase R (Figure 1A) that are in good agreement with previous studies (Matos et al., 2011). The three top scoring models from I-TASSER (Figure 1B) support the notion that RNase R has a structurally conserved core domain that is highly similar to the known crystal structure of RNase II (Frazao et al., 2006; Zuo et al., 2006). The I-TASSER models predicted that the unique N- and C-terminal domains, which do not share significant homology to any known structures, flank the core catalytic domain of RNase R. Notably, the C-terminal K-rich domain is predicted to contain an extensive loop segment that likely imparts significant flexibility to this domain. To further characterize the 3D shape and spatial arrangement of the N- and C-terminal domains, we collected SAXS data on RNase R^WT and RNase R⁷²³, an RNase R variant lacking the entire K-rich domain. The SAXS envelope for RNase R⁷²³ has a shape that is highly similar to RNase II. Docking of the conserved core domain into the SAXS envelope using SITUS resulted in a robust fit with a correlation coefficient of 0.88 (Figure 1C). Additional density is apparent near the N-terminus that most likely corresponds to residues 1–65 of RNase R. Having confirmed that the solution structure of RNase R displays a conserved molecular shape with RNase II, we collected SAXS data on RNase R^WT, and carried out rigid-body modeling with the conserved core domain of RNase R. We combined the core domain modeling with ab initio C-α bead modeling for the missing N- and C-terminal domains using BUNCH (Figure 1D). The resulting SAXS model of RNase R^WT reveals a tri-lobed structure that is somewhat elongated, and is consistent with the presence of extensive flexible loops, as predicted by homology modeling. The scattering patterns from RNase R⁷²³ and RNase R^WT (Figure 1E) were quite similar, reflecting their largely analogous topologies, and the SAXS model of the N-terminal and core domains of RNase R^WT correlated well with the overall shape of the RNase R⁷²³ envelope. Both models demonstrate an excellent fit to the raw scattering data (Figure 1E), with discrepancies (χ²) of 0.68 for RNase R⁷²³ and 1.43 for RNase R^WT. Visual comparison of the two models reveals the relative positioning of the K-rich C-terminal lobe with respect to the core and N-terminal domains. Both the I-TASSER and SAXS models imply intimately close links between the catalytic core and the K-rich domains of RNase R, suggesting a potential regulatory role for the K-rich domain. However, the extent of such contacts and the identity of functionally important amino acid residues in the K-rich domain that participate in any potential intra- or inter-molecular interactions have not been determined.

In an effort to uncover the identities of functionally important residues in the K-rich domain, we generated a series of alanine substitution variants of RNase R. We were guided by previous studies of the K-rich domain, which demonstrated that C-terminal truncations from residue 813 to 750 (RNase R⁷⁵⁰) had no effect on the trans-translation related activity of RNase R (Ge et al., 2010). These studies also showed that a further truncation of 15 residues in the K-rich domain, to amino acid residue 735 (RNase R⁷³⁵), resulted in a discernible defect in the ability of the enzyme to degrade non-stop mRNA. RNase R truncation variant RNase R⁷²³, which lacks the entire K-rich region, exhibited a similar pronounced defect in degradation of non-stop mRNA (Ge et al., 2010). Based on these results, we reasoned that some of the critical amino acid residues, responsible for the trans-translation activity of RNase R, must lie in the region encompassing residues 735 to 750. To evaluate the validity of this inference, we performed alanine-scanning mutagenesis of conserved amino acid residues in this region and assessed their contribution to the SmpB-tmRNA mediated targeted degradation of non-stop mRNA. To facilitate this assessment, we used plasmid-encoded reporter transcripts that are comprised of the coding sequence for the N-terminal domain of the bacteriophage λ cI gene followed by the trp operon transcriptional terminator (trpAt). To ensure evaluation of a non-stop mRNA specific process, we utilized two related versions of this reporter (Figure 2A), a non-stop transcript lacking in-frame stop codons (λ-cI-NS), and a “normal” control transcript possessing an in-frame stop codon (λ-cI-S). The λ-cI-NS non-stop transcript has been well characterized and is known to promote ribosome stalling at its 3′ end, whereas the stop codon containing λ-cI-S variant is readily translated and promotes efficient recycling of the translation machinery (Sundermeier et al., 2008; Mehta et al., 2012). The competence of RNase R alanine substitution variants in reporter mRNA decay was analyzed in a strain lacking chromosomal rnr and bearing plasmids for controlled expression of both the reporter transcript and one of the RNase R alanine substitution variants. Expression of RNase R, and its variants, was from a pACYCDuet-based plasmid bearing the rnr coding sequence under the control of the arabinose inducible P_BAD promoter.

We expressed each RNase R alanine substitution variant alongside the λ-cI-NS reporter mRNA, and measured the relative steady state abundance of the non-stop transcript. For all mRNA stability assessments, we used RNase R^WT as a positive control and the previously validated RNase R⁷²³ truncation variant (Ge et al., 2010) as a negative control. This analysis showed that various combinations of single and double alanine substitutions at residues K736, T737, and R739 had little or no effect on the steady state level of the reporter mRNA (Figures 2B,C). Similarly, single and double alanine substitutions at residues K743 and K744 had no effect on the steady state level of the non-stop reporter mRNA (Figure 2C). Alanine substitution at E740 or K741, either individually or in combination with other neighboring residues (RNase R^R739A/E740A and RNase R^K741A/K742A), resulted in measurable increases in the accumulation of non-stop mRNA (Figure 2C), indicating a functional role for these adjoining residues in the mRNA decay activity of RNase R. To further investigate their functional contributions, we generated RNase R^E740A/K741A, a double alanine substitution variant at residues E740 and K741. Assessment of RNase R^E740A/K741A revealed profound defects in non-stop mRNA degradation (Figure 2C). Alanine substitution of two additional conserved residues (K749 and K750) in the region between 735 and 750 (RNase R^K749A/K750A) resulted in extensive accumulation of non-stop mRNA, suggesting that these residues also played an important role in the selective non-stop mRNA decay activity of RNase R (Figure 3A). Collectively, these data implied that conserved residues E740–K741 and K749–K750 play a crucial role in the trans-translation mediated decay of non-stop mRNA.

To obtain a more quantitative measure of the non-stop mRNA degradation defect of the alanine substitution mutants, we selected three double-alanine substitution variants (RNase R^T737A/R739A, RNase R^E740A/K741A, and RNase R^K749A/K750A) for further evaluation. First, we examined the relative differences in the steady state abundance of non-stop mRNA among these variants, using RNase R^WT and RNase R⁷²³ as controls (Figure 3A). This analysis showed no significant difference in the ability of RNase R^WT and RNase R^T737A/R739A in the selective degradation of the λ-cI-NS non-stop mRNA. In contrast, alanine substitution variants RNase R^E740A/K741A and RNase R^K749A/K750A consistently exhibited a 2.5 to 3-fold defect in degradation of the λ-cI-NS reporter; a defect that was statistically significant and similar to the defect exhibited by RNase R⁷²³ (Figure 3A). To eliminate the possibility that the observed non-stop mRNA stability defects of these RNase R variants were due to difference in their relative abundance, we compared their steady state levels to RNase R^WT and RNase R⁷²³ under identical experimental conditions. This analysis revealed no significant differences in the steady state levels of these RNase R variants (Figure 3B), suggesting expression or stability difference could account for the observed non-stop mRNA decay defects exhibited by these RNase R variants.

Next, we measured the in vivo decay rate and half-life (t_1/2) of the λ-cI-NS non-stop mRNA in the presence of RNase R^WT, RNase R^E740A/K741A, and RNase R^K749A/K750A (Figure 4). Consistent with previous studies, RNase R^WT degraded the λ-cI-NS report transcript very efficiently, with a t_1/2 of 0.8 ± 0.3 min (Figure 4A). RNase R^E740A/K741A had a strong defect in the non-stop mRNA degradation, increasing t_1/2 by roughly 4-fold to 3.2 ± 0.4 min (Figures 4A,B). We found this deficiency in non-stop mRNA decay reminiscent of the defect observed with the RNase R⁷²³ truncation variant. Similarly, in the presence of RNase R^K749A/K750A, we detected a substantial increase in the stability of the non-stop reporter mRNA with a t_1/2 of 2.0 ± 0.1 min (Figure 4C). Based on these findings, we conclude that residues E740–K741 and K749–K750 in the K-rich domain of RNase R play important but distinct roles in the selective degradation of non-stop mRNA.

One possible mechanistic explanation for the observed non-stop mRNA decay defect of RNase R^E740A/K741A and RNase R^K749A/K750A could be that these residues (K740–K741 and K749–K750) interact directly with the catalytic core of RNase R to modulate its ribonuclease activity. To assess this possibility, we examined the exoribonuclease activity of each alanine substitution variant in an in vitro RNA degradation assay. To this end, we overexpressed RNase R^WT, RNase R^E740A/K741A, and RNase R^K749A/K750A, in an rnr^- rnb^- strain, and purified each protein to near homogeneity (Figure 5A). The ribonuclease activity of each purified enzyme was assessed by its ability to degrade the trpAt-A₁₀ substrate, a 32-nucleotide long structured RNA containing the trpA-terminator stem-loop followed by a stretch of 10 adenines at its 3′ end. Degradation reactions were initiated by the addition of the respective RNase R variant and terminated at the indicated time points by the addition of denaturing sample buffer. The reaction products were resolved by electrophoresis on denaturing polyacrylamide gels (Figure 5B), with the activity of each RNase R variant indicated by the disappearance of the full-length substrates and appearance of lower-molecular-weight 2–5 nucleotide products. Using this assay, we observed that RNase R^E740A/K741A and RNase R^K749A/K750A were as proficient in degradation of this structured RNA substrate as RNase R^WT (Figures 5B,C). This result indicated that alanine substitutions at amino acid pairs E740–K741 and K749–K750 did not compromise the ability of RNase R to unwind and degrade this RNA substrate, implying that these defined alterations did not impact intramolecular interactions of the K-rich domain with the catalytic core but rather had a specific effect on the trans-translation dependent activity of RNase R.

The SmpB-tmRNA system rescues ribosomes stalled at the 3′ end of non-stop mRNAs and recruits RNase R to initiate non-stop mRNA decay (Richards et al., 2006; Ge et al., 2010; Mehta et al., 2012). Our previous studies revealed that the inability of C-terminal truncation variant RNase R⁷²³ to degrade non-stop mRNA was due to a defect in ribosome enrichment (Ge et al., 2010). This result suggested that the K-rich domain was essential for association of RNase R with rescued ribosomes. We hypothesized that the stabilization of non-stop mRNA in the presence of RNase R^K749A/K750A and RNase R^E740A/K741A could similarly be due to their inability to productively engage tmRNA-rescued ribosomes. To examine this inference, we used the ribosome enrichment assay (Mehta et al., 2012) to analyze the effect of the K-rich domain alanine substitution variants on the association of RNase R with tmRNA-rescued ribosomes. To ensure the RNase R recruitment process was specific to the λ-cI-NS non-stop transcript, and to verify that we were examining tmRNA-rescued ribosomes, we used the stop codon containing λ-cI-S reporter as negative control, and routinely examined the captured ribosome fractions for the presence and enrichment of SmpB protein. As expected, analysis of the λ-cI-NS reporter in rnr^- cells expressing the plasmid borne full-length enzyme showed enrichment of RNase R on ribosomes expressing the non-stop reporter (Figure 6A). Quantification of ribosome enrichment data, from several independent experiments, showed greater than 3-fold RNase R^WT enrichment on ribosome translating the λ-cI-NS non-stop reporter as compared to the λ-cI-S control transcript (Figure 6B). In contrast, ribosome enrichment analysis of the RNase R^E740A/K741A variant in rnr^- cells revealed only background level of ribosome association in cells expressing either the λ-cI-NS or λ-cI-S reporter transcript (Figure 6). Although we consistently detected a small but reproducible increase in enrichment of the RNase R^K749A/K750A variant on stalled ribosomes (igures 6A,B), this level of enrichment did not fully reflect the observed increase in non-stop mRNA half-life. We attribute this difference to the fact that the ribosome enrichment assay, which involves several in vitro processing steps, is not sensitive enough to detect intermediate level enrichment of RNase R variants, whereas the rifampicin-chase assay is more sensitive in detecting subtle differences in non-stop mRNA degradation rates. Based on these findings, we conclude that the non-stop mRNA decay defect of RNase R^E740A/K741A and RNase R^K749A/K750A is due to the failure of these variants to productively engage tmRNA-rescued ribosomes.

Alanine substitution at residues E740 and K741 appears to have a profound effect on the trans-translation related activity of RNase R. Given the nature of these amino acids, it was conceivable that they interacted with each other, perhaps by forming a salt bridge required for maintaining intramolecular contacts. The presence of these contacts might be important for the non-stop mRNA degradation activity of RNase R. We reasoned that swapping the positions of these residues could preserve these interactions, and hence the trans-translation activity of RNase R on stalled ribosomes. To assess this possibility, we constructed the RNase R^E740K/K741E switch variant and performed ribosome enrichment assays. This analysis showed that the E–K switch variant (RNase R^E740K/K741E) had a strong defect in association with tmRNA-rescued ribosomes (Figure 7). This finding indicated that the position and orientation of residues E740 and K741 are important for productive engagement of RNase R with the trans-translation machinery.