A dedicated diribonucleotidase resolves a key bottleneck for the terminal step of RNA degradation

Abstract

Degradation of RNA polymers, an ubiquitous process in all cells, is catalyzed by specific subsets of endo- and exoribonucleases that together recycle RNA fragments into nucleotide monophosphate. In γ-proteobacteria, 3-'5' exoribonucleases comprise up to eight distinct enzymes. Among them, Oligoribonuclease (Orn) is unique as its activity is required for clearing short RNA fragments, which is important for cellular fitness. However, the molecular basis of Orn's unique cellular function remained unclear. Here, we show that Orn exhibits exquisite substrate preference for diribonucleotides. Crystal structures of substrate-bound Orn reveal an active site optimized for diribonucleotides. While other cellular RNases process oligoribonucleotides down to diribonucleotide entities, Orn is the one and only diribonucleotidase that completes the terminal step of RNA degradation. Together, our studies indicate RNA degradation as a step-wise process with a dedicated enzyme for the clearance of a specific intermediate pool, diribonucleotides, that affects cellular physiology and viability.

Keywords: Pseudomonas aeruginosa; Vibrio cholerae; diribonuclease; diribonucleotides; infectious disease; microbiology; oligoribonuclease.

Figure 1.. Orn has a stark preference for diribonucleotide cleavage in vitro.

RNA nucleotides two to seven residues in length (1 μM, containing the corresponding ³²P-labeled RNA tracer) were each subjected to cleavage over time with 5 nM (A, B) or 1000 nM Orn_Vc (C, D). Aliquots of each reaction were stopped at indicated times (min), and assessed by denaturing 20% PAGE (A, C). Quantification of the intensities of bands corresponding to the amount of uncleaved initial oligonucleotide over time are plotted as the average and SD of three independent experiments (B, D).

Figure 1—figure supplement 1.. Catalytic residues of Orn_Vc are required for degradation for ³²P-AAAAAGG in vitro.

Degradation of ³²P-AAAAAGG by purified indicated His₁₀-MBP-Orn_Vc variants with 5 nM (A) or 1 µM (B). Samples were stopped at indicated time (min) and analyzed by denaturing 20% PAGE. Representative gel images of two independent assays are shown.

Figure 2.. Structures reveal Orn’s conserved substrate preference for diribonucleotides.

Crystal structures of pGpG-bound V. cholerae Orn (A) and human REXO2 (B) are shown in comparison to E. coli RNase T bound to substrate (PDB 3nh1; Hsiao et al., 2011) (C), another DnaQ-fold 3’−5’ exoribonuclease with a DEDD(h) active site motif. The top panels show ribbon representations of the dimeric enzymes. The insets are surface representations of the enzymes’ active sites shown in similar orientations. The bottom panel describes the active site residues involved in RNA binding and catalysis. Residue numbering for REXO2 refers to its cytosolic isoform lacking the mitochondria-targeting pre-sequence. The sequence logos in (D) were constructed based on multi-sequence alignments of Orn and RNase T orthologs. Overall sequence identity ranges from 43 to 70% for Orn and 46 to 69% for RNase T. Sequence identifiers are provided in Figure 2—source data 2. Sequence logos were plotted using WebLogo (Crooks et al., 2004). Conserved residues of the active site’s DEDD motif (‘act.’; red), for ribonucleotide base binding (‘nuc.’; gray), and of the phosphate cap (‘P-cap’; purple) are highlighted.

Figure 2—figure supplement 1.. Structural comparison of Orn orthologs, RNase T and Exo I.

(A) Superposition of Orn_Vc with REXO2 (left) and the substrate-free state of Orn from Xanthomonas campestris (PDB 2gbz; Chin et al., 2006) (right). Rmsd values are reported for the alignment of a monomer. (B) Surface properties of Orn enzymes. An alignment of Orn orthologs was used to map conservation scores onto the solvent accessible surface of Orn_Vc (left). High degree of conservation overlaps with the acidic active site observed in the electrostatic potential map calculated with the APBS software (Baker et al., 2001) and based on Orn_Vc (right). (C) Active-site view of E. coli RNase T bound to dAAC (PDB 3v9z; Hsiao et al., 2012). (D) Active-site view of E. coli Exo I bound to dT13 (PDB 4jrp; Korada et al., 2013).

Figure 2—figure supplement 2.. Structural comparison of Orn_Vc and REXO2 bound to diverse diribonucleotides.

(A) Crystal structures of REXO2 bound to pApA and pApG. The structural overview (left) shows a C-terminal helix that is unique to Orn orthologs in higher eukaryotes and in the current crystal forms is stabilized by crystal packing contacts. (B) Crystal structures of Orn_Vc bound to di-purine (pApA, pApG, pGpA), purine-pyrimidine (pGpC, pCpG), and di-pyrimidine (pCpU) substrates. Views are identical to those shown in Figure 2 for the REXO2:pGpG and Orn:pGpG complexes.

Figure 3.. Functional importance of active-site and phosphate-cap residues for Orn function.

(A) In vitro enzyme activity. Degradation of ³²P-pGpG (1 μM) by purified wild-type Orn_Vc or variants with alanine substitutions (5 nM) at the indicated sites was assessed. Samples were stopped at the indicated times (min) and analyzed by denaturing 20% PAGE. Representative gel images are shown with the indicated RNA size. Figure 3—figure supplement 4 shows the graphs of the means and SD of three independent experiments. (B) In vivo activity of alanine substituted orn_Vc alleles to complement P. aeruginosa Δorn. Overnight cultures of the indicated strains were allowed to stand for 10 min without agitation to allow bacterial aggregates to sediment. Representative images of the cultures of triplicated assays are shown.

Figure 3—figure supplement 1.. Contribution of catalytic residues and His₁₀-MBP tag to binding interaction between Orn_Vc and pGpG.

The dissociation constants (K_d) of indicated purified (A) His₁₀-MBP-Orn_Vc variants and (B) untagged Orn_Vc to pGpG were measured by DRaCALA. Fraction bound from data was plotted and determined by nonlinear regression. All data shown represent average and SD of three independent experiments.

Figure 3—figure supplement 2.. Immunoblot analysis of Orn in P. aeruginosa ∆orn expressing indicated orn alleles.

Proteins in whole cell lysates of strains expressing the indicated HA-tagged orn allele were separated by 12% SDS-PAGE and probed with anti-HA antibodies. All images were performed in duplicate independently. A representative image of two independent experiments is shown.

Figure 3—figure supplement 3.. Molecular weight determination indicates that Orn_Vc variants with point mutations at the phosphate cap and active site remain dimeric.

(A) Absolute molecular weights (black data points across elution peaks are plotted on the right axis; theoretical monomer and dimer molecular weights, horizontal dashed lines) of wild-type Orn_Vc and point mutants were determined using SEC-MALS (90°-light scattering: colored, solid lines; refractive index signal: black, dashed lines; plotted on the left axis).

Figure 3—figure supplement 4.. pGpG degradation by purified wild-type Orn_Vc or variants with indicated alanine substitutions.

1 µM pGpG was treated with 5 nM of the indicated Orn protein and sampled at indicated time points. The graphs show quantification of three independent data sets, of which representative radiographs are shown in Figure 3A.

Figure 4.. Orn_Vc cleaves all diribonucleotides.

(A) Orn_Vc (5 nM) was incubated with di-purine (pRpR), purine-pyrimidine (pRpY or pYpR), or di-pyrimidine (pYpY) diribonucleotides (1 μM) containing the corresponding ³²P-labeled RNA tracer. Aliquots of each reaction were stopped at the indicated times (min) and assessed by denaturing 20% PAGE. (B) Quantification of the intensities of bands corresponding to the amount of diribonucleotide cleaved at the 10 min time point. Results are the average and SD of duplicate independent experiments. Orn cleaves all diribonucleotides to nucleoside monosphosphates, albeit to varying extents. Diribonucleotides consisting of two purines (pRpR) were hydrolyzed most efficiently, with over 90% of starting RNAs processed by 10 min. A majority (>75%) of diribonucleotides with a 5' purine (pRpY) were also processed by the 10 min endpoint. However, diribonucleotides with a 5' pyrimidines exhibited moderately reduced levels of cleavage. Di-pyrimidine (pYpY) substrates, and in particular pUpC and pCpC, showed the slowest turnover from the substrates tested. These results demonstrate that while all diribonucleotides are acceptable substrates for Orn, the enzyme is likely to exhibit moderate preferences for diribonucleotides that contain a 5' purine.

Figure 4—figure supplement 1.. Effect of GpG and pAp on Orn activity.

(A) Cleavage of 130 nM ³²P-pGpG by 5 nM Orn in the absence or presence of unlabeled pAp, GpG or GMP (100 µM). (B) Fraction bound of ³²P-pGpG to 200 nM Orn in presence of no competitor, 100 µM pGpG or 100 µM GpG. (C) Fraction bound of ³²P-pAp to different concentration of Orn_Vc was determined by DRaCALA. K_d could not be measured as ³²P-pAp binding to Orn was not saturated. (B) (C) All data shown represent duplicate-independent experiments.

Figure 5.. Orn acts as a diribonuclease in cell lysates.

Degradation of ³²P-GG (A) and ³²P-AAAAAGG (B) by whole cell lysates of wild-type, orn mutant, orn mutant complemented with orn_Vc, or orn_Vc D¹²A. For ∆orn and ∆orn complemented with orn_Vc D¹²A,100 nM of purified Orn_Vc was added at 30 min time point and incubated for an additional 10 min. Samples were stopped at the indicated time and analyzed by 20% denaturing PAGE. Representative gel images of triplicated assays are shown with the indicated RNA size. Graphs show quantitation of triplicate data for indicated RNA species over time.

Figure 5—figure supplement 1.. Catalytic residues of Orn_Vc are required for degradation for ³²P-AAAAAGG in whole-cell lysates.

Degradation of ³²P-AAAAAGG by whole-cell lysates of the indicated strain (PA14, ∆orn mutant complemented with pMMB-orn_Vc L¹⁸A or pMMB-orn_Vc H¹⁵⁸A. For ∆orn expressing orn_Vc L¹⁸A or orn_Vc H¹⁵⁸A, 100 nM of purified, wild-type Orn_Vc protein was added at the 30 min time point and reactions were stopped 10 min thereafter. Samples were stopped at indicated time and analyzed by denaturing 20% PAGE. Representative gel images of two independent experiments are shown.

Figure 6.. Small-colony phenotype of ∆orn is independent of c-di-GMP signaling.

Bacterial cultures were diluted and dripped on LB agar plates with the indicated concentration of IPTG. After overnight incubation, representative images of the plates were taken and shown for (A) PA14 and PA14 ∆pelA harboring pMMB vector with the indicated diguanylate cyclase gene and (B) PA14 ∆orn and PA14 ∆orn ∆pelA-G with pMMB and pMMB-orn. Experiments were performed in triplicate.

Figure 6—figure supplement 1.. Small-colony phenotype of ∆orn is complemented by functional orn_Vc alleles.

Bacterial cultures were diluted, 10 µL was spotted on LB agar plates with the indicated concentration of IPTG and drip to form parallel lines of bacteria. After overnight incubation, representative images of the plates were taken of PA14 ∆orn and PA14 ∆orn ∆pelA-G with pMMB alone or containing the indicated allele of orn_Vc. Experiments were performed in triplicate and one representative image of the plates is shown.

Figure 6—figure supplement 2.. P. aeruginosa PAK ∆orn mutants show a small-colony phenotype.

Indicated bacterial strains were diluted and dripped on (A) LB agar plates or (B) LB agar plates containing 50 µg/mL carbenicillin with indicated concentration of IPTG and incubated overnight at 37°C. All experiments were performed in triplicate and one representative image of the plates is shown.

Figure 7.. Model for Orn’s cellular function as a diribonuclease.

RNA degradation is initiated by fragmentation via endoribonucleases (e.g. RNase E and RNase III) that cleave unstructured or structured RNA sequences. RNA fragments are processed further at their 3’ termini by 3’−5’ exoribonucleases (e.g. PNPase, RNase R, and RNase II). Their combined activity produces mononucleotides and various terminal diribonucleotides from the original RNA fragments. The pGpG (GG) linear diribonucleotide is also produced by linearization of c-di-GMP by specific phosphodiesterases, EAL- and HD-GYP-domain-containing enzymes, which terminate c-di-GMP signaling. In Pseudomonas aeruginosa and likely other organisms that rely on Orn for growth, Orn is the only diribonuclease that cleaves diribonucleotides to mononucleotides. In the absence of Orn, diribonucleotides accumulate with a drastic impact on cellular physiology, ranging from transcriptional changes, small-colony phenotypes and growth arrest, depending on the organism. Orn is also unique because it acts as the second phosphodiesterase in the degradation of c-di-GMP by clearing the pGpG intermediate. In an orn mutant, c-di-GMP levels are elevated through feedback inhibition of the c-di-GMP-degrading phosphodiesterases by pGpG, leading to the associated biofilm phenotypes.

Author response image 1.. Additives to cell lystaes do not contribute to cleavage of ³²P-AAAAAGG oligoribonucleotide.

10 µg/ml DNase, 25 µg/ml lysozyme, 1 mM PMSF, and the combination of all three additives were test for cleavage of ³²P-AAAAAGG. The samples were stopped at indicated times and analyzed by 20% Urea gel. All data shown represent duplicate independent experiments.