(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



Attachment of the RNA degradosome to the bacterial inner cytoplasmic membrane prevents wasteful degradation of rRNA in ribosome assembly intermediates [1]

['Lydia Hadjeras', 'Lmgm', 'Université De Toulouse', 'Cnrs', 'Ups', 'Cbi', 'Toulouse', 'Marie Bouvier', 'Isabelle Canal', 'Leonora Poljak']

Date: 2023-01

RNA processing and degradation shape the transcriptome by generating stable molecules that are necessary for translation (rRNA and tRNA) and by facilitating the turnover of mRNA, which is necessary for the posttranscriptional control of gene expression. In bacteria and the plant chloroplast, RNA degradosomes are multienzyme complexes that process and degrade RNA. In many bacterial species, the endoribonuclease RNase E is the central component of the RNA degradosome. RNase E-based RNA degradosomes are inner membrane proteins in a large family of gram-negative bacteria (β- and γ-Proteobacteria). Until now, the reason for membrane localization was not understood. Here, we show that a mutant strain of Escherichia coli, in which the RNA degradosome is localized to the interior of the cell, has high levels of 20S and 40S particles that are defective intermediates in ribosome assembly. These particles have aberrant protein composition and contain rRNA precursors that have been cleaved by RNase E. After RNase E cleavage, rRNA fragments are degraded to nucleotides by exoribonucleases. In vitro, rRNA in intact ribosomes is resistant to RNase E cleavage, whereas protein-free rRNA is readily degraded. We conclude that RNA degradosomes in the nucleoid of the mutant strain interfere with cotranscriptional ribosome assembly. We propose that membrane-attached RNA degradosomes in wild-type cells control the quality of ribosome assembly after intermediates are released from the nucleoid. That is, the compact structure of mature ribosomes protects rRNA against cleavage by RNase E. Turnover of a proportion of intermediates in ribosome assembly explains slow growth of the mutant strain. Competition between mRNA and rRNA degradation could be the cause of slower mRNA degradation in the mutant strain. We conclude that attachment of the RNA degradosome to the bacterial inner cytoplasmic membrane prevents wasteful degradation of rRNA precursors, thus explaining the reason for conservation of membrane-attached RNA degradosomes throughout the β- and γ-Proteobacteria.

Funding: This work was supported by grants from the French National Research Agency (ANR‐13‐BSV6‐0005 to AJC; ANR-16-CE12-0014-02 AJC). LH was awarded a predoctoral fellowship from the French Ministry of Education. The work was also supported in part by the French Ministry of Research with the Investissement d’Avenir Infrastructures Nationales en Biologie et Santé program (ProFI, Proteomics French Infrastructure project, ANR-10-INBS-08 to OBS). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Considering the complexity of the ribosome assembly process, it is not surprising that occasional errors result in dead-end intermediates, which are then removed by a quality control process. Recent work has shown that degradation of rRNA in ribosomal subunits is initiated by RNase E cleavage [ 36 – 39 ]. Here, we show that ncRNase E in the bacterial nucleoid interferes with normal ribosome assembly resulting in wasteful degradation of newly synthesized rRNA. We suggest that the slowdown in mRNA degradation in the rne(ΔMTS) mutant strain is an indirect effect involving the degradation of rRNA in ribosome assembly intermediates. In the wild-type bacterial cell, we propose that degradation of dead-end intermediates in ribosome assembly is initiated by imRNase E after they are released from the bacterial nucleoid. We compare and contrast our results to recently published work with another family of gram-negative bacteria (α-Proteobacteria) in which RNase E-based RNA degradosomes are localized to ribonucleoprotein granules in the interior of the cell.

Evidence accumulating over the past 2 decades suggests that ribosome assembly and rRNA processing are spatially organized [ 24 ]. In E. coli, despite their separation by hundreds of kilobase pairs of DNA on the bacterial chromosome, most rRNA operons are in close proximity, suggesting that there is a bacterial nucleolus [ 25 ]. Ribosome assembly is a complex process that requires the coordinated synthesis of rRNA and ribosomal proteins (r-proteins) [ 26 , 27 ]. Ribosomal RNA is transcribed as a 30S precursor, which then undergoes extensive processing carried out by a battery of ribonucleases including RNase III and RNase E [ 26 , 28 – 30 ]. Imaging of chromosome spreads has shown that r-proteins bind cotranscriptionally to nascent rRNA [ 31 ]. Recent work has elucidated the structure of an rRNA transcription elongation complex that promotes cotranscriptional rRNA folding and r-protein binding [ 32 ]. Some maturation steps, such as processing by RNase III, take place on nascent rRNA in the bacterial nucleoid [ 31 , 33 ]. In addition to r-protein binding, there are many ribosome assembly factors, including rRNA modification enzymes, RNA helicases, and protein chaperons. Kinetic analyses have identified 2 intermediates in the 30S assembly (p 1 30S and p 2 30S) and 3 intermediates in the 50S assembly (p 1 50S, p 2 50S, and p 3 50S) [ 34 ]. However, recent work employing quantitative mass spectrometry and single-particle cryoEM has shown that ribosome assembly pathways are more complex than previously believed [ 26 , 27 , 35 ]. Ribosome assembly involves cooperative rRNA folding blocks that correspond to structural domains in mature 30S and 50S ribosomal subunits. Therefore, multiple parallel assembly pathways result in a heterogeneous population of intermediates.

Although RNase E is an essential enzyme in E. coli, mutant strains encoding variants in which part or all of the C-terminal noncatalytic region is deleted are viable [ 10 , 20 – 22 ]. Binding to the bacterial inner cytoplasmic membrane is disrupted in mutant strains in which the amphipathic α-helix formed by the MTS is disrupted by amino acid substitution or deletion [ 6 ]. In an rne(ΔMTS) mutant strain, the rne gene encodes a variant in which the MTS has been deleted [ 12 , 13 ]. This variant is distributed uniformly throughout the cell. For clarity, we will refer to the wild-type enzyme as inner-membrane-RNase E (imRNase E) and the variant as nucleo-cytoplasmic-RNase E (ncRNase E). In the mutant strain, growth and mRNA degradation are slower than in the isogenic wild-type control strain [ 23 ]. Although a previous study suggested that membrane localization of RNase E preferentially destabilizes mRNA encoding inner membrane proteins [ 13 ], this preference was not considered significant in a subsequent study, which concluded that the slowdown of mRNA degradation in the rne(ΔMTS) mutant strain is global [ 23 ].

Recent work in E. coli has shown that the RNA degradosome can be displaced from the inner cytoplasmic membrane under conditions of stress. Upon transition from aerobic to anaerobic growth, cells filament and RNase E localizes to the interior of the cell in a diffuse pattern [ 18 ]. Starvation for a nitrogen source results in the formation of a single large RNase E focus [ 19 ]. Treatment of cells with the protein synthesis inhibitor chloramphenicol results in the formation of RNase E foci that are not attached to the inner cytoplasmic membrane [ 16 ]. These observations suggest that stress-induced detachment of RNase E from the inner membrane could control RNase E activity or accessibility to RNA substrates. The mechanism that triggers detachment of the RNA degradosome from the bacterial inner cytoplasmic membrane remains to be elucidated.

Fluorescence microscopy of live E. coli cells has shown that RNase E is localized to the periphery with no detectable protein at the center of the cell where the nucleoid is located [ 6 , 12 , 13 ]. Systematic analyses of the inner membrane proteome have confirmed that RNase E is an inner membrane protein [ 14 , 15 ]. RNase E forms small clusters (puncta) on the inner cytoplasmic membrane [ 12 , 16 ]. RhlB and PNPase colocalize with RNase E, thus confirming association with these enzymes in live cells [ 16 ]. Inhibition of transcription by rifampicin, which depletes mRNA as well as precursors of rRNA and tRNA, results in the disappearance of puncta, thus suggesting that RNA substrate is required for RNA degradosome clustering [ 12 ]. However, recent work has shown that kasugamycin, which inhibits the initiation of translation, also results in disappearance of RNA degradosome puncta [ 16 , 17 ]. Since the synthesis of mRNA, rRNA, and tRNA continues in the presence of kasugamycin, it has been proposed that polyribosomes are necessary for RNA degradosome clustering. This experimental work suggests that inner membrane puncta are sites of mRNA degradation in which the initial step involves capture of polyribosomes by the RNA degradosome.

In Escherichia coli, ribonuclease E (RNase E) has a central role in the processing of stable RNA (rRNA and tRNA) and in the degradation of mRNA. RNase E homologues are widely distributed in bacteria and the plant chloroplast [ 1 , 2 ]. The N-terminal half of E. coli RNase E has a compact structure that is the site of endoribonuclease activity, whereas the C-terminal noncatalytic region is predominantly natively unstructured protein [ 3 – 5 ]. The noncatalytic region has small motifs (15 to 40 residues) known as microdomains or Short Linear Motifs (SLiMs) that serve as sites of interaction with proteins, RNA, and phospholipid bilayers [ 1 , 2 , 5 , 6 ]. In E. coli, the exoribonuclease PNPase, the glycolytic enzyme enolase, and the DEAD-box RNA helicase B (RhlB) bind to RNase E microdomains to form the multienzyme RNA degradosome [ 7 – 11 ]. Another microdomain, known as the Membrane-Targeting Sequence (MTS), forms a 15-residue amphipathic alpha-helix that attaches RNase E to the phospholipid bilayer of the bacterial inner cytoplasmic membrane [ 6 , 12 ]. Protein sequence comparisons have shown that RNase E homologues in the β- and γ-Proteobacteria have a conserved N-terminal catalytic domain and a large natively unstructured C-terminal half with microdomains including an MTS [ 1 ]. Thus, throughout these phyla of bacteria, RNA degradosomes are predicted to be membrane-attached.

50S particles from the rneΔMTS strain are enriched in ncRNase E, PNPase, RhlB, and enolase, which are components of the RNA degradosome, as well as SrmB and RluB ( Fig 6B ). SrmB is a DEAD-box RNA helicase that acts early in the assembly of the 50S subunit; RluB is a pseudouridine synthetase that acts late in the assembly of the 50S subunit [ 52 , 53 ]. Since it is likely that the gradient fractions analyzed here contain a mixture of particles (see Discussion ), these results suggest that the 50S fraction from the rneΔMTS contains a proportion of immature/defective particles whose degradation is initiated by the associated ncRNase E. 40S particles ( Fig 6D ) are also enriched in SrmB and RluB as well as RNase R and RimM, which is a factor involved in the assembly of the 30S subunit [ 52 , 53 ]. The enrichment of RimM suggests a noncanonical interaction with aberrant 40S particles. The 20S particle is enriched for RNase III, RNase PH, and oligoribonuclease ( Fig 6C and 6E ) as well as the protein chaperones GroEL-GroES, which have been shown to have a role in assembly of the 50S ribosomal subunit [ 54 ]. Oddly, in the rneΔMTS strain, the 30S fraction is enriched in a subset of large subunit proteins ( Fig 6C ). Since the particle peaks in velocity sedimentation on sucrose gradients are broad due to diffusion, the trailing edge of faster sedimenting particles overlaps the leading edge of slower sedimenting particles, and the 40S particle is a heterogeneous mixture of degradation intermediates (see Discussion ), the detection of large subunit proteins in the 30S subunit is likely due to contamination by 40S degradation intermediates. Proteins involved in ribosome assembly including enzymes that modify rRNA, ribonucleases, DEAD-box RNA helicases, the GroEL protein chaperone, and the ClpXP protease are associated with the 20S and 40S particles, and most of these factors are underrepresented in the 30S and 50S ribosomal subunits ( Fig 6E and S2 Table ). The underrepresented r-proteins and the associated ribosome assembly factors are supporting evidence for the conclusion that the 20S and 40S particles are aberrant dead-end intermediates in ribosome assembly.

Protein content of the ribosomal particles from rne + and rneΔMTS strains. Extracted proteins from sucrose gradient fractions were identified and quantified using a label-free quantitative mass spectrometry approach. Volcano plots showing significantly variant proteins (striped plots) in 30S particles from rne + versus rneΔMTS strains ( A ), in 50S particles from rne + versus rneΔMTS strains ( B ), in 20S versus 30S particles from rneΔMTS strains ( C ), and in 40S versus 50S particles from rneΔMTS strains ( D ) are presented. An unpaired bilateral Student t test with equal variance was used. Variant significance thresholds are represented by an absolute log2-transformed fold-change (FC) greater than 1 and a -log10-transformed (p-value) greater than 1.3 (see Materials and methods ). Small subunit proteins (green), large subunit proteins (orange), and ribosome biogenesis factors (purple) are indicated. (E ) Abundance levels of the quantified factors involved in ribosome biogenesis are represented as a percentage in 30S particle from rne + strain (light green), in 30S particle from rneΔMTS strain (medium green), in 20S particle (dark green), 50S particle from rne + strain (light orange), in 50S particle from rneΔMTS strain (medium orange), and in 40S particle (dark orange). The data underlying the graph shown in Fig 6 can be found in S2 Table .

We next analyzed protein content of the 20S and 40S particles from the rneΔMTS strain. Proteins from sucrose gradient fractions corresponding to these particles as well as to the 50S and 30S subunits of the rne + and rneΔMTS strains were extracted, digested with trypsin, and then subjected to chromatography tandem mass spectrometry (nanoLC-MS/MS), leading to the identification and quantification of 1,286 proteins ( S2 Table ). To evaluate changes in protein compositions, pairwise comparisons based on MS intensity values were performed for each quantified protein, firstly, between rne + and rneΔMTS strains for 30S and 50S particles, secondly, between 20S and 30S particles as well as 40S and 50S particles in rneΔMTS strain. Variant proteins were selected based on their significant protein abundance variations between the compared ribosomal particles (fold-change (FC) > 2 and < 0.5, and Student t test P < 0.05). Volcano plots in Fig 6A and 6B ( S2 Table ) show that composition of the 30S and 50S particles is globally the same in the 2 strains. The wild-type 30S and 50S particles are enriched in integral and associated membrane proteins (pstG, secY, sdaC, bamD, murF, ubiG, proY, ccmE, gadC, mipA, ftsY, and accY) [ 52 , 53 ]. Since the preparation of lysates for sucrose gradient analysis involves the use of sodium deoxycholate to solubilize membrane-associated ribosomes, an interaction of imRNase E as part of detergent micelles containing other membrane proteins with the ribosomal subunits could account for the detection of membrane proteins. In the rneΔMTS strain, 17 small subunit proteins and 21 large subunit proteins are significantly underrepresented in the 20S and 40S particles, respectively ( Fig 6C and 6D and S2 Table ).

RNA fragments from the gels in Fig 4 were extracted, purified, and circularized, and the region containing the 3′-5′ junction was PCR amplified (cRACE), as indicated in S6 Fig . After cloning the PCR fragments into a plasmid vector, the 3′-5′ junction was sequenced and the ends aligned with the sequence of 16S or 23S rRNA. In the diagrams representing 16S and 23S rRNA, III (blue), E (green), G (green), AM (orange), and T (orange) represent, respectively, rRNA processing sites for RNase III, RNase E, and RNase G, which are endoribonucleases and RNase AM and RNase T, which are exoribonucleases that trim intermediates to the final mature species. (A and B) Identification of cleavage sites in 16S rRNA and 23S rRNA. The color-coded key indicates cleavages that were mapped in vivo, in vitro, or both in vivo and in vitro. The number of times a site was sequenced is indicated (n). The data underlying the schemes shown in Fig 5A and 5B can be found in S1 Table . (C) Consensus sequence of rRNA cleavage sites that were mapped in vivo.

Using cRACE (circular Rapid Amplification of cDNA Ends), we mapped 16S and 23S rRNA cleavage sites in vivo in the rneΔMTS-exo − strain and in vitro using purified RNase E and ribosomes. The strategy employed in this analysis is described in S6 Fig . S1 Table , which is a tabulation of the cRACE results, shows that the 3′ ends of in vivo fragments often contain noncoded oligo(A) additions. Fig 5A and 5B are schematic diagrams indicating RNase E cleavage sites mapped by cRACE. The frequency (n) represents the number of times an end was sequenced. The color-coded key indicates cleavages that were mapped in vivo, in vitro, or both in vivo and in vitro. Cleavages in vivo in the +22 to +32 of 16S rRNA results in a nested set of fragments with raggedy 3′ ends that are likely due to partial degradation by residual 3′ exonuclease activity in the rneΔMTS-exo − strain. In Fig 5A , many of the internal cleavages in 16S rRNA were detected both in vivo and in vitro, which validates the in vitro cleavage of partially unfolded ribosomes by RNase E as a faithful representation of cleavage in vivo. Fig 5C shows the consensus sequence of rRNA cleavage sites that were mapped in vivo. The sequence is similar to the genome-wide consensus obtained from 22,000 RNase E sites in Salmonella mRNA, including the highly conserved U at position +2. [ 51 ]. These results are strong circumstantial evidence that ncRNase is responsible for cleavages of rRNA in the aberrant 20S and 40S particles.

We tested the activity of RNase E on ribosomes or rRNA in vitro ( Fig 4C ). In a high ionic strength buffer, which is necessary for stability 70S ribosomes, rRNA is resistant to RNase E cleavage. Lanes at the right of the panel show that RNase E readily degrades protein-free rRNA in the high ionic strength buffer. Digestion, which results in a smear of fragments less than approximately 600 nt in length, shows that rRNA has a large number of RNase E cleavage sites. Resistance of ribosomes to cleavage in the high ionic strength buffer shows that rRNA secondary and tertiary interactions and r-proteins protect rRNA from RNase E cleavage. The cleavage of protein-free rRNA by RNase E is slower in the low ionic strength buffer due to the limiting amount of Mg ++ , which is necessary for RNase E activity. In the low ionic strength buffer, rRNA in ribosomes is nicked by RNase E to give a series of fragments ranging from 500 to 2,000 nt in length. From these results, we conclude that a subset of RNase E cleavage sites is accessible when the ribosome is destabilized in the low ionic strength buffer. Since low ionic strength dissociates the 30S and 50S subunits, RNase E cleavage could involve exposure of the subunit interfaces as has been described in previous work on ribosome degradation under conditions of nutrient starvation [ 36 ].

Northern blots of RNA from the sucrose gradients were probed with oligonucleotides specific to the 5′ end of 16S rRNA and the 3′ end of 23S rRNA ( Fig 4B ). These blots show that the 1,000-nt RNA fragment in the 20S particles contains the 5′ end of 16S rRNA, and the 1,700-nt RNA fragment in the 40S particle contains the 3′ end of 23S rRNA. These results strongly suggest that rRNA in the 20S and 40S particles is fragmented by endonucleolytic cleavage in the body of 16S and 23S rRNA.

(A) Equal volumes of cell lysates from the rne + , exo − (left) and rneΔMTS, exo − (right) strains were separated by velocity sedimentation 5% to 20% sucrose gradients. RNA from each fraction was separated by gel electrophoresis. (B) Northern blots with probes specific to the 5′ end of 16S rRNA and the 3′ end of 23S rRNA as indicated in the diagram at the bottom of each panel. (C) Degradation of rRNA in vitro. RNase E cleavage assays were performed with purified 70S ribosomes or protein-free rRNA. A representative experiment is presented. After incubation at the indicated temperature and times, RNA was extracted, separated on 1% agarose gels, and stained with SybrGreen. Each reaction contained 0.22 μM 70S ribosome or rRNA and 0.3 μM RNase E(1–598)-HIS6. Control lanes without RNase E (−) were also included. The position of the 23S and 16S rRNA are indicated to the right of the panel. Uncropped gels of Fig 4A and 4C and northern blots in Fig 4B can be found in S1 Raw Images .

Since previous work has shown that RNase E has an essential role in ribosome quality control [ 38 ], we asked if 16S and 23S rRNA are fragmented in the 20S and 40S particles, respectively. To identify internal RNase E cleavages in 16S and 23S rRNA, we used an exo − strain background to knock down 3′ exonuclease activity and thereby increase the level of rRNA fragments. Although there are a large number of 3′ exonucleases in E. coli, RNase R and PNPase have a major role in the degradation of rRNA. Since inactivation of both genes encoding these enzymes is lethal, the exo − background combines a knockout of the rnr gene with the pnp-200 allele, which expresses a partially inactive variant of PNPase [ 49 , 50 ]. Fig 4A shows sedimentation profiles of ribosomes from strains in which the Δrnr and pnp-200 alleles were moved into the rne + and rneΔMTS strains. Total RNA was extracted from each fraction of the gradient and then separated by gel electrophoresis. As expected for both strains, full-length 23S and 16S rRNAs are found in 70S ribosomes and are present in 50S and 30S subunits, respectively. In the rneΔMTS strain, the 20S particle, which is essentially devoid of intact 16S rRNA, contains shorter RNA species that are about 1000 and 500 nt in length. The 40S particle contains 23S rRNA as well as shorter RNA species that are about 1,700 and 1,000 nt in length. In addition, fragments of about 500 nt are conspicuous in the LMW RNA fractions of the rneΔMTS strain.

Oligoadenylation of RNA 3′ ends by poly(A) polymerase, which is encoded by the pcnB gene, is known to facilitate exonucleolytic degradation by creating a single-stranded extension that acts as a binding site for exonucleases such as PNPase [ 43 – 45 ]. We therefore compared the electrophoretic profiles of total RNA in the mutant strains lacking PNPase and poly(A) polymerase ( Fig 3A ) and determined levels of 5S* rRNA by primer extension ( Fig 3B ). Isogeneic rne + and rneΔMTS strains were compared to reveal phenotypes specifically associated with the MTS deletion. In the rne + strain, there are small amounts of 5S* rRNA. Levels expressed as the ratio of 5S*/5S shows that deletion of the genes encoding poly(A)polymerase and PNPase results in a large increase in 5S* rRNA. Deletion of the gene encoding PNPase alone also results in an increase in 5S* rRNA. Deletion of the gene encoding poly(A)polymerase results in the lower levels of 5S* rRNA, which is consistent with blockage of oligoadenylation. A kinetic analysis of 5S* rRNA decay after rifampicin treatment shows that p5S rRNA in the rneΔMTS strain is degraded in a 3′ exonucleolytic pathway involving the activities of poly(A) polymerase and PNPase ( Fig 3C and S2 Data ). Since the effect of deleting both enzymes is cumulative, the poly(A)polymerase-dependent pathway likely involves RNase R activity (see [ 46 – 48 ]).

(A) Total RNA (10 μg) was extracted from strains that were grown in LB at 37°C to OD 600 = 0.4, separated on a denaturing polyacrylamide gel (10%, 7M urea) for 5 h at 300 V in 1× TBE and then stained with SybrGreen dye. (B) Total RNA (1 μg) was analyzed by primer extension with a probe specific for 5S rRNA. The position of mature 5S rRNA and its precursors are indicated on the right. The levels of 5S and 5S* were quantified by phosphorimaging. The 5S*/5S ratio is indicated at the bottom of each lane. (C) The decay of 5S* rRNA was measured after the inhibition of transcription by rifampicin (left panel). Strains and half-lives are indicated to the left of each panel. Semi-log plot of quantification by phosphoimaging used to calculate half-lives (right panel). Mean half-live and standard deviation were determined from 2 or 3 independent experiments for each strain. The data underlying the graph shown in Fig 3C can be found in S2 Data . Uncropped gels of Fig 3A-3C can be found in S1 Raw Images .

Separation of total RNA on denaturing polyacrylamide gels, which resolve small RNA species in the range of 50 to 500 nt, revealed a prominent RNA in the rneΔMTS strain, migrating slightly slower than 5S rRNA, which we named 5S* rRNA. ( Fig 3A , lane 2). Primer extension of total RNA with an oligonucleotide specific to 5S rRNA detected the presence of mature 5S rRNA 5′ ends as well as species with 5′ end extension ( Fig 3B , lane 2). We have consistently seen 2 bands located between 5S and 5S* rRNA corresponding to species with 1 or 2 nt extensions, which agrees with work showing minor heterogeneity in the 5′ end of mature 5S rRNA [ 30 , 42 ]. We gel purified 5S rRNA from the rne + and rneΔMTS strains and 5S* rRNA from the rneΔMTS strain and used RACE analyses to map the 5′ and 3′ ends of these molecules. A large proportion of the 5S rRNAs has a 5′ end corresponding to the mature molecule ( S4A Fig ). In contrast, most of the 5S* rRNAs has a 5′ AUU extension that corresponds to the p5S precursor, which is generated by RNase E cleavage of 9S rRNA precursor. Analysis of 3′ ends showed that nearly all 5S rRNA molecules have a mature 3′ end, whereas the 5S* molecules have heterogeneous 3′ ends ( S4B Fig ). A large proportion of these molecules have the 3′ CAA extension that corresponds to the p5S rRNA precursor as well as untemplated additions ranging from 1 to 4 nt. Untemplated 3′ additions are not detected in the ΔpcnB background, which lacks poly(A) polymerase activity ( S4C Fig ). The bands corresponding to 5S* rRNA ( Fig 3A , lanes 2) therefore corresponds to a mixture of p5S rRNA and oligoadenylated p5S rRNA. 3′ end analysis of RNA extracted from sucrose gradient fractions showed that p5S as well as p23S rRNA are oligoadenylated in the 40S particles from the rneΔMTS strain, whereas these RNAs were not oligoadenylated in 50S ribosomal subunits from the rne + strain ( S5 Fig ).

The presence of p5S rRNA in the rneΔMTS strain in the LMW fractions at the top of the gradient is striking ( Fig 2B ). In addition, p5S rRNA cosediments with L5 and L18 ( Fig 2A , western blots), which are r-proteins known to bind to 5S rRNA [ 41 ]. The cosedimentation of p5S with L5/L18 in the LMW fractions is specific to the rneΔMTS strain since they are almost undetectable in the rne + strain. As a control, the sedimentation of S3, a 30S r-protein, shows no differences in the rne + and rneΔMTS strains, suggesting low levels of most r-proteins in the LMW fractions. Taken together, these results suggest that a proportion of p5S rRNA that is complexed with the L5/L18 r-proteins fails to incorporate into mature 50S ribosomal subunit in the rneΔMTS strain. Since p5S rRNA is the product of RNase E cleavage, these results show that the defect in ribosome assembly is not due to a defect in RNase E processing of the 9S rRNA precursor.

Equal volumes of clarified cell lysates from rne + (left) and rneΔMTS (right) strains were fractionated by velocity sedimentation. Conditions were optimized for separation in the range of 20S to 70S. Sucrose gradient fraction numbers are indicated below the UV absorption profiles. (A) RNA from the sucrose gradient (fractions 10 to 29 for rne + strain and fractions 9 to 28 for rneΔMTS strain) was analyzed by slot blots using oligonucleotides specific to the RNA species indicated to the right of each panel. Protein from the sucrose gradient (fractions 10 to 28 for rne + strain and fractions 9 to 27 for rneΔMTS strain) was analyzed by western blotting using antibodies against the ribosomal proteins indicated to the right of each panel. (B) RNA from the sucrose gradient fractions was analyzed by primer extensions using [ 32 P] end-labelled oligonucleotides specific to the 5′ ends of 5S, 23S, 17S, and 16S rRNA. After extension by reverse transcriptase, the products were separated by denaturing gel electrophoresis. The 5′ end of mature rRNA and that of the prominent precursors are indicated to the right of each panel. Bands located between the p5S and 5S ends correspond to 5S+1 and 5S+2 species. Uncropped slot blots, western blots, and gels of Fig 2A and 2B can be found in S1 Raw Images .

To characterize the RNA composition of the 20S and 40S particles in the rneΔMTS strain, sucrose gradient sedimentation was optimized to resolve the 20S to 70S region. RNA extracted from each sucrose gradient fraction was analyzed by slot blots probed with oligonucleotides specific to 17S, p16S, 16S, p23S, 23S, and 5S rRNAs ( Fig 2A ). For comparison, we have included an analysis of sucrose gradient fractions from the wild-type strain. For both strains, the 30S subunit (fractions 18/19/20) contains 17S, p16S, and 16S rRNA, and the 50S subunit (fractions 24/25/26) contains p23S, 23S, and 5S rRNA. 17S, p23S, and 9S rRNA are released from the primary rRNA transcript by RNase III cleavage. The 17S and 9S intermediates are cleaved by RNase E to produces p16S and p5S rRNA, respectively, which together with the p23S intermediate are trimmed at their 5′ and 3′ ends by a battery of ribonucleases to produce mature 16S, 23S, and 5S rRNA [ 28 , 30 ]. The presence of the 17S intermediate in the 30S subunit and the p23S and p5S intermediates in the 50S subunit was confirmed by primer extension ( Fig 2B ). As a control, we analyzed the 5′ ends of polyribosomal rRNA by primer extension ( S2 Fig ). As expected, mature 16S, 23S, and 5S rRNA were detected in polyribosomes. Detection of rRNA intermediates in the 30S and 50S subunits shows that these are newly synthesized particles containing immature rRNA [ 26 , 34 ]. In the rneΔMTS strain, the 20S particle contains the 17S and p16S precursors of 16S rRNA; the 40S particle contains the p23S and p5S precursors of 23S and 5S rRNA, respectively ( Fig 2B ). The mapping of rRNA precursors in Figs 2B and S2 was further confirmed by 5′ RACE analyses ( S3 Fig ). Together, these results show that the 20S and 40S particles are intermediates in ribosome assembly as evidenced by the presence of precursor rRNA species.

Comparable levels of 23S and 16S rRNA in the rne + and rneΔMTS strains strongly suggests that ribosome content in the mutant and wild-type strains are comparable. Nevertheless, the slow growth phenotype could be due to a defect in translation resulting in lower protein synthesis rates. We therefore analyzed polyribosome profiles by velocity sedimentation on sucrose gradients to compare the level of 70S ribosomes to polyribosomes. Fig 1E shows representative profiles from the rne + and rneΔMTS strains. Quantification of 4 biological replicates from each strain shows that there is no significant difference in the ratio of 70S monosomes to 70S monosomes plus polysomes (M/M+P), thus arguing against a defect in translation; data available in S2 Data . However, the appearance of aberrant particles in the mutant strain with sedimentation coefficients of approximately 20S and 40S is striking. This result suggests a defect in ribosome assembly in the rneΔMTS strain that could explain the slow growth phenotype.

During the preparation of RNA for transcriptome analyses [ 23 ], we noticed an approximately 60% increase in Low Molecular Weight (LMW) RNA in the rneΔMTS strain during growth on minimal glucose medium ( S1 Fig ). When we extracted total RNA from exponentially growing strains in LB medium, we consistently obtained about 50% more RNA from the rneΔMTS strain ( Fig 1C and S2 Data ). Since RNA was extracted from cultures grown to the same density (OD 600 = 0.4) and there is only a small difference in cell size between the rneΔMTS and rne + strains, these results show a significant increase in total RNA levels in the rneΔMTS mutant strain. We fractionated total RNA on an agarose gel by loading RNA extracted from equal volumes of cultures grown to the same density ( Fig 1D ). The levels of 23S and 16S rRNA are comparable, whereas the level of LMW RNA is about 30% higher in the mutant strain. The results in Fig 1C and 1D show that the 50% increase in total RNA is at least partly due to an increase in LMW RNA. Comparison of the data in Figs 1D and S1 shows that the percent increase in LMW RNA in the mutant strain depends on growth medium. Although we have previously reported a slowdown in mRNA degradation in the rneΔMTS strain [ 23 ], it seems unlikely that the accumulation of mRNA degradation intermediates could by themselves explain the large increase in LMW RNA.

(A) Phase-contrast images. Micrographs of strains expressing either membrane-bound (rne + ) or cytoplasmic (rneΔMTS) RNase E were made at the same magnification. (B) Cell size. Lengths and widths were measured as described [ 39 ]. Scatter plots showing median cell length and width of rne + and rneΔMTS strains grown in either LB or MOPS media. Cells from 2 independent experiments (n > 100) were analyzed by ImageJ using the MicrobeJ plugin. Median length and widths (μm) are shown below each plot. P values were calculated using a parametric unpaired t test (GraphPad Prism): **** = P < 0.0001; *** = 0.0001<; P < 0.001; ** = 0.001 < P <0.01. (C) RNA yield. Cultures of the rne + and rneΔMTS strains were grown to OD 600 = 0.4 in LB medium. RNA was extracted from equal volumes of culture. Purified total RNA was eluted in equal volumes of water, and concentrations were determined by UV absorption at 260 nm. Average and standard deviation of RNA concentration from 3 independent experiments are shown. (D) Ribosomal RNA levels. Equal volumes of total RNA (Fig 1C) separated by agarose gel electrophoresis and staining with SybrGreen. Levels of 16S and 23S are comparable in the 2 strains, whereas there is 30% more Low Molecular Weigh RNA (LMW RNA) in the rneΔMTS strain as estimated by quantification of fluorescence levels (Image Lab, Bio-Rad). (E) Polysome profiles. Clarified cell lysates prepared from equal volumes of cell cultures grown to OD 600 = 0.4 in LB medium were fractionated by velocity sedimentation on 10%–40% sucrose gradients. Sedimentation is from left to right. Upper panel, rne + strain; lower panel, rneΔMTS strain. Peaks corresponding to 30S and 50S ribosomal subunits, 70S ribosomes and polysomes are indicated. 20S and 40S particles in the rneΔMTS strain are indicated. The ratio of 70S Monosomes (M) to 70S Monosomes + Polysomes (P) was measured by integrating the area of the profile corresponding to monosomes or polysomes. The average and standard deviation of the M/M+P of 4 biological replicates is shown below each profile. The data underlying the graphs shown in Fig 1B, 1C and 1E can be found in S1 and S2 Data, respectively. Uncropped gel of Fig 1D can be found in S1 Raw Images .

We previously observed that the rneΔMTS strain, which expresses ncRNase E, grows at approximately 80% of the rate of the isogenic rne + strain expressing imRNase E [ 23 ]. To investigate the slow growth rate, we first analyzed cell shape and size. Visual inspection of the micrographs in Fig 1A shows no obvious morphological difference between the rneΔMTS and rne + strains. This result suggests that the slower growth rate is not due to defective cell wall synthesis or cell division since the morphology is normal. Next, we measured cell size. In LB medium, there is a small decrease in cell length and width in the mutant strain that results in about a 10% decrease in cell size ( Fig 1B and S1 Data ). Similar results were obtained in MOPS-glycerol medium, although the difference in cell width is negligible. From these results, we conclude that the slower rate of growth of the rneΔMTS strain correlates with a small decrease in cell size, which is consistent with the known correlation between growth rate and cell size in E. coli [ 40 ].

Discussion

Here we have shown that the E. coli rneΔMTS strain expressing ncRNase E has an abnormal ribosome profile with high levels of 20S and 40S particles. 5′ and 3′ end analysis showed that the particles contain precursors of 16S, 23S, and 5S rRNA, thus supporting the conclusion that they are intermediates in ribosome assembly as opposed to intermediates in the degradation of mature ribosomal particles. rRNA in the 20S and 40S particles is fragmented by ncRNase E cleavage within the 16S and 23S sequences. Mapping of ncRNase E cleavages in the 20S and 40S particles revealed sites whose sequences correspond to the consensus previously determined by genome-wide mapping of cleavages in Salmonella [55]. In vitro experiments with purified RNase E and ribosomes showed that properly folded ribosomes are resistant to RNase E cleavage, whereas protein-free rRNA is readily degraded by RNase E. rRNA in ribosomes that are partially unfolded in vitro under low ionic strength conditions is cleaved by RNase E at sites that were mapped in vivo. From these results, we conclude that rRNA cleavage sites in intact 70S ribosomes are sequestered by rRNA folding and r-protein binding. Furthermore, as has been described in work on ribosome degradation under conditions of nutrient limitation, the 30S to 50S interface of the 70S ribosome could sequester rRNA from RNase E cleavage that has been proposed to initiate ribosome degradation [36–38].

In the rneΔMTS strain, fragments of 16S and 23 S rRNA as well as p5S rRNA have 3′ untemplated oligo(A) extensions. Oligoadenylated p5S rRNA migrates electrophoretically as a distinct species that we named 5S*. In vivo results with mutant strains showed that 3′ exonucleolytic degradation of 5S* rRNA involves the activities of PNPase and poly(A) polymerase. Measurements of 5S* degradation after rifampicin treatment showed an approximately 5-fold increase in half-life in a pnp−-pcnB− strain relative to the isogenic rneΔMTS control. It is also noteworthy that the exonucleolytic degradation pathway for 5S* rRNA is the same as previously described for several sRNA molecules as well as mRNA degradation intermediates containing REP elements [46,47,56,57]. Our velocity sedimentation results show that, in the LMW fraction at the top of the sucrose gradient, there are significant amounts of 5S rRNA precursors that cosediment with r-proteins L5 and L18, which are known to bind to 5S rRNA [41]. These results suggest that the p5S-L5-L18 complex accumulates as an intermediate in the rneΔMTS strain and that its failure to incorporate into the 50S ribosomal subunit triggers degradation.

The 20S and 40S particles are, respectively, nominally equivalent to the p 1 30S intermediate, which sediments as a 21S particle, and the p 2 50S, which sediments as a 43S particle. Nevertheless, our proteomics analysis showing that 17 small subunit proteins and 21 large subunit proteins are underrepresented in the 20S and 40S particles, respectively, is inconsistent with their identification as bona fide intermediates in ribosome assembly. We believe they are defective intermediates that accumulate due to damage by RNase E cleavage, which blocks correct rRNA folding and r-protein binding. Furthermore, it is unlikely that the particles in the 20S and 40S sucrose gradient fractions are homogeneous in composition. Recent analysis of sucrose gradient fractions in the trailing edge of the 30S peak in a wild-type strain showed that they contain a heterogeneous mixture of intermediates in ribosome assembly [58]. Similar results with intermediates in assembly of the 50S particle has led to the conclusion that ribosome assembly involves cooperative rRNA folding blocks that correspond to structural domains in the mature 30S and 50S ribosomal subunits and that there are multiple parallel pathways leading to mature 30S and 50S ribosomal subunits [27,35].

Considering the large number of ncRNase E cleavages of rRNA that we have mapped in the rneΔMTS strain, we suspect that there are multiple pathways for the interference of ncRNase E with ribosome assembly. Although RNase E cleavage sites are single-stranded, the enzyme has the capacity to bind to structured RNA [59,60]. We therefore propose that ncRNase E competes directly with cotranscriptional r-protein binding resulting in misfolded intermediates lacking r-proteins. This proposal is consistent with the finding that ncRNase E is distributed uniformly through the interior of the cell including the region in the center of the cell where the nucleoid is located [13]. These defective intermediates are then cleaved by ncRNase E, which initiates their degradation. Although cotranscriptional interference with r-protein binding might be expected to trigger rho-dependent transcription termination, the rRNA transcription elongation complex is insensitive to rho-mediated termination [32,61]. We believe that ncRNase E interference and rRNA cleavage are stochastic processes leading to a large number of different dead-end intermediates. The association of ribosome assembly factors with the 20S and 40S particles suggest that these factors are trying to “rescue” defective intermediates. However, the degradation of rRNA in these particles suggests that the damage is mostly irreversible.

Our results strongly suggest that quality control of ribosomes is mediated by imRNase E. Fig 7 is a cartoon depicting how the compartmentalization of the RNA degradosome to the inner cytoplasmic membrane protects partially unfolded intermediates in ribosome assembly from wasteful degradation. In this model, membrane attached RNA degradosomes are involved in the “trimming” of 17S rRNA to p16S rRNA and 9S rRNA to p5S rRNA [62–64]. We propose that trimming of intermediates in ribosome assembly on the inner cytoplasmic membrane occurs after the subunits are properly folded and contain a full complement of r-proteins. This leads to the suggestion that the membrane attached RNA degradosome acts as a sensor that discriminates between properly folded, functional ribosomes and partially unfolded, inactive ribosomes that are degraded by the membrane-attached RNA degradosome. However, we believe that the interference of ncRNase E with ribosome assembly is likely to be mostly cotranscriptional in the nucleoid and that normal ribosome quality control starts after intermediates are released from the nucleoid.

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 7. Quality control of ribosome assembly by the membrane-attached RNA degradosome. Cartoon depicting the synthesis of rRNA in the nucleoid, the release of early intermediates in ribosome assembly from nucleoid, maturation of late intermediates in the cytoplasm, and trimming of 17S and 9S rRNA by the membrane-attached RNA degradosome. Defective ribosomal particles are degraded by the membrane attached RNA degradosome. In this model, compartmentalization of ribosome assembly to the interior of the cell and the RNA degradosome on the inner cytoplasmic membrane shields intermediates in ribosome assembly from degradation, thus avoiding wasteful turnover of rRNA. Defective particles can be either newly synthesized intermediates that have failed to properly fold or mature ribosomal subunits that are inactive (see Discussion). https://doi.org/10.1371/journal.pbio.3001942.g007

Ribosome biogenesis is a major activity in growing cells. The time it takes for a cell to double is directly related to the time it takes to double ribosome content. Since rRNA synthesis is the limiting step in ribosome biogenesis [65], the wasteful degradation of rRNA likely explains the slower rate of growth of the rneΔMTS strain compared to the rne+ strain [6,23]. Enzymes involved in rRNA and mRNA degradation are the same [36–39,66]. Since recent work has shown the importance of competition between RNase E substrates in setting rates of mRNA degradation [67,68], competition between rRNA and mRNA degradation could explain the global slowdown in mRNA degradation in the rne(ΔMTS) strain [23]. The work reported here shows that membrane attachment of RNase E as a component of the RNA degradosome is necessary to avoid a futile cycle of wasteful degradation of intermediates in ribosome assembly. Conservation of membrane-associated RNase E throughout the β- and γ-Proteobacteria is likely due to selective pressure to avoid interference with ribosome assembly.

RNase E homologues in the α-Proteobacteria lack identifiable MTS sequences [2,17]. Recent work, principally in Caulobacter crescentus, has shown that these proteins are not attached to the inner cytoplasmic membrane. The RNA degradosome of C. crescentus is localized to the interior of the cell in condensates known as BR-bodies (Bacterial Ribonucleoprotein-bodies) that have properties similar to eukaryotic stress granules and P-bodies [69,70]. Assembly of BR-bodies is dynamic and requires RNA substrate as evidenced by rifampicin treatment. The endoribonuclease activity of RNase E is necessary for the disassembly of BR-bodies as evidenced by catalytically inactive variants of the enzyme. The intrinsically unstructured C-terminal region of RNase E, which is conserved in the α-Proteobacteria, is necessary and sufficient for BR-body formation. Characterization of the RNA content of Caulobacter BR-bodies showed that they are enriched in mRNAs and that rRNA and tRNA are excluded [71]. It was thus proposed that the Caulobacter BR-body is a compartment nucleated by the RNA degradosome in which mRNA is degraded. Selective permeability of the BR-body results in the enrichment of mRNA and mRNA decay intermediates, thus increasing their concentration and driving degradation to nucleotides, which is important for maintaining nucleotide pools for transcription and DNA replication in growing cells. Importantly, BR-bodies form a compartment that is distinct from the nucleoid and cytoplasm. These results suggest that intermediates in ribosome assembly in Caulobacter are protected from cleavage by Caulobacter RNase E due to the sequestration of the RNA degradosome into condensates that exclude ribosome precursors, ribosomes, and polysomes.

Transcription and mRNA degradation in E. coli and C. crescentus are physically separated in membraneless compartments. These bacteria, which are separated by billions of years of evolution, use different strategies to achieve similar outcomes [17]. Short-lived RNA degradosome puncta on the inner cytoplasmic membrane of E. coli are centers of mRNA degradation. The membrane attached RNA degradosome is also involved in the processing of rRNA and quality control of ribosomes. The physical separation of the RNA degradosome on the inner membrane from early steps in ribosome assembly in the nucleoid is necessary to prevent degradation of intermediates in ribosome assembly. The compartmentalization of RNA degradosomes in Caulobacter BR-bodies has functions similar to the membrane attachment of E. coli RNase E. BR-bodies are condensates in which the RNA degradosome and ribosome-free mRNA are concentrated, thus driving degradation to nucleotides. BR-bodies exclude rRNA, ribosomes, and polysomes, thus segregating ribosome assembly from mRNA degradation. Compartmentalization of the mRNA degrading machinery in E. coli and Caulobacter is a fascinating example of evolution in which different cellular organizations result in solutions to similar problems involving the accessibility of RNA substrates to the RNA degradosome and the concerted degradation of mRNA to nucleotides.

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3001942

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.