Contents

INTRODUCTION

It may seem surprising that in almost all known life-forms, information-encoding transcripts are actively annihilated. Although at first glance this seems to be a potential waste of resources and loss of information, the anticipated advantages of restricting transcript lifetimes include fast response rates and a capacity to rapidly redirect gene expression pathways. In this way, destroying individual transcripts in a modulated manner might effectively enhance the collective information capacity of the living system. Escherichia coli has proven to be a useful model system to study such processes, and nearly 45 years ago, a hypothetical endoribonuclease was proposed by Apirion as the key missing factor that might account for the observed degradation patterns of mRNA in that bacterium. At the time this hypothesis was formulated, transcript decay in E. coli was best described as a series of endonucleolytic cleavages and subsequent fragment scavenging by 3′ exonucleases (1). A few years later, Apirion and colleagues reported the discovery of the endoribonuclease RNase E and showed it to be involved in processing of rRNA precursors (2–4), and the enzyme was subsequently discovered to also cleave an mRNA from T4 phage into a stable intermediate (5). Thus, RNase E seemed to be an ideal candidate for the proposed endonuclease factor to initiate RNA decay in bacteria. What made these findings surprising was that it had previously been thought that RNases might be specialized, with one set presumed responsible for mRNA decay and another set dedicated to stable RNA processing, whereas RNase E could perform both of these distinct tasks (6). This broad functionality has been found to be a recurrent feature of other RNases in E. coli and evolutionarily distant bacteria (7).

In the ensuing decades following the discovery of RNase E, more evidence and deeper insights have been gained into the function and importance of the enzyme in RNA metabolism. The data corroborate the numerous roles played by the RNase, including the initiation of turnover for many mRNA species (8–12) and the maturation of precursors of tRNA (13, 14) and rRNA (15, 16). The roles for RNase E have been expanded to include processing and degradation of regulatory RNAs (17, 18) and rRNA quality control (19).

It is important to note that RNase E is not the sole RNase that can initiate turnover in E. coli, as others can catalyze the initial cleavage of mRNAs, including RNase G, RNase P, the double-strand-specific RNase III, and RNases from the toxin/antitoxin families (20) (see also chapter 2). Whether any particular RNA will be engaged by RNase E or another RNase is determined by enzyme specificity, substrate accessibility, and which arrives first at the scene. RNase E appears to have privileged access to substrates, and despite the apparent functional overlap with other RNases for all RNA metabolic processes, the enzyme is essential under most growth conditions (21, 22), implicating a unique and dominating role.

The access of RNase E and other RNases to substrates can be modulated by RNA-binding proteins (23). For instance, the ribosome protein S1 can shield RNase E recognition sites and protect mRNAs against cleavage (24). Ribosomes can protect mRNAs from enzyme attack during translation, but speculatively these might become accessible in a process of cotranslational decay. RNase E recognition sites can become either buried or exposed in locally formed RNA structures (25). These local structures can be induced or remodeled by base-pairing interactions formed in cis or trans, or by other binding proteins and the unwinding/remodeling activity of helicases. The actions of all these factors modulate substrate access.

In the degradation pathway of mRNA for E. coli, the initial cleavage of a transcript by RNase E is followed closely by exonucleolytic degradation of the products by PNPase (polynucleotide phosphorylase), RNase II, or RNase R (26) (Fig. 1, left). Depending on the organism, RNase E forms a complex with some of these exoribonucleases as part of a cooperative system referred to as the RNA degradosome (27). These assemblies have some mechanistic parallels and, for one component in particular—namely, PNPase—evolutionary relationship to the exosome of archaea and eukaryotes (28). The exosome complex, like RNase E, recruits RNA helicases and accessory RNases that help to achieve efficient and complete substrate degradation (29, 30) (Table 1).

Figure 1: RNase-dependent processes in bacteria. RNases play crucial roles in efficient removal of defective or unnecessary RNAs, regulation of gene expression by sRNAs, and processing of various types of RNAs. (Left) RNA degradation is initiated by endoribonucleolytic cleavage, which can be preceded by pyrophosphate removal from the primary transcript. The majority of degradation initiation events are RNase E dependent. The initial cleavage generates monophosphorylated RNA fragments that can either boost subsequent RNase E cleavage or become substrates for cellular exoribonucleases. Fragments resulting from exoribonucleolytic degradation are further converted to nucleotides by oligoribonuclease. (Middle) When RNA degradation is mediated by sRNA, sRNA-chaperone complexes (such as sRNA-Hfq) can recognize a complementary sequence near the translation initiation region and prevent ribosome association on the transcript (left branch). Naked mRNA is rapidly scavenged by endo- and exoribonucleases. The sRNA-Hfq complex can also bind within the coding region of mRNA, recruiting RNase E and promoting transcript decay (right branch). (Right) In the case of substrates for processing, the order of RNA processing can be defined by the structure of precursors and the specificity of the RNases. The processing can form a cascade of interdependent events where some target sites are being revealed only upon specific initial cleavage. RNA, dark blue; endoribonucleases, purple; exoribonucleases, light blue; sRNA, red; ribosomes, gray ovals; Hfq, orange.

Table 1 Components of the bacterial RNA degradosome and analogous or homologous assemblies from archaea and eukaryotes

a Interaction with Hfq might be indirect and mediated by RNA.

b Hydrolytic RNase component of the exosome, a member of the RNase R family and unrelated to RNase E.

In E. coli and numerous other bacteria that are phylogenetically diverse, RNase E is membrane associated (31–33), and this compartmentalization is expected to give an intrinsic temporal delay between transcription and the onset of decay (34). mRNAs encoding membrane-bound proteins were found to have shorter half-lives due to cotranslational migration of the mRNA to the membrane [33], where they might potentially become more accessible substrates for the membrane-associated degradosome.

RNase E can work in conjunction with small regulatory RNAs (sRNAs) and RNA chaperones like the Hfq protein to identify specific transcripts (35) (Fig. 1, middle). The recognition specificity is achieved through base pairing of an element in the sRNA, referred to as the seed, to the complementary region of the target. This mode of recognition resembles the use of seed pairing for target recognition by the microRNAs and small interfering RNAs of eukaryotes, and by the bacterial CRISPR (clustered regularly interspaced short palindromic repeat)-Cas9 antiphage system (36). The silencing process also bears some mechanistic analogy to the eukaryotic regulatory RNA-induced silencing complex (RISC) formed by proteins of the Argonaute family, which uses ∼22-nucleotide guide RNAs. However, bacterial sRNAs are much larger than the guide RNAs used by eukaryotes, and include structural and sequence elements used for recognition by Hfq and other facilitator proteins, such as ProQ and cold shock proteins (37, 38). The question arises how the seed-target pairings are guiding RNase E and the degradosome machinery to recognize transcripts tagged for degradation.

RNase E is also implicated in RNA processing, releasing mRNA from polycistronic transcripts and participating in sRNA, tRNA, and rRNA maturation (Fig. 1, right). In these processes RNase E is often assisted by other endo- and exoribonucleases, in tandem with which it leads to release of ready-made RNA (Fig. 1, right).

An overview is provided here of the salient structural, functional, and mechanistic features of RNase E from the perspective of accounting for the many in vivo functions of the enzyme. The accumulating data provide insights into how RNase E might operate in a cellular context as an intricate machine and interconnected hub of regulatory networks that finds and acts upon general substrates as well as specific targets with optimal speed and accuracy.

A BRIEF EVOLUTIONARY HISTORY OF RNASE E

At 1,061 amino acids, RNase E is one of the largest proteins encoded by the E. coli genome (Fig. 2). The enzyme has a conserved N-terminal domain (NTD) of roughly 510 residues that encompasses the endonucleolytic active site. The remaining C-terminal portion is predicted to be predominantly natively unstructured, and consequently this region is anticipated to lack a compact and globular character (39). Again, this gives RNase E the distinction of containing one of the largest encoded segments with this predicted disordered property within E. coli (40) (Fig. 2). E. coli and many other species also express the nonessential RNase G, a paralog of RNase E. In E. coli, RNase G is 489 amino acids in length and has roughly 30% identity and close to 50% similarity in amino acid sequence over their common 430 residues, with maintenance of the key residues involved in substrate recognition and catalysis. RNase G and E are likely to have diverged after a duplication within a chromosome early in the evolution of the gammaproteobacteria, but the enzymes still share some similar activities, including rRNA processing and mRNA turnover (34). Orthologs of RNase E and RNase G are found in roughly 80% of all bacterial genomes sequenced to date (40). Although some Gram-positive bacterial species such as Bacillus subtilis encode neither RNase E nor RNase G, they nonetheless express functionally equivalent enzymes such as RNase J1/J2 or RNase Y (41–43) (see also chapter 3).

Figure 2: RNase E catalytic domain and a model of the organization of the E. coli RNA degradosome. (Top) RNase E is a tetramer (purple, with a single protomer highlighted in dark purple), and the quaternary organization is secured through zinc coordination (black spot) linking NTDs. The CTD is predicted to be predominantly unstructured and provides binding sites for the other degradosome components: RhlB (green), enolase (yellow), and PNPase (blue). The C terminus also harbors two RNA-binding sites (red) and a membrane anchor (dark gray). (Bottom) Structure of the RNase E catalytic domain, with the subdomains of one protomer color-coded. Close view of the phosphate binding pocket (left) and the active site (right), with the main amino acids of functional importance labeled (47).

In the chloroplasts of plants, an RNase E homolog is implicated in polycistronic RNA cleavage (44), and counterintuitively, deficiency of this enzyme appears to have a destabilizing effect on some transcripts (45). Perhaps RNase E and its plant homolog could have similar roles with respect to noncoding RNAs. Some E. coli noncoding RNAs are also destabilized in the absence of RNase E, which may be due to direct interactions of RNase E and sRNA or result from changes in other mRNAs or unidentified interactors (46).

DOMAINS, MICRODOMAINS, AND PSEUDODOMAINS OF RNASE E

X-ray crystallography analysis reveals a patchwork of domains for the N-terminal catalytic region of RNase E. There is an RNase H-like subdomain at the N terminus that probably fulfills a structural function (residues 1 to 35 and 215 to 278), an RNA-binding S1 domain (36 to 118), a 5′ sensor responsible for the enzyme preference of monophosphorylated substrates (119 to 214), a DNase I domain (279 to 400), a Zn-binding domain (401 to 414) that stabilizes dimer formation, and a small domain (415 to 510) that mediates tetramer formation at the dimer-dimer interface (47) (Fig. 2).

The crystal structure of the catalytic domain of RNase E identifies two aspartic acid residues involved in metal ion coordination at the active site: D303 and D346 (Fig. 2). The metal bound at this site was proposed to be magnesium and/or manganese, which also supports RNA cleavage (48). These metals are likely to help activate water for nucleophilic attack of the phosphate backbone. From the crystallographic data and simulations, it is inferred that the scissile phosphate must approach the nucleophile in a defined geometry that is incompatible with the A-form conformation found in duplex regions (12, 47). This requirement accounts for the observation that RNase E prefers single-stranded substrates. The sites recognized by RNase E can be accompanied by stem-loop structures, situated either upstream or downstream of the cleavage site.

The C-terminal domain of RNase E (CTD; amino acids 511 to 1061) forms a scaffold region that serves as a platform for the degradosome complex (Fig. 2). Because it is predicted to be predominantly unstructured, it is not a conventional domain with compacted, folded character bearing defined secondary structural elements like strands and helices. However, the CTD is punctuated by small microdomains that have structural propensity and that recruit protein partners (39). Biophysical analyses indicate that recruitment of the partner proteins may help to partially compact the natively unstructured portion of the CTD into a conformation that may facilitate RNA interactions (49).

THE INTERACTION PARTNERS OF RNASE E

As the scaffolding core of the RNA degradosome, RNase E interacts with several types of proteins; its repertoire of partners can vary depending on growth conditions (50–53). RNase E homologs in divergent bacterial lineages have different interaction modules and partners, making the degradosome a variable machinery with capacity to perform specialized tasks depending on biological context (27, 54, 55) (Table 1).

The E. coli RNase E forms a degradosome assembly in which the canonical components associated with the CTD are a DEAD-box RNA helicase (RhlB), the glycolytic enzyme enolase, and the exoribonuclease PNPase (Fig. 2 and 3; Table 1). The Caulobacter crescentus RNase E forms an RNA degradosome complex together with RhlB, the metabolic enzyme aconitase, PNPase, and the exoribonuclease RNase D (56, 57) (Table 1). As with the E. coli degradosome, the proteins of the C. crescentus degradosome bind to RNase E mainly through its unstructured CTD, with the exception of RhlB, which binds to a partially helical insert in the S1 domain within the globular N-terminal part of RNase E. We describe the degradosomal interactions and their functional consequences in the following subsections.

Figure 3: RNase E and interactions with RNA substrates. (A) RNase E activation by 5′-monophosphorylated substrate binding (5′P depicted as a yellow star). Only the principle dimer of the RNase E tetramer (purple) is shown for clarity. The S1 domain together with 5′ sensor (red bar) is capturing the substrate (dark blue) and aligning it in the active site by structural changes induced by RNA binding. (B) Substrate (dark blue) channeling by the ATP helicase (green) to the active site of PNPase (blue). Its action may thread substrate down the channel into the active site, as occurs for the exosome and the mitochondrial exoribonuclease-helicase complex of yeast. The helicase is also likely to provide the same threading function for RNase E (purple).

The CTD of E. coli RNase E has two RNA-binding domains flanking the helicase binding site (Fig. 2). Accumulating evidence indicates that this C-terminal portion of RNase E plays an important role in sRNA-mediated regulation. In vivo, truncation of RNase E to disrupt the degradosome assembly impacts on the kinetics of substrate cleavage, as shown from results of single-molecule studies of the action of sRNA SgrS on the ptsG transcript (58). Removing the RNase E CTD diminishes the efficiency of codegradation of the sRNA-mRNA pair RyhB-sodB, implying that the degradosome might be important for presenting the RNA duplex to the catalytic domain of RNase E (17, 59). The CTD also has a membrane-association motif that compartmentalizes the assembly to the cytoplasmic membrane (27, 31, 32, 34) (Fig. 2 and 4B).

INTERACTION OF RNASE E WITH REGULATORY RNAs AND CHAPERONES

Ever since the discovery of the sRNA MicF, which represses the expression of the major outer membrane porin of E. coli, ompF (86), hundreds of small regulatory RNAs have been discovered that help to regulate gene expression in bacteria (87). Prokaryotic sRNAs vary tremendously in size and are involved in various aspects of gene regulation. These sRNAs are generally synthesized in response to stress or metabolic conditions and act by pairing to target mRNAs and regulating their translation and stability with the assistance of Hfq, a conserved RNA-binding protein from the extensive Lsm/Sm protein family. Many sRNAs studied so far lead to translational repression by binding to the translation initiation region of target mRNAs and occluding the ribosome-binding site (RBS) (88, 89) (Fig. 1, middle). sRNA-mediated degradation of target mRNAs often involves RNase E (17, 90, 91), and sRNAs paired with target mRNAs can be degraded together in an RNase E-dependent manner (17). Hfq, along with Hfq-binding sRNAs, associates with RNase E, resulting in the formation of an RNP effector complex that allows for the tethering of RNase E near the base-pairing region of target mRNAs (92).

Bacterial sRNAs that bind to their target within the coding region can alter transcript stability by creating a RNase cleavage site (93). Additionally, sRNAs can differentially alter transcript accumulation within an operon. An example is RyhB, which downregulates the iscSUA genes within the iscRSUA operon to allow independent accumulation of iscR (94). sRNAs can be degraded by RNase E and PNPase, or by RNase III if the sRNA has long stretches of complementarity (95). sRNA can act in trans to allosterically activate RNase E, which we will describe further in the next section (96).

The question arises where and how the RNase E-dependent degradation of the mRNA-sRNA hybrid is initiated (97). The key challenge faced in answering this question is that the complexes are formed transiently and are therefore difficult to capture. One experimental approach to overcome the challenge is to cross-link the RNase E-RNA complexes in vivo, in their cellular context, and then identify the components. Such an approach was employed with a variant of RNA sequencing, the CLASH methodology (UV-cross-linking, ligation, and sequencing of hybrids) (35). The data obtained using this method support an interaction/displacement model whereby RNase E binds closely to Hfq interaction sites on sRNAs, thereby displacing Hfq from the sRNA-target RNA pair. This is followed by cleavage that occurs maximally 13 nucleotides downstream of the pairing site. The model is consistent with the RNase E cleavage 6 nucleotides downstream of the MicC-ompD sRNA-mRNA duplex (93, 96).

In addition to Hfq, RNase E can cooperate with several other known RNA-binding proteins to specifically regulate sRNA stability. One example is RapZ, which was identified in E. coli as an adaptor guiding RNase E for processing of the sRNA GlmZ, as part of the mechanism controlling amino-sugar metabolism. As this metabolic pathway provides essential components required for cell envelope biogenesis, it must be synchronized with other cell division processes and part of a highly interconnected regulatory network. It has been hypothesized that RapZ guides RNase E through modification of the structure of the sRNA so it can be recognized by RNase E, or by functioning as an interaction platform by delivering the sRNA to RNase E (98, 99). Another protein cooperating with RNase E is CsrD, an RNA-binding protein that destabilizes the sRNAs CsrB and CsrC in an RNase E-dependent way, and it was suggested that CsrD might induce structural changes in the RNAs that make them more susceptible to attack by the RNase (100). RNase E cleavage sites have been mapped in RsmZ, an analog of CsrB (101). Most likely, there are more proteins like RapZ and CsrD that bind sRNAs and target them for degradation by RNase E and other enzymes.

The CTD of RNase E can also interact with regulators such as the inhibitory proteins RraA and RraB, which bind the CTD at distinct sites (102). RraA was shown to modulate degradosome activity by altering its composition through interaction with RhlB and the RNA-binding regions of the CTD (98). RraB binds within amino acids 694 to 727 of the C terminus of RNase E and also influences the degradosome composition in vivo (103). In Pseudomonas aeruginosa, the RNase E CTD is targeted by a phage protein, Dip, that inhibits its activity during phage infection (104, 105). Crystallographic and biophysical data suggest that the interaction site resides within the RNA-binding region of the CTD.

SUBSTRATE PREFERENCES OF RNASE E

Given the pivotal role of RNase E in RNA processing, turnover, and RNA-mediated regulation, the question naturally arises whether the enzyme has cleavage preferences and if these might encode information on the lifetime of the substrate. One salient signature of RNase E cleavage sites is that they are AU rich and single stranded (106–108), with strong preference for U at position +2 with respect to the scissile phosphate. U+2 is predicted to make favorable and conserved interactions with the enzyme (12). The site of cleavage by RNase E was originally mapped to a consensus sequence, A/GAUUA/U (5), in single-stranded region (5, 109, 110). This consensus sequence has been confirmed and extended by the recent mapping of >22,000 cleavage sites in Salmonella (12). This preference for single-stranded substrates is in accord with the crystallographic data, as the catalytic site cannot fit double-stranded substrates. Recently it was noted that RNase E cleavage sites are flanked by stem-loop structures (111).

A notable feature of RNase E is its 5′-end dependence for certain substrates, for which it strongly prefers 5′-mono- over the -triphosphorylated group (34, 112). This feature explains why stable 5′ stem-loops that mask the 5′ end from the enzyme protect mRNAs from decay. Crystallographic studies show that the terminal phosphate in a single-stranded region is recognized by hydrogen bonding interactions with R169 and T170 in a 5′ sensor domain, and this interaction is proposed to favor a closed conformational state that boosts enzyme activity (Fig. 2 and 3A). Comparison of crystal structures of enzyme with and without an RNA substrate analog bound shows that RNase E seems to be in an open form in the latter that closes upon substrate binding (113). This movement enables optimal orientation of substrate for catalysis and is favored by the interaction of the 5′ end with the sensing pocket (Fig. 3A). While the 5′ sensing pocket can accommodate 5′ triphosphate, due to steric hindrance the enzyme cannot close upon the substrate, and in consequence the catalytic activity is impeded. Studies with short oligonucleotides indicate that the apparent boost in catalytic efficiency for substrates with 5′ monophosphate compared to those with a 5′ hydroxyl group arises principally from the reduced Michaelis-Menten parameter, Km (114), suggesting that the effect of 5′-end sensing is mostly to contribute to substrate binding.

A second, potentially large class of RNase E substrates exists for which the enzyme activity is not affected at all by the chemical status of the 5′ end (106, 114, 115). The existence of this class has led to the hypothesis of two potential pathways for substrate recognition by RNase E: 5′-end-dependent and internal entry (116). The first pathway relies on 5′-monophosphate recognition, described above. Data from experiments with substrates with complex secondary structure suggest that the NTD can engage such substrates as a potential mechanism that contributes to the 5′ bypass mode of operation (114). Moreover, for some substrates, the internal entry pathway could be favored by interactions of RNA substrates with the NTD and arginine-rich segments present in the CTD (39, 117). While two different ways of recognizing the substrate exist, neither of those pathways is strictly essential and the pathways may work in a cooperative manner (21, 116). It is interesting in this regard to note that the truncation of the CTD of RNase E is lethal when combined with mutations in the 5′ sensing pocket in the catalytic domain, suggesting that the pathways of degradation involving 5′-end activation and RNA fold recognition are not redundant and there is a mechanistically important interplay between the CTD and NTD during substrate recognition and cleavage (21).

Early studies of RNA processing of complex substrates by RNase E indicate a role for secondary structure recognition by the enzyme. The 9S gene product is cleaved by RNase E twice to yield the 5S precursor, which is further trimmed by nucleases to give the mature 5S rRNA (15, 110, 118). Evidence indicates that a secondary structure in the 5′ region of 9S is essential for its recognition by RNase E (110). More recently, high-throughput sequencing analysis on transcriptome-wide scale in E. coli revealed that, in many mRNAs that are RNase E substrates, a stem-loop is present upstream of the cleavage site (111). It was also found that RNase E cleaves ompD mRNA in the presence of a small noncoding RNA, MicC, which guides mRNA degradation in a specific manner (96). Those two RNAs form a duplex by imperfect base pairing upstream of the cleavage site, and changes to this duplex influence RNase E activity (K. Bandyra, K. Froehlich, J. Vogel, B. Luisi, in preparation). These findings indicate that structural motifs in RNA substrates might be crucial for recognition by RNase E to help align single-stranded regions for cleavage (119, 120).

ENCOUNTERING AND ACTING UPON SUBSTRATES

RNases act on any RNA to which they have access and a match to their specificity. As they have redundant activities, the RNase that performs the first cleavage will depend on which first encounters the substrate in a permissive conformation (7, 121). As RNase E seems to be the main enzyme for initiating the onset of degradation, it must have some type of privileged access. Since degradosomes are tethered on the cytoplasmic membrane during aerobic growth, RNAs destined for processing or decay must be delivered to them, or encounter them by random diffusion.

It has been noted that the overexpression of certain transcripts results in rapid degradation by RNase E, most likely because they are produced faster than could be accommodated by the ribosomes and therefore are exposed (122). Regulatory RNA-binding proteins can modulate the efficiency of translation initiation by directly competing with ribosomes for binding to the ribosome interaction region or by initiating a change in the secondary structure of the mRNA sequence near this region (123–125). The resulting reduction in translation initiation efficiency often decreases mRNA stability as well. Moreover, emerging data suggest that codon optimality may also regulate mRNA degradation pace by influencing the rate of ribosome elongation (126). Data support a model in which mRNA degradation is prevented by efficient translation and operates through close coordination of transcription with translation. However, it might be expected that at some stage in committed translation, the transcript is no longer needed as a template, and consequently the mRNA would move off a polysome into the degradation pathway. One model for substrate access is that RNase E might encounter the emerging 5′ end of transcripts as it spools off the end of a polysome. This could be activated by an sRNA and would lead to a process of cotranslational decay (127) (Fig. 4A). Another mode of degradation might occur when the rates of translation initiation efficiency or elongation are reduced, so that the spacing of the translating ribosomes on the mRNA is less compact. In this case, it is more likely that RNase recognition sites in the mRNA would become exposed, causing transcript decay (128) (Fig. 4A). Such a degradation mode would demand the cooperation of rescue mechanisms such as the nonstop-decay process.

Figure 4: Model of the degradosome interaction with polysomes and the cellular membrane. (A) Speculative model of degradosome interaction with the polysome. RNase E can gain access to translated transcripts (dark blue) upon sRNA action (top), when an sRNA (red) in complex with Hfq (orange) targets the translation initiation region (TIR, black) and by inhibiting assembly of ribosomes provides access for RNase E. The enzyme can also gain access to translated mRNA on its own (bottom). Adapted from reference 140. (B) Association with the inner membrane (gray) is mediated by an amphipathic helix (dark gray) localized in the CTD of RNase E. RNase E, purple; RhlB, green; enolase, yellow; PNPase, blue.

CELLULAR LOCALIZATION OF RNASE E AND THE DEGRADOSOME

Adjacent to the N-terminal catalytic domain is an amphipathic α-helix that tethers the RNA degradosome to the bacterial cell membrane in E. coli and is expected to affect the way that the four natively unstructured C-terminal regions would extend outwards the tetrameric catalytic center (Fig. 4B). Strikingly, the functionally analogous (but not homologous) enzyme of Staphylococcus aureus, RNase Y, is also membrane associated (129, 130). Under anaerobic conditions, RNase E becomes destabilized and may come off the membrane (85).

Although the association of RNase E with the cytoplasmic membrane is required for optimal cell growth in E. coli, for other bacteria the RNA degradosome is not membrane localized. For example, in the alphaproteobacterium C. crescentus, RNase E forms patchy foci that associate with the nucleoid instead of the membrane (131).

The membrane-associated E. coli RNase E generates transient foci that form on transcripts in vivo (32). These foci may be cooperative degradation centers formed by several degradosome particles, and they share remarkable similarities and functional analogy with the eukaryotic RNP granules formed by RNA-binding and -processing enzymes (132). These RNP granules are microscopic structures resembling phase-separated droplets and are proposed to act as “nano-organelles” that are partitioned from the cytoplasm without the requirement for a lipid membrane. The liquid-liquid phase separation is postulated to be mediated by disordered regions of RNA-binding proteins that can form new interactions within such droplets. This phase separation not only brings about compartmentalization of the enzymes and RNA-binding proteins but also influences their specificities for nucleic acids. In the context of the degradosome, extensive disordered regions in the C-terminal tail of RNase E could promote phase separation through self-interaction or distributed contacts with RNA and association with unstructured regions of other degradosome components (97). Thus, clustering of the RNase on the cytoplasmic membrane may be a two-dimensional analog of the phase-transition behavior proposed for RNP granules and could yield highly cooperative activities on a bound substrate. The CTD of RNase E has capacity to accommodate numerous RNA species of different sequence and structure, and could contribute to the formation of such proposed granule-like foci. The CTD might help to intercept polysomes, free RNA, or structured RNA precursors for interaction with the degradosome, and orchestrate the constantly alternating RNA universe.

SUMMARY AND PERSPECTIVES

Analyses of bacterial RNA metabolism continue to reveal its numerous links to key regulatory processes. Many of the mechanisms that enable efficient RNA processing and degradation in bacteria have analogous processes in organisms of other domains of life, including metazoans. RNase E has been a key paradigm in understanding the complexity of RNA-mediated regulation and metabolism in bacteria. RNA-binding proteins, which often act in concert with RNase E, can act as global regulators to help orchestrate complex behavior. Detailed profiling of cellular targets of the sRNA chaperone Hfq, the translational repressor CsrA, the ProQ protein, and cold shock proteins has shown that there are extensive posttranscriptional networks in bacteria (38, 133–138).

There is growing appreciation of the contribution of RNA metabolism in mediating complex behavior of individual cells and cooperative communities. This behavior is also manifested during bacterial infection—for example, in the rapid response provided by sRNAs in pathogenic bacteria to coordinate invasion steps and adjust quickly to demanding and hostile environments inside the host (139). Although understanding of the processes of RNA metabolism and riboregulation is growing, there is still scope to discover some extraordinary solutions that nature has developed to improve cellular efficiency and capacity to adapt, develop, and evolve.