SEARCH

SEARCH BY CITATION

Keywords:

  • mRNA degradation;
  • Degradosome;
  • RNase E;
  • PNPase;
  • Ribonuclease

Abstract

  1. Top of page
  2. Abstract
  3. 1mRNA degradation in bacteria
  4. 2The 5′-end of mRNA
  5. 3The endoribonuclease E
  6. 4PNPase
  7. 5Polyphosphate kinase
  8. 6Helicases
  9. 7The structure of the degradosome
  10. 8The 3′-end
  11. 9Prospects
  12. References

Messenger RNAs in prokaryotes exhibit short half-lives when compared with eukaryotic mRNAs. Considerable progress has been made during recent years in our understanding of mRNA degradation in bacteria. Two major aspects determine the life span of a messenger in the bacterial cell. On the side of the substrate, the structural features of mRNA have a profound influence on the stability of the molecule. On the other hand, there is the degradative machinery. Progress in the biochemical characterization of proteins involved in mRNA degradation has made clear that RNA degradation is a highly organized cellular process in which several protein components, and not only nucleases, are involved. In Escherichia coli, these proteins are organized in a high molecular mass complex, the degradosome. The key enzyme for initial events in mRNA degradation and for the assembly of the degradosome is endoribonuclease E. We discuss the identified components of the degradosome and its mode of action. Since research in mRNA degradation suffers from dominance of E. coli-related observations we also look to other organisms to ask whether they could possibly follow the E. coli standard model.


1mRNA degradation in bacteria

  1. Top of page
  2. Abstract
  3. 1mRNA degradation in bacteria
  4. 2The 5′-end of mRNA
  5. 3The endoribonuclease E
  6. 4PNPase
  7. 5Polyphosphate kinase
  8. 6Helicases
  9. 7The structure of the degradosome
  10. 8The 3′-end
  11. 9Prospects
  12. References

Compared to eukaryotes, exhibiting messenger RNA (mRNA) half-lives in the hour range, bacteria show considerable mRNA instability in vivo with average half-lives in the order of a few minutes only. The steady-state concentration of a messenger is directly proportional to its half-life. This is part of the bacterial strategy to respond rapidly to quickly changing environmental parameters. To meet the changing requirements of the cellular household, protein synthesis can be rapidly reprogrammed by quick changes in the mRNA pattern.

mRNA degradation appears to be a major player in the regulation of gene expression in two ways. First, the time an mRNA molecule can serve as a template for ribosomal translation is determined by its inherent stability. The cellular RNA-degradation machinery removes each messenger sooner or later from the cellular circulation, depending on stabilizing or destabilizing structural features of the RNA. Secondly, mRNA degradation, regulated itself by environmental parameters, can provide a means to actively adapt the cellular translation to changing bacterial needs. While there is no evidence for environmental parameters directly controlling translational rates in prokaryotes, such an influence on the stability of mRNA species has been observed. The rate of the puf mRNA decay in Rhodobacter capsulatus for example is significantly enhanced in the presence of oxygen [1]. This polycistronic messenger encodes several protein components of the photosynthetic complex in this purple bacterium and is optimally expressed only under conditions of low oxygen tension. The stability of the cspA mRNA, encoding the major cold-shock protein in Escherichia coli, is transiently increased upon lowering growth temperature [2]. The mRNA for the RNA binding protein Rbp of Anabaena variabilis is specifically stabilized upon lowering growth temperature [3]. E. coli cat and Bacillus subtilis thrS mRNA show increased stability under starving conditions [4,5]. This possibly hints at a general phenomenon for other mRNAs, when cells are under conditions of C or N starvation [6,7]. It is notable that in eukaryotic systems RNA sequences have been identified as the direct sensors of environmental stress [8].

The abundance of polycistronic primary transcripts is a prokaryotic peculiarity. Not only 5′- and 3′-untranslated but also intercistronic regions organize the sequence of coding regions. The association of several genes in one transcript does not necessarily mean that equimolar amounts of each protein are translated. Ribosomes probably bind independently to the beginning of each individual cistron and disengage from the transcript at various internal termination sites. Fine tuning of differential translation initiation, elongation, and termination kinetics in individual segments can lead to different molar amounts of translated protein. Autogenous regulation of subsets of proteins in polycistronic messengers adds to the repertoire of translational control mechanisms. Such mechanisms have been extensively studied in operons like those for E. coli ribosomal proteins or the E. coli atp operon (ATPase) [9,10].

In polycistronic transcripts mRNA degradation makes a significant contribution to differential gene expression. Extensive studies over the last years on the polycistronic puf mRNA of R. capsulatus revealed that differential segmental stability exerts one of the determining influences on the protein stoichiometry during translation of the respective fragments (see Fig. 1) [11]. Proteins coded in this polycistronic transcript are needed in different stoichiometric amounts in order to assemble functional photosynthetic complexes. As a consequence of a variety of mRNA stabilizing and destabilizing structural elements within the primary transcript, nucleolytic activities split the polycistronic RNA into smaller mRNAs. These segments exhibit quite different half-lives, which in turn leads to different molar amounts of translated protein.

image

Figure 1. The puf operon as an example for stabilizing and destabilizing RNA structures. These structures influence the differential segmental stability of the polycistronic messenger. This operon encodes the LH I and RC proteins of the photosynthetic complex in R. capsulatus. Proteins coded by B and A are needed in a molar ratio of 15:1 over the proteins coded by segments L and M. This is reflected in the half-lives of the corresponding mRNA segments. While the primary transcript and the transcript trimmed to 2.7 kb have half-lives of less than half a minute or 8 min, respectively, the downsized pufBA has a half-life of 33 min and is therefore available much longer for translation. Red hairpins stabilize upstream sequences. The green hairpins form a complex 5′ stabilizer. The yellow box signifies an internal RNase E cleavage site responsible for rate limiting decay of the 2.7-kb messenger.

Download figure to PowerPoint

Similar effects have been studied in a number of other polycistronic transcripts like the E. coli malEFG, the Salmonella histidine transport operon, the E. coli ftsA-ftsZ transcript, the Acinetobacter calcoaceticus mop operon, and the Alcaligenes eutrophus hoxS mRNA [12–16].

To describe the mRNA degradation machinery in the bacterial cell it is necessary to look at two aspects. First, we have to understand the nature and distribution of stabilizing or destabilizing structural elements in mRNA. We then also have to investigate which ribonucleases are involved in mRNA degradation. Although the cleavage of an internucleotide bond requires the action of an endo- or exoribonuclease, recent results have made it clear that RNA degradation also requires enzymes other than that.

Since the subject of mRNA stability was reviewed thoroughly for the last time results of most recent research have begun to provide a general picture of a highly ordered mRNA degradation in the bacterial cell [17,18]. The field, though, suffers somewhat from the fact that E. coli dominates the scene. Results obtained from E. coli, using a handful of model RNAs, may not be representative at all for other bacteria.

mRNA degradation experiments very often require measuring the half-life of an mRNA. This can be achieved either with in vivo or with in vitro experiments. During in vivo studies mRNA transcription initiation in a living cell is brought to a halt by the addition of a transcription inhibitor to a life culture at time zero. Aliquots of the culture are then removed at various time points to isolate RNA and separate it on a gel. Hybridization techniques using labeled probes, specific for a particular mRNA of interest, are then used to detect and quantify the presence and decay kinetics of this mRNA (chemical half-life). Since RNases show overlapping activities or can even replace each other, it is difficult to distinguish the individual effects they may have on mRNA degradation. The elimination of one or more nuclease genes from a strain with an otherwise identical genetic background can improve the interpretation of complex degradation patterns of model substrates. Bacterial strains carrying temperature-sensitive mutant genes of proteins involved in mRNA degradation are subjected to temperature shifts thus revealing the impact of the ts gene on degradation pathways and kinetics of mRNA decay [19]. In vitro experiments are taking advantage of the ease of RNA transcript availability, once an RNA is cloned under the control of a phage promoter like T7 or SP6. Labeled run-off transcripts can be prepared with in vitro transcriptions. Depending on the construct, only parts of a messenger, 5′- and 3′-untranslated regions (UTR), or complete polycistronic regions can be included. These in vitro substrates are then incubated with bacterial extracts or purified proteins during in vitro degradation assays. This approach is of great advantage when studying the effects of mutated sites in RNA substrates on mRNA degradation. Using extracts rather than purified activities may lead to ambiguous results due to overlapping activities from various components. The 5′- and 3′-ends of observed RNA fragments are determined using primer extension analysis and S1 nuclease mapping, respectively, thus revealing the position of cleavage sites in the in vitro substrate.

It is particularly misleading to think of an mRNA molecule as being straight like a knitting needle. Most RNases act as RNA maturation nucleases. The majority of RNA molecules in an organism are transcribed as biologically inactive precursor molecules, which in order to come up to their biological tasks first require extensive processing, internally or at their termini. As is the case for most protein-RNA interactions, the recognition between an RNA substrate and the nucleolytic processing activity during maturation and also degradation does not simply rely on the recognition of the primary sequence. Very often the actual nucleolytic cleavage site comes already prepackaged with quite a lot of surrounding secondary and tertiary structure. The standard repertoire of secondary structure elements like stems, loops, bulges and ss regions of various thermodynamic stability is sufficient to create a variety of structured RNA signals.

To understand the regulated expression of a gene product it can be helpful to study the degradation of the corresponding mRNA. An analysis of possible secondary structures, the comparison with corresponding regions in the genomes of other organisms can point to possible regulatory features of this RNA. Once such structures are postulated, they will of course require verification using chemical and enzymatic probing methods for RNA structure. Proof for postulated structures in stability-related regulation can also be obtained using mutational analysis.

2The 5′-end of mRNA

  1. Top of page
  2. Abstract
  3. 1mRNA degradation in bacteria
  4. 2The 5′-end of mRNA
  5. 3The endoribonuclease E
  6. 4PNPase
  7. 5Polyphosphate kinase
  8. 6Helicases
  9. 7The structure of the degradosome
  10. 8The 3′-end
  11. 9Prospects
  12. References

To organize our way through mRNA degradation we start at the 5′-end of the messenger. The triphosphate in the 5′ position of the first nucleotide is the very beginning of a messenger. Internal fragments, which result from downsized transcripts, instead carry a 5′ monophosphate. That this is not trivial will become clear once we look at the catalytic activity of an enzyme like RNase E. The coding region with the translation start codon is preceded by the 5′-UTR. This 5′-UTR extends up to several hundred nucleotides and can exhibit quite a bit of secondary structure. Also part of this 5′-UTR is the Shine-Dalgarno (SD) sequence, closely spaced 5′ to the start codon. The 5′-end not only serves as a potential entry point for degradative nucleases, thus becoming automatically a preferred point to deploy protective measures, but is of course also the entry point for ribosomes during translation. It is therefore not surprising that research on possible 5′ stabilizing structures always had to address the possible influence of translating ribosomes. A ribosome protects a bound messenger RNA to an extent of 15 nt downstream of the first P-site nucleotide [20]. Although, contrary to eukaryotes, a 5′[RIGHTWARDS ARROW]3′ exonuclease activity in prokaryotes was never found, mRNA degradation often shows a net 5′[RIGHTWARDS ARROW]3′ directionality. For many years the standard model of prokaryotic mRNA decay depicted decay as the result of the combined action of endo- and exoribonucleases. In a first step an endonuclease – we now generally assume this to be RNase E – would recognize a cleavage site and a rate determining endonucleolytic cut occurs. Frequently these sites are upstream of stem-loop protected 3′-ends, but often they are close to the extreme 5′-end, even upstream of 5′-stem loops. In the former case, cleavage is then followed by a second step of rapid exonucleolytic degradation through a 3′[RIGHTWARDS ARROW]5′ exoribonuclease. In the latter case, RNase E recognizes an accessible 5′-end and after cleavage further downstream, degradation of fragments by a 3′[RIGHTWARDS ARROW]5′ exonuclease follows [21,22]. After an initial cut at the 5′-end, the endonuclease would follow the progressing chain of ribosomes to the next RNase E site. Scanning mRNA from the 5′-end for further cleavage sites, RNase E would thus create the impression of 5′[RIGHTWARDS ARROW]3′ progressing degradation. Events at the 5′-end have been studied in several models. One of them is RNA 1, the antisense regulator of ColE1. This RNA comprises nine single-stranded nucleotides at its 5′-end, which include an RNase E site, followed by 99 nt folded into three stem-loops (see Fig. 2). The introduction of a stem-loop at the extreme 5′-terminus stabilizes this RNA, even when the single-stranded stretch is extended [21]. Such a stem-loop would therefore act as a 5′ stabilizer. The degradation rate of transcripts with single-stranded 5′-ends is independent of the sequence of the ss region, the distance to the downstream internal RNase E cleavage site, and the potential of the sequence leading to this cleavage site to form secondary structures. With this system considerable evidence was provided that it is endonuclease RNase E which is responsible for the degradation pattern at 5′-ends of bacterial messengers. While it is easier to comprehend that 3′[RIGHTWARDS ARROW]5′ digestion of the 3′-end is often blocked by 3′ stem-loops, it is far more difficult to understand the degradation initiation originating from the 5′-end, since 5′[RIGHTWARDS ARROW]3′ exonucleases could not be found in prokaryotes until now. Bacterial 5′ stabilizers thus cannot be directed against 5′[RIGHTWARDS ARROW]3′ exoribonucleases. To give meaning to the concept of 5′ stabilizing RNA elements, a true 5′ stabilizer should be able to confer stability in trans on other, heterologous transcripts when attached to them, otherwise stability may just simply reflect a particular local situation of a messenger.

image

Figure 2. 5′ Stabilizer in the RNA I substrate. The highly structured RNA I substrate can be stabilized with a 5′ stabilizer hairpin when inserted at the extreme 5′-terminus. With a destabilizing ss region of at least 3–5 nt at the 5′-end a 5′ stabilizer is no longer able to block degradation.

Download figure to PowerPoint

In E. coli the ompA 5′-UTR acts as the cis mRNA stabilizer of the growth-rate-regulated ompA. Under rapid growth conditions a half-life of the message of approx. 17 min is observed, as opposed to 4 min under slow growth. Indeed, the 5′-UTR of the ompA mRNA acts as a transferable mRNA stabilizer when abutted to the 5′-end of other, even long labile mRNAs [23,24]. For example, the lacZ message can thus be stabilized while still being normally translated. It is primarily the ompA 5′-UTR structure which stabilizes, not a change in the density of translating ribosomes. As part of a chimeric ompA 5′-UTR-bla gene, the ompA 5′-UTR can stabilize the short-lived bla message [25]. Ribosomes in this case initiate at the ompA SD sequence and translate into the coding region of bla. Attaching the entire bla transcriptional unit (including its own 5′-UTR) to the ompA gene (including its 5′-UTR stabilizer) creates a chimeric dicistron. Even in this dicistron the translationally independent bla message is markedly stabilized compared to wild-type bla, even when early codons in bla are replaced with stop codons, forcing premature release of ribosomes [26]. In wild-type bla replacing early codons with stop codons further destabilizes the already labile message [27].

The important structural features of the ompA 5′-UTR have been studied using mutational analysis of parts of this structure (see Fig. 3). A stem-loop, no more than 2–4 nt from the extreme 5′-terminus, and a nearby positioned ss region 2 are important. This SS2 region contains the ribosomal binding site and a AUG start codon. More than 4 nt 5′ to hairpin I cause destabilization of the messenger. The detailed stem-loop features are rather unimportant as long as the stem-loop is at the direct terminus. The primary sequences of stem-loop and ss region are not evolutionarily conserved in ompA 5′-UTRs. A missing stem-loop I causes degradation downstream of the AUG. The second stem-loop, 11 nt downstream of the first, is already not important.

image

Figure 3. Structural organization of a 5′-UTR. The ompA 5′-UTR consists of two functionally important regions: the 5′ proximal hairpin I and the single-stranded region SS2 which contains a Shine-Dalgarno sequence and the start codon AUG.

Download figure to PowerPoint

Mutational analysis of the ss segment 2 indicates that it stabilizes by mere saturation binding of ribosomes to a strong SD element rather than by translation initiation. Replacing the AUG with an inefficient CUG start codon reduces translation dramatically without influencing the stability of the messenger. The fact that one 5′ stem-loop can completely protect the full length of an entire mRNA against downstream nucleolytic events strongly supports the assumption of an initial 5′ contact of the endonucleolytic activity prior to any downstream degradative action. It does not seem, though, that 5′ stem-loops have to be unwound to allow passage of a nuclease. The question whether the downstream nucleolytic activity is RNase E remains open for this model system [24]. Using a bla construct that contains the ompA 5′-UTR in place of the bla 5′-UTR, Arnold et al. [24] state that although ribosome density in the untranslated region is more important, ribosomes also have to pass into the coding region. Forced premature translation termination on early codons destabilizes the messenger.

5′ Stabilizers have also been described for the Rhodobacter pufBA messenger and the E. coli rnc operon RNA [28,29]. In the case of the pufBA message an apparently more complex 5′ stem-loop structure is necessary to stabilize this internal dicistron which is generated through removal of additional upstream and downstream cistrons. In the rnc operon a 5′ stabilizing hairpin works in conjunction with a downstream RNase III cleavage site.

The most convincing model for degradation originating at the 5′-end, the role of 5′ stabilizers, and 5′[RIGHTWARDS ARROW]3′ directionality of degradation has been developed for the rpsT transcript which encodes ribosomal protein S20 [30,113]. The amount of rpsT is strongly controlled by translational efficiency even in the absence of key RNases. The translation initiation step itself, rather than simply covering the full messenger length with ribosomes, must serve as an explanation for stabilization [31]. It became clear that the 5′ stem-loop of rpsT impedes RNase E cleavage more than 20 nt downstream. Initially RNase E recognizes ss 5′-ends containing a cleavage site. Mackie and colleagues propose that in a second step the assumed RNase E dimer would start scanning for downstream cleavage sites while still attached to the first site [32]. Very recently, through a series of rather neat experiments using circularized mRNAs and 5′ antisense oligos, Mackie could show that substrates with 5′-ends masked in this way are inefficiently processed by RNase E [30]. The clear result was that RNase E requires an unpaired 5′-end for unrestricted access to cleavage sites further downstream in the coding region. It does not make a difference whether RNase E is present as an isolated enzyme or as part of the multi-enzyme complex, the degradosome, catalyzing degradation (see below). It also became clear that RNase E requires the terminal nucleotide to carry a 5′ monophosphate. Similar effects of the phosphorylation status of the terminal base have been noted before, but have not been attributed to RNase E [33]. These data immediately imply the need for a specially dedicated phosphatase activity converting 5′ triphosphates to monophosphates, thus earmarking such messengers for recognition by RNase E. It would therefore appear to be justified to call the prokaryotic 5′ triphosphate the equivalent of the eukaryotic protective 5′ cap structure. This model would make an elegant explanation for the observed 5′[RIGHTWARDS ARROW]3′ directionality of RNase E and the influence of translational efficiency on messenger stability. The RNase E is directed to the 5′-terminus through monophosphate recognition. Scanning for a cleavage site from this position would then interfere with the translation initiation process or the translation of early codons. Internal fragments which result from nucleolytic processing of polycistronic transcripts have 5′ monophosphates at their respective 5′-ends. This may require different strategies for their 5′ protection against degradation [28].

Currently there is no consensus whether the 5′ stabilizer model is generally valid for the majority of mRNAs and for all bacteria. Joyce and Dreyfus [34] take quite a different stand describing 5′ protection in lacZ mRNA of E. coli requiring ribosomes to travel at least 1 kb into the coding region for stabilization of the full-length transcript. In the absence of translation, 5′ stabilizers do not offer effective protection against downstream RNase E cleavage in this messenger. The authors rather question the concept of E. coli 5′ stabilizers protecting a majority of messengers against degradation. In their opinion E. coli degradation protection relies primarily on translation. A series of conditions is cited where downshifting translation efficiency causes mRNA instability. In Bacillus instead, several cases are reported where stalled ribosomes in the 5′-UTR stabilize long downstream sequences, whether translated or not [35].

3The endoribonuclease E

  1. Top of page
  2. Abstract
  3. 1mRNA degradation in bacteria
  4. 2The 5′-end of mRNA
  5. 3The endoribonuclease E
  6. 4PNPase
  7. 5Polyphosphate kinase
  8. 6Helicases
  9. 7The structure of the degradosome
  10. 8The 3′-end
  11. 9Prospects
  12. References

This enzyme has been already mentioned several times as the apparent key enzyme for the initiation of mRNA degradation. It has been the research on RNase E which has provided the most substantial progress in our understanding of mRNA degradation during recent years. In a recent review on RNase E the enzyme was called “a still wonderfully mysterious enzyme”[36]. The essential rne gene in E. coli encodes an enzyme of 1061 amino acids (see Fig. 4). rne autoregulates its own synthesis by repressing the concentration of the transcript through increased decay rates [37]. The calculated molecular mass of the coded protein is 118 kDa, but RNase E runs at a 180-kDa position, probably due to a proline-rich C-terminal region [38]. This aberration was the reason for many confusions of the past. The N-terminal half contains the catalytic center. Although initially evidence was presented that the central RNA binding region from amino acids 580 to 700 is required for catalytic activity, it is now generally assumed that the presence of this arginine-rich RNA binding domain is not necessary for activity [39–41]. This domain has similarity to the RNA binding site of the 70-kDa human U1 snRNP protein [38,42].

image

Figure 4. RNase E as the key enzyme in bacterial RNA degradation. The N-terminus carries regions of significant homology with the CafA and S1 proteins (see text). The C-terminal half shows instead only poor conservation with islands of clear amino acid composition preference. It is dedicated to protein-protein interactions with other degradosome components and to RNA binding.

Download figure to PowerPoint

The RNA cleaving activity of RNase E has been studied with a variety of standard substrates such as the 9S rRNA precursor of 5S rRNA, RNA I (the antisense regulator of ColE1), the mRNAs encoding OmpA, TrxA, and ribosomal proteins S20 and S15 (rpsT and rpsO). Cleavage products contain 5′ phosphates and 3′ hydroxyl groups, the cleavage preferentially occurs in single-stranded A/U-rich regions [43]. The original proposals describing the structural constraints of an RNase E cleavage site in E. coli required a loose single-stranded consensus sequence (A/G)AUU(A/U) and additional lateral stabilizing stem-loop anchors [44,45]. The stem-loops would either serve as entry points or keep the target site single-stranded [46]. Interestingly, a cleavage site of this type was also identified in R. capsulatus, an organism with a much higher GC content than E. coli[47]. McDowall et al. [48,49] recently reexamined the cleavage requirements using oligonucleotides mimicking the RNase E cleavage site in RNA I. Although the enzyme shows a preference for single-stranded A/U-rich regions, a simple consensus sequence or a need of secondary structure (stem-loops) in the immediate 5′ or 3′ vicinity of the cleavage site could not be confirmed with this in vitro system. An influence of the overall substrate conformation instead is possible. A systematic survey by Arraiano et al. [50] of the endonucleolytic cleavage sites in the E. coli dicistronic pyrF-orfF transcript revealed a total of 27 sites and a 12-nt potential consensus sequence with a 4-nt A/U-rich core, without clear need for additional secondary structure. The question what RNase E recognizes clearly deserves more attention in the future. Possibly SELEX-based approaches (for systematic evolution of ligands by exponential enrichment) will help to define an RNA structure which serves as a perfect substrate. Again, structure and not primary sequence appears to be the key for this RNA-protein recognition.

For those interested in protein structure RNase E offers particular riches, not surprising for a protein of this size. The N-terminal half shows extensive homology with the E. coli CafA protein (for cytoplasmic axial filament protein) [51]. CafA is a 50-kDa protein of unknown function possibly acting on cytoskeletal elements [52]. RNase E could possibly have additional functions in intracellular RNA transport. It cross-reacts with myosin heavy chain AB and shows sequence similarity with myosin-type molecules [38,51]. The CafA region of homology also embraces an S1 domain homology. When Bycroft et al. [53] determined the structure of the E. coli PNPase S1 RNA binding domain (approx. 70 amino acids long) they showed that this domain, which was originally identified in ribosomal Protein S1, was also present in the immediate N-terminus of RNase E and at the C-terminus of PNPase and RNase II. The fact that the S1 RBD (for RNA binding domain) is present in three different RNases may indicate that they share common regulatory features. The authors furthermore point out that the S1 domain fold has a high similarity to the cold shock domain (CSD) found in cold shock proteins (CSP). These proteins apparently function as RNA chaperones in the eubacterial cell under a variety of stress situations. The CSD, in conjunction with other RNA binding domains, mediates specific RNA binding. Possibly, CSPs lock RNAs in a linear state making them accessible for translation or degradation [54].

The year 1994 brought a real breakthrough in RNase E related biochemistry. Carpousis et al. – using proper salt, detergent, and protease inhibitor conditions – could for the first time present a purification scheme which consistently produces full length RNase E from E. coli of 118 kDa (i.e. 180 kDa on a PAA-SDS gel) [55]. The purification not only produced intact RNase E, but also made it immediately clear that RNase E to a major extent exists in the bacterial cell as part of a high molecular mass complex. First it was shown that in this complex RNase E is associated with polynucleotide phosphorylase (PNPase) and other unknown proteins which coprecipitate with anti-RNase E antibodies. PNPase migrates as 85-kDa (α subunit) and 48-kDa (sometimes called β subunit) bands. The identity of a 50-kDa band remained unknown. The complex, christened ‘degradosome’, is another example of a complex multi-component protein machine like the proteasome, spliceosome, ribosome, and nuclear pore. These machines drive the most important cellular processes. With their internalized conformational changes of highly coordinated protein parts they create efficiently working micro-environments for catalysis, in this case for RNA degradation [56]. The observed complex has considerable size heterogeneity of 8–16S or 160–460 kDa, respectively. RNase E proteolysis disrupts the complex, but also uncomplexed forms of full-length RNase E are observed. The authors suggest a 2:3:1.3:3.4 stoichiometry for RNase E:PNPase α:50-kDa:PNPase β.

Shortly after, Py et al. purified from E. coli a 500-kDa complex which contains RNase E, PNPase, and other unidentified proteins [57]. Immunoprecipitation of the complex with anti-PNPase antibodies precipitated bands of 180 (RNase E), 104, 85 (PNPase α subunit), 74, and 49 kDa. Minor bands of 92, 74, 52, 47, 45, and 30 kDa were also observed. In this complex the authors could also identify an activity which impeded the most important 3′[RIGHTWARDS ARROW]5′ exoribonuclease, PNPase. This exoribonuclease-impeding-factor does not seem to be an RNA binding protein itself but rather copurifies with PNPase [58]. Similar effects of 3′[RIGHTWARDS ARROW]5′ exonuclease stalling at stem-loops located at the 3′-end of a messenger RNA have also been observed in B. subtilis[59]. The need for an additional protein factor to explain the strong protection kinetics against 3′[RIGHTWARDS ARROW]5′ exonucleolytic processivity by 3′ stem-loops was already expressed by McLaren et al. [60]. Two protein bands of the original Carpousis preparation were subsequently identified as the 50-kDa RhlB protein (for RNA helicase-like), and the 48-kDa enolase [61]. This would make enolase the β subunit of PNPase, but this nomenclature has been called into question [62]. We will discuss the role for helicases in mRNA degradation further on.

4PNPase

  1. Top of page
  2. Abstract
  3. 1mRNA degradation in bacteria
  4. 2The 5′-end of mRNA
  5. 3The endoribonuclease E
  6. 4PNPase
  7. 5Polyphosphate kinase
  8. 6Helicases
  9. 7The structure of the degradosome
  10. 8The 3′-end
  11. 9Prospects
  12. References

E. coli PNPase (pnp) is generally described as a homotrimer of 78-kDa subunits, running at a slightly higher gel position of 86 kDa [63]. PNPase and RNase II (rnb) are the two most important 3′[RIGHTWARDS ARROW]5′ exonucleases in E. coli[64]. PNPase is a processive Pi-dependent exonuclease with polyadenylation activity. The enzyme in its phosphorolytic mode makes mononucleotide diphosphates which inhibit the protein. RNase II instead is a 70-kDa monomeric protein which creates 5′ NMP. The distribution of total exonuclease activity in E. coli between the two enzymes is 90% for RNase II vs. 10% for PNPase [65]. Both enzymes appear to be inter-regulated in E. coli[66]. A double-minus mutant in E. coli is lethal [67]. In B. subtilis the importance of both enzymes is rather reversed [68]. We will also return to PNPase, once we discuss the complex reactions of the degradosome at the 3′-end of mRNA.

5Polyphosphate kinase

  1. Top of page
  2. Abstract
  3. 1mRNA degradation in bacteria
  4. 2The 5′-end of mRNA
  5. 3The endoribonuclease E
  6. 4PNPase
  7. 5Polyphosphate kinase
  8. 6Helicases
  9. 7The structure of the degradosome
  10. 8The 3′-end
  11. 9Prospects
  12. References

Polyphosphate kinase (PPK) is the most recently identified component of the degradosome [69]. The homotetrameric enzyme, build from 80-kDa subunits, catalyzes the reversible conversion of poly(P) and ADP to ATP [70]. E. coli is a rich source of this enzyme. Interestingly, the protein sequence is highly conserved between numerous bacteria, while it has not been found in several completely sequenced bacterial genomes [71]. Blum et al. [69] could show that PPK interacts with RNase E and that a ppk-minus strain shows increased mRNA stability. Polyphosphate which could originate from various sources has a strong inhibitory effect on RNA degradation [70]. PPK may remove inhibitory poly(P) and NDPs in the degradosome.

Other proteins which are discussed as putative components of the degradosome are GroEL and DnaK [62,69,72,73].

Not discussed as a potential component of the degradosome, and of rather limited importance in mRNA degradation, is RNase III (see review Nicholson, this volume). RNase III is a dsRNA-specific RNase. The consensus sequence of the cleavage sites is rather preliminary [74]. This RNase cleaves only a small set of mRNAs, most prominently at hairpins in the 5′-UTR of pnp and rnc. Initial cleavage of these 5′ stabilizers is a prerequisite for subsequent further degradation.

6Helicases

  1. Top of page
  2. Abstract
  3. 1mRNA degradation in bacteria
  4. 2The 5′-end of mRNA
  5. 3The endoribonuclease E
  6. 4PNPase
  7. 5Polyphosphate kinase
  8. 6Helicases
  9. 7The structure of the degradosome
  10. 8The 3′-end
  11. 9Prospects
  12. References

When Py et al. [61] identified the 50-kDa protein of the degradosome as the RhlB protein they were also able to show the function of this ATP-dependent DEAD-box helicase (DEAD stands for the conserved amino acid motif Asp-Glu-Ala-Asp found in this protein family). Helicases are proteins that unwind nucleic acid duplexes, whether DNA or RNA. Based on sequence homology they can be grouped into five vast superfamilies (SF) [75,76]. RNA helicases fall into SF I and II. They have sets of seven conserved sequence motifs involved in Mg2+, ATP, and RNA binding. These helicases are presumably dimeric proteins and they require single-stranded overhangs to attack a duplex. The DEAD-box RNA helicases are a subset of the SF II helicases and share a core region of approximately 300 amino acids, with eight highly conserved motifs [77]. Subfamilies are DExH- or DEAH-box proteins. Based on this the authors used an RNA substrate with a hairpin separating 5′ and 3′ single-stranded regions. This substrate is derived from the malE-malF intercistronic region. When incubated with degradosome fractions in the absence of ATP, this substrate is degraded by PNPase 3′[RIGHTWARDS ARROW]5′ processivity until PNPase hits the base of the stem (see Fig. 5). Further degradation is only observed when ATP is added. Apparently, RhlB unwinds the hairpin in an ATP-consuming reaction and thus aids the degradation of structured RNA by PNPase, which is otherwise arrested when it meets secondary structures such as hairpins. Helicases as part of the degradosome must therefore be assigned a major role in RNA turnover [78].

image

Figure 5. Helicase action in the degradosome. Degradosome-dependent 3′[RIGHTWARDS ARROW]5′ degradation of stable 3′ stem-loops requires the ATP-dependent unwinding activity of the degradosome's helicase component.

Download figure to PowerPoint

Instead of the RNase E preference for 5′ monophosphate ends, the catalytic force of the degradosome in this context appears to be directed primarily at the 3′-end of an RNA [30]. How does the structural organization of the complex fit into the picture?

7The structure of the degradosome

  1. Top of page
  2. Abstract
  3. 1mRNA degradation in bacteria
  4. 2The 5′-end of mRNA
  5. 3The endoribonuclease E
  6. 4PNPase
  7. 5Polyphosphate kinase
  8. 6Helicases
  9. 7The structure of the degradosome
  10. 8The 3′-end
  11. 9Prospects
  12. References

The catalytic center for endonucleolytic activity in RNase E is exclusively located in the N-terminal half of the protein (see Figs. 4 and 6) [40,41,79]. In a study directed at understanding the domain structure of RNase E, Vanzo et al. [62] analyzed the role RNase E plays in the assembly of the E. coli degradosome. The C-terminal half of RNase E contains distinct binding sites for the three degradosome components DEAD-box-helicase RhlB, enolase, and PNPase. Contact apparently is made on a one to one basis between RNase E and the ligand, while there is no direct interaction between the three. In particular, no interaction between PNPase α subunit and enolase was observed, making further questionable the fact that originally enolase was named β subunit of PNPase [61]. Vanzo et al. also showed direct evidence for oligomerization of RNase E, which was already proposed by Mackie [32]. The RhlB helicase is highly activated when in contact with the binding domain of RNase E. RNase E, apart from its own catalytic abilities, with its C-terminus is the assembly platform for the degradosome. A truncated C-terminal half of E. coli RNase E can bind the degradosome components PNPase, RhlB, and enolase [80].

image

Figure 6. The prokaryotic degradosome. This scheme presents a model for the structural organization of the degradosome acting on 3′-ends. The various pools for phosphate in this micro-environment are ortho-phosphate Pi, poly-phosphate (Pi)n, mononucleotide diphosphates NDP, and ATP. The inhibitory or stimulatory influence of these pools on mRNA degradation is indicated by + or −. NDPs inhibit PNPase, poly-phosphate probably inhibits the helicase. The model reflects the current ideas about the interaction of known degradosome components. PPK, poly-phosphate kinase; PNPase, polynucleotide kinase.

Download figure to PowerPoint

Our lab has recently purified the degradosome from R. capsulatus (Fuhrmann, Rauhut and Klug, unpublished results). An RNase E of the apparent ‘180’-kDa type, RhlB helicase, enolase, and PNPase are present in the complex and, most interestingly, a second DEAD-box RNA helicase of 65 kDa. This complex would therefore be able to operate in a similar mode as that proposed for E. coli.

It is now time to look at the currently known RNases E to see whether all of them are fit to act as degradosome assemblers. A recent compilation of RNases E compares the enzymes from E. coli, Haemophilus influenzae, Synechocystis, Porphyria chloroplasts, and Mycobacterium tuberculosis[80]. The enzymes from Synechocystis and Porphyria chloroplast only have a reported length of approximately 50% of the E. coli and H. influenzae proteins. The reported parts are clearly homologous with the N-terminus of the latter two. The E. coli, H. influenzae, and the M. tuberculosis enzymes share the same length. The former two have highly homologous N-termini, their C-termini being highly divergent. M. tuberculosis shows limited homology over the full length. The just published sequence of the Rickettsia prowazekii genome revealed the presence of an only 683 amino acid long RNase E. Only the N-terminal two thirds show the usual high sequence homology (with a 90-amino acid insert in the ultimate N-terminus). If the situation of dramatically shortened C-termini alone should already give cause to doubt a general presence of an RNase E-mediated degradosome, then the more so when we consider the fact that almost all of the completely sequenced bacterial genomes lack a detectable RNase E like protein. Evolution not only found different solutions for the C-terminus of RNase E, but also for the entire activity exemplified by RNase E. The 13.3-kDa protein ARD-1 from human cells is a functional analogue of E. coli RNase E, shares structural features, and can rescue rne-deficient E. coli strains [81,82].

It is interesting, though, to look at degradosome-like complexes which have been described in cellular organelles like chloroplasts and mitochondria, both of bacterial origin. A chloroplast high molecular mass complex with PNPase and an RNase E-like enzyme was described [83,84]. The RNase E-like activity has a molecular mass of 67 kDa and is recognized by E. coli anti-RNase E antibodies. In yeast mitochondria the mtEXO complex contains three major proteins with 3′[RIGHTWARDS ARROW]5′ exonuclease activity and is involved in mtRNA degradation [85]. The complex shows NTP requirement for nuclease activity and NTPase activity, both indicative of the presence of a helicase. One component was recently identified as DExH-box RNA helicase [86]. The yeast nucleus contains a different type of high molecular mass complex, the exosome [87]. This complex of 300–400 kDa harbors five 3′[RIGHTWARDS ARROW]5′ exonucleases, mostly homologues of E. coli 3′[RIGHTWARDS ARROW]5′ exonucleases, but no helicases. All nucleases are vital parts needed for precursor rRNA processing. Interacting with ATP-dependent RNA helicases like the yeast Dob1p and Ski2p, the exosome probably has an important role in the mRNA turnover of all eukaryotes [88–90]. Also the E. coli degradosome was shown recently to be involved in rRNA degradation in E. coli[91].

8The 3′-end

  1. Top of page
  2. Abstract
  3. 1mRNA degradation in bacteria
  4. 2The 5′-end of mRNA
  5. 3The endoribonuclease E
  6. 4PNPase
  7. 5Polyphosphate kinase
  8. 6Helicases
  9. 7The structure of the degradosome
  10. 8The 3′-end
  11. 9Prospects
  12. References

In our journey through the RNA molecule we now have to look into structures and processes particular to the 3′-end. And it is presumably this end which underwent the most far-reaching rethinking. Stem-loops at the 3′-end of transcription units and in intercistronic regions are a prominent feature in prokaryotes [92]. Both types can serve as a protective measure against 3′[RIGHTWARDS ARROW]5′ exonucleases [12,93,94]. Most prokaryotic mRNAs display a 3′-UTR downstream of the translational stop codon. 3′-UTRs contain intrinsic rho-independent transcription termination stem-loops which are difficult to distinguish from true 3′ stabilizers [60,64,95]. A recent computer analysis of completed prokaryotic genome projects revealed that the majority of the organisms do not form hairpins in their 3′-UTRs [96]. This finding has profound implications, not only for transcription termination, but also for mRNA degradation.

Surprisingly, there is more at the end. During recent years it has become clear that also prokaryotic mRNAs are polyadenylated [97]. In prokaryotes the poly(A) tails of mRNA can be up to 60 nt long. In E. coli poly(A) tails are probably between 15 and 50 nt long [98]. Li et al. [99] proposed that polyadenylation is a general feature in E. coli, even for the small stable RNAs. While poly(A) tails in eukaryotes exert a stabilizing influence on messenger RNAs, they seem to do just the opposite in prokaryotic RNAs (see also the article on eukaryotic polyadenylation in this volume) [100]. Two poly(A) polymerases of 53 and 35 kDa have been identified in E. coli, PAPI and PAPII [101,102]. A similar situation is observed in B. subtilis[103].

PNPase and RNase II can only exert 3′[RIGHTWARDS ARROW]5′ processivity on 3′-ends of more than a dozen or so unpaired nucleotides, located 3′ to the 3′ stem-loop [104]. And this is where polyadenylation enters the scene. Possibly it is polyadenylation which facilitates access to a transcript for one or more 3′[RIGHTWARDS ARROW]5′ exonucleases. In the presence of poly(A) tails the amount of substrate sequence accessible for PNPase digestion increases dramatically. In the absence of poly(A) polymerase the degradation of ompA, trxA, and rpsO mRNAs in E. coli slows down instead [98,105]. Similarly, RNA I decay accelerates through 3′ adenylation [33]. PNPase- or RNase II-deficient E. coli strains show a strong increase in number and length of poly(A) tails [106]. Adding 3′ As and thus recruiting PNPase and the degradosome complex to the 3′-end provides an attractive model to overcome the roadblock of 3′ stabilizing structures. It is therefore tempting to speculate about additional poly(A) binding proteins and poly(A) polymerase itself as part of the degradosome. Until now none have been found in the degradosome [61]. On the other hand, it has been reported for the RNA I substrate that adding long poly(A) tails prevents endonucleolytic cleavage of this substrate by RNase E. RNase E can shorten such poly(A) tails 3′[RIGHTWARDS ARROW]5′ in an exonucleolytic but not strictly processive mode! In this case, shortening of the poly(A) tail down to a critical threshold length makes the RNA accessible for degradation [41].

An important part of our knowledge about events in the 3′-end originates from observations made in an organelle of prokaryotic heritage, the chloroplast. Most chloroplast mRNAs have a stem-loop in their 3′-UTR which functions as an upstream stabilizer, but they are not polyadenylated. Those that are, can exhibit poly(A) tails (or rather poly(A)-rich tails with interspersed Gs) of considerable length [107]. In spinach chloroplasts polyadenylation apparently destines an RNA for degradation by exonucleases. A protein homologue of bacterial PNPase, the exoribonuclease 100RNP, which preferentially cleaves poly(A) tails, and a PAP-like activity have been found in chloroplasts. The chloroplast PNPase was shown to be associated with an RNase E-like protein in a high molecular mass complex. This complex contains additional proteins which bind to the 3′ stem-loop thus providing protection from degradation [84]. PNPase shows a high affinity for poly(A) tails and is closely associated with PAP [83,108]. Polyadenylation and the presence of a degradosome-like structure in chloroplasts lend some credibility to the idea that chloroplasts and bacteria still use a common mechanism for mRNA degradation. First, endonucleolytic cleavage in the 3′ coding region or the 3′-UTR of mature RNAs by an RNase E-like endonuclease would produce RNA without a 3′ stem-loop. Subsequent polyadenylation of these ‘free’ 3′-ends would make them the preferred high affinity substrate for PNPase [83,109].

If a bacterial 3′ stem-loop is very stable or associated with exonuclease impeding factors, thus strongly protecting against degradation by stalling exonucleases at the 3′ base of the stem, a rate-limiting RNA cleavage at an RNase E site upstream of the stem-loop is another means to remove this obstacle [58,60,110]. The new unprotected 3′ entry site leads to rapid degradation of the messenger by RNase II and PNPase. Coburn and Mackie [111] recently tried to put the 3′-end model to the test in an in vitro reconstituted degradation system. They used the 3′-terminal degradation fragment of the rpsT mRNA (ribosomal protein S20). This fragment is a product of RNase E-dependent degradation and comprises part of the coding region and the 3′-UTR. Degradosomes cannot degrade this fragment even when adenylated. Instead, cycles of polyadenylation and PNPase action completely degrade the fragment in vitro. So the in vitro reconstituted degradation system requires only RNase E (initial internal cut), PNPase, PAP I, ATP and Pi. The fact that ATP can be substituted with non-hydrolyzable ATPγS, suitable for PAP I but not for RhlB, indicates that the action of RhlB helicase is not required. Surprisingly, RNase II can counteract the degradation pathway by removing poly(A) tails from PNPase-degradable intermediates [110]. Without poly(A) tails degradosomes should no longer act from the 3′-end, but instead are forced to use the endonucleolytic pathway which depends on RNase E cleavage of upstream RNase E sites. It has been reported for RNA I RNA that the presence of a long poly(A) tail indeed prevents endonucleolytic cleavage of the primary transcript by RNase E. Transcripts with short or no tails are cleaved instead [41].

Citing, among other evidence, that in vivo half-lives and mRNA levels for rpsT are the same in PAP I-deficient strains, Coburn and Mackie [111,113] do not believe that polyadenylation is a general prerequisite for mRNA decay.

Depending on the stability of the 3′-terminal structures a well-tuned set of degradative tools is applied instead: unstructured ends are readily available for exonucleolytic attack (RNase II, PNPase), more stable stem-loop structures like in RNA I require a one-time polyadenylation [33]. Stable stem-loops, as in rpsT, require repeated action of PAP I and PNPase as shown in Fig. 7. Only the most stable structures require the degradosome for degradation, complete with helicases and ATP hydrolysis [61]. Cleavage initiated by RNase E here at the 3′-end can be well combined with degradation protection by translating ribosomes [112]. The rpsO messenger (encoding ribosomal protein S15) forms a dicistron with the downstream pnp message. RNase E cleaves 10 nucleotides downstream of the rpsO stop codon. This cleavage removes a hairpin downstream of the cleavage site which in turn is a prerequisite for subsequent degradation by 3′[RIGHTWARDS ARROW]5′ PNPase. The distance between the stop codon and the cleavage site is critical for stability. Extending it destabilizes the messenger, bringing the cleavage site outside the protection range of a bound ribosome.

image

Figure 7. Polyadenylation as an alternative for the degradosome. This scheme, adapted from [111], depicts the complex interaction of 3′[RIGHTWARDS ARROW]5′ exonucleases and PNPase at the 3′-end of the rpsT mRNA (ribosomal protein S20). Polyadenylation and exonuclease activities are sufficient for degradation, while helicase activity is not required for this hairpin. RNase II is depicted as poly(A) tail shortening exonuclease, thus making RNAs less susceptible to degradation. PAP I, poly(A) polymerase I

Download figure to PowerPoint

9Prospects

  1. Top of page
  2. Abstract
  3. 1mRNA degradation in bacteria
  4. 2The 5′-end of mRNA
  5. 3The endoribonuclease E
  6. 4PNPase
  7. 5Polyphosphate kinase
  8. 6Helicases
  9. 7The structure of the degradosome
  10. 8The 3′-end
  11. 9Prospects
  12. References

What will be the future direction of RNA degradation research? If RNA degradation is a process truly responsive to environmental changes, the mechanisms have to be understood which transmit a particular environmental signal such as lower temperature or decreasing oxygen partial pressure to a distinct target molecule of the degradative process, so that specific changes in the half-life of only certain RNAs can be achieved.

Considerable progress has been made in the purification and physical characterization of some of the key enzymes involved, especially RNase E. By now, after many years of troubled purification history, we have a better understanding of the protein than we have of the substrates of RNase E. The structural requirements which determine a cleavage site need to be investigated in more detail employing new techniques and the rather limited set of substrates currently used for RNA degradation studies needs to be extended.

The future will show whether the degradosome-like machinery is really the prototype for bacterial RNA degradation. More organisms should be studied to correct a picture severely biased by E. coli. The degradosome possibly contains more components which need to be identified. A variety of minor protein bands in current purification protocols are promising candidates for new exiting functions, to give an answer to questions such as what attracts a degradosome to the 3′-UTR, the 5′ monophosphate or an internal cleavage site.

References

  1. Top of page
  2. Abstract
  3. 1mRNA degradation in bacteria
  4. 2The 5′-end of mRNA
  5. 3The endoribonuclease E
  6. 4PNPase
  7. 5Polyphosphate kinase
  8. 6Helicases
  9. 7The structure of the degradosome
  10. 8The 3′-end
  11. 9Prospects
  12. References
  • [1]
    Klug, G. (1991) Endonucleolytic degradation of puf mRNA in Rhodobacter capsulatus is influenced by oxygen. Proc. Natl. Acad. Sci. USA 88, 17651769.
  • [2]
    Goldenberg, D., Azar, I., Oppenheim, A.B. (1996) Differential mRNA stability of the cspA gene in the cold-shock response of Escherichia coli. Mol. Microbiol. 19, 241248.
  • [3]
    Sato, N., Nakamura, A. (1998) Involvement of the 5′-untranslated region in cold-regulated expression of the rbpA1 gene in the cyanobacterium Anabaena variabilis M3. Nucleic Acids Res. 26, 21922199.
  • [4]
    Kuzj, A.E., Medberry, P.S., Schottel, J.L. (1998) Stationary phase, amino acid limitation and recovery from stationary phase modulate the stability and translation of chloramphenicol acetyltransferase mRNA and total mRNA in Escherichia coli. Microbiology 144, 739750.
  • [5]
    Condon, C., Putzer, H., Grunberg-Manago, M. (1996) Processing of the leader mRNA plays a major role in the induction of thrS expression following threonine starvation in Bacillus subtilis. Proc. Natl. Acad. Sci. USA 93, 69926997.
  • [6]
    Albertson, N.H., Nystrom, T. (1994) Effects of starvation for exogenous carbon on functional mRNA stability and rate of peptide chain elongation in Escherichia coli. FEMS Microbiol. Lett. 117, 181187.
  • [7]
    Kahn, D., Hawkins, M., Eady, R.R. (1982) Metabolic control of Klebsiella pneumoniae mRNA degradation by the availability of fixed nitrogen. J. Gen. Microbiol. 128, 30113018.
  • [8]
    Spicher, A., Guicherit, O.M., Duret, L., Aslanian, A., Sanjines, E.M., Denko, N.C., Giaccia, A.J., Blau, H.M. (1998) Highly conserved RNA sequences that are sensors of environmental stress. Mol. Cell Biol. 18, 73717382.
  • [9]
    McCarthy, J.E., Gualerzi, C. (1990) Translational control of prokaryotic gene expression. Trends Genet. 6, 7885.
  • [10]
    McCarthy, J.E. (1990) Post-transcriptional control in the polycistronic operon environment: studies of the atp operon of Escherichia coli. Mol. Microbiol. 4, 12331240.
  • [11]
    Klug, G. (1993) The role of mRNA degradation in the regulated expression of bacterial photosynthesis genes. Mol. Microbiol. 9, 17.
  • [12]
    Newbury, S.F., Smith, N.H., Higgins, C.F. (1987) Differential mRNA stability controls relative gene expression within a polycistronic operon. Cell 51, 11311143.
  • [13]
    Stern, M.J., Prossnitz, E., Ames, G.F. (1988) Role of the intercistronic region in post-transcriptional control of gene expression in the histidine transport operon of Salmonella typhimurium: involvement of REP sequences. Mol. Microbiol. 2, 141152.
  • [14]
    Cam, K., Rome, G., Krisch, H.M., Bouche, J.P. (1996) RNase E processing of essential cell division genes mRNA in Escherichia coli. Nucleic Acids Res. 24, 30653070.
  • [15]
    Schirmer, F., Hillen, W. (1998) The Acinetobacter calcoaceticus NCIB8250 mop operon mRNA is differentially degraded, resulting in a higher level of the 3′ CatA-encoding segment than of the 5′ phenolhydroxylase-encoding portion. Mol. Gen. Genet. 257, 330337.
  • [16]
    Oelmuller, U., Schlegel, H.G., Friedrich, C.G. (1990) Differential stability of mRNA species of Alcaligenes eutrophus soluble and particulate hydrogenases. J. Bacteriol. 172, 70577064.
  • [17]
    Nierlich, D.P., Murakawa, G.J. (1996) The decay of bacterial messenger RNA. Prog. Nucleic Acid Res. Mol. Biol. 52, 153216.
  • [18]
    Belasco, J.G. and Brawerman, G. (Eds.) (1993) Control of Messenger RNA Stability. Academic Press, San Diego, CA.
  • [19]
    Arraiano, C.M., Yancey, S.D., Kushner, S.R. (1988) Stabilization of discrete mRNA breakdown products in ams pnp rnb multiple mutants of Escherichia coli K-12. J. Bacteriol. 170, 46254633.
  • [20]
    Hartz, D., McPheeters, D.S., Green, L., Gold, L. (1991) Detection of Escherichia coli ribosome binding at translation initiation sites in the absence of tRNA. J. Mol. Biol. 218, 99105.
  • [21]
    Bouvet, P., Belasco, J.G. (1992) Control of RNase E-mediated RNA degradation by 5′-terminal base pairing in E. coli. Nature 360, 488491.
  • [22]
    Lin-Chao, S., Cohen, S.N. (1991) The rate of processing and degradation of antisense RNAI regulates the replication of ColE1-type plasmids in vivo. Cell 65, 12331242.
  • [23]
    Emory, S.A., Bouvet, P., Belasco, J.G. (1992) A 5′-terminal stem-loop structure can stabilize mRNA in Escherichia coli. Genes Dev. 6, 135148.
  • [24]
    Arnold, T.E., Yu, J., Belasco, J.G. (1998) mRNA stabilization by the ompA 5′ untranslated region: two protective elements hinder distinct pathways for mRNA degradation. RNA 4, 319330.
  • [25]
    Belasco, J.G., Nilsson, G., von Gabain, A., Cohen, S.N. (1986) The stability of E. coli gene transcripts is dependent on determinants localized to specific mRNA segments. Cell 46, 245251.
  • [26]
    Hansen, M.J., Chen, L.H., Fejzo, M.L., Belasco, J.G. (1994) The ompA 5′ untranslated region impedes a major pathway for mRNA degradation in Escherichia coli. Mol. Microbiol. 12, 707716.
  • [27]
    Nilsson, G., Belasco, J.G., Cohen, S.N., von Gabain, A. (1987) Effect of premature termination of translation on mRNA stability depends on the site of ribosome release. Proc. Natl. Acad. Sci. USA 84, 48904894.
  • [28]
    Heck, C., Rothfuchs, R., Jager, A., Rauhut, R., Klug, G. (1996) Effect of the pufQ-pufB intercistronic region on puf mRNA stability in Rhodobacter capsulatus. Mol. Microbiol. 20, 11651178.
  • [29]
    Matsunaga, J., Simons, E.L., Simons, R.W. (1996) RNase III autoregulation: structure and function of rncO, the posttranscriptional ‘operator’. RNA 2, 12281240.
  • [30]
    Mackie, G.A. (1998) Ribonuclease E is a 5′-end-dependent endonuclease. Nature 395, 720723.
  • [31]
    Rapaport, L.R., Mackie, G.A. (1994) Influence of translational efficiency on the stability of the mRNA for ribosomal protein S20 in Escherichia coli. J. Bacteriol. 176, 992998.
  • [32]
    Mackie, G.A., Genereaux, J.L., Masterman, S.K. (1997) Modulation of the activity of RNase E in vitro by RNA sequences and secondary structures 5′ to cleavage sites. J. Biol. Chem. 272, 609616.
  • [33]
    Xu, F., Cohen, S.N. (1995) RNA degradation in Escherichia coli regulated by 3′ adenylation and 5′ phosphorylation. Nature 374, 180183.
  • [34]
    Joyce, S.A., Dreyfus, M. (1998) In the absence of translation, RNase E can bypass 5′ mRNA stabilizers in Escherichia coli. J. Mol. Biol. 282, 241254.
  • [35]
    Sandler, P., Weisblum, B. (1989) Erythromycin-induced ribosome stall in the ermA leader: a barricade to 5′-to-3′ nucleolytic cleavage of the ermA transcript. J. Bacteriol. 171, 66806688.
  • [36]
    Cohen, S.N., McDowall, K.J. (1997) RNase E: still a wonderfully mysterious enzyme. Mol. Microbiol. 23, 10991106.
  • [37]
    Jain, C., Belasco, J.G. (1995) RNase E autoregulates its synthesis by controlling the degradation rate of its own mRNA in Escherichia coli: unusual sensitivity of the rne transcript to RNase E activity. Genes Dev. 9, 8496.
  • [38]
    Casaregola, S., Jacq, A., Laoudj, D., McGurk, G., Margarson, S., Tempete, M., Norris, V., Holland, I.B. (1992) Cloning and analysis of the entire Escherichia coli ams gene. ams is identical to hmp1 and encodes a 114 kDa protein that migrates as a 180 kDa protein. J. Mol. Biol. 228, 3040.
  • [39]
    Taraseviciene, L., Bjork, G.R., Uhlin, B.E. (1995) Evidence for an RNA binding region in the Escherichia coli processing endoribonuclease RNase E. J. Biol. Chem. 270, 2639126398.
  • [40]
    McDowall, K.J., Cohen, S.N. (1996) The N-terminal domain of the rne gene product has RNase E activity and is non-overlapping with the arginine-rich RNA-binding site. J. Mol. Biol. 255, 349355.
  • [41]
    Huang, H., Liao, J., Cohen, S.N. (1998) Poly(A)- and poly(U)-specific RNA 3′ tail shortening by E. coli ribonuclease E. Nature 391, 99102.
  • [42]
    Claverie-Martin, F., Diaz-Torres, M.R., Yancey, S.D., Kushner, S.R. (1991) Analysis of the altered mRNA stability (ams) gene from Escherichia coli. Nucleotide sequence, transcriptional analysis, and homology of its product to MRP3, a mitochondrial ribosomal protein from Neurospora crassa. J. Biol. Chem. 266, 28432851.
  • [43]
    Mackie, G.A. (1992) Secondary structure of the mRNA for ribosomal protein S20. Implications for cleavage by ribonuclease E. J. Biol. Chem. 267, 10541061.
  • [44]
    Ehretsmann, C.P., Carpousis, A.J., Krisch, H.M. (1992) Specificity of Escherichia coli endoribonuclease RNase E: in vivo and in vitro analysis of mutants in a bacteriophage T4 mRNA processing site. Genes Dev. 149159.
  • [45]
    Cormack, R.S., Mackie, G.A. (1992) Structural requirements for the processing of Escherichia coli 5S ribosomal RNA by RNase E in vitro. J. Mol. Biol. 228, 10781090.
  • [46]
    Mackie, G.A., Genereaux, J.L. (1993) The role of RNA structure in determining RNase E-dependent cleavage sites in the mRNA for ribosomal protein S20 in vitro. J. Mol. Biol. 234, 9981012.
  • [47]
    Fritsch, J., Rothfuchs, R., Rauhut, R., Klug, G. (1995) Identification of an mRNA element promoting rate-limiting cleavage of the polycistronic puf mRNA in Rhodobacter capsulatus by an enzyme similar to RNase E. Mol. Microbiol. 15, 10171029.
  • [48]
    McDowall, K.J., Lin-Chao, S., Cohen, S.N. (1994) A+U content rather than a particular nucleotide order determines the specificity of RNase E cleavage. J. Biol. Chem. 269, 1079010796.
  • [49]
    McDowall, K.J., Kaberdin, V.R., Wu, S.W., Cohen, S.N., Lin-Chao, S. (1995) Site-specific RNase E cleavage of oligonucleotides and inhibition by stem-loops. Nature 374, 287290.
  • [50]
    Arraiano, C.M., Cruz, A.A., Kushner, S.R. (1997) Analysis of the in vivo decay of the Escherichia coli dicistronic pyrF-orfF transcript: evidence for multiple degradation pathways. J. Mol. Biol. 268, 261272.
  • [51]
    McDowall, K.J., Hernandez, R.G., Lin-Chao, S., Cohen, S.N. (1993) The ams-1 and rne-3071 temperature-sensitive mutations in the ams gene are in close proximity to each other and cause substitutions within a domain that resembles a product of the Escherichia coli mre locus. J. Bacteriol. 175, 42454249.
  • [52]
    Wachi, M., Umitsuki, G., Nagai, K. (1997) Functional relationship between Escherichia coli RNase E and the CafA protein. Mol. Gen. Genet. 253, 515519.
  • [53]
    Bycroft, M., Hubbard, T.J., Proctor, M., Freund, S.M., Murzin, A.G. (1997) The solution structure of the S1 RNA binding domain: a member of an ancient nucleic acid-binding fold. Cell 88, 235242.
  • [54]
    Graumann, P.L., Marahiel, M.A. (1998) A superfamily of proteins that contain the cold-shock domain. Trends Biochem. Sci. 23, 286290.
  • [55]
    Carpousis, A.J., Van Houwe, G., Ehretsmann, C., Krisch, H.M. (1994) Copurification of E. coli RNAase E and PNPase: evidence for a specific association between two enzymes important in RNA processing and degradation. Cell 76, 889900.
  • [56]
    Alberts, B. (1998) The cell as a collection of protein machines: preparing the next generation of molecular biologists. Cell 92, 291294.
  • [57]
    Py, B., Causton, H., Mudd, E.A., Higgins, C.F. (1994) A protein complex mediating mRNA degradation in Escherichia coli. Mol. Microbiol. 14, 717729.
  • [58]
    Causton, H., Py, B., McLaren, R.S., Higgins, C.F. (1994) mRNA degradation in Escherichia coli: a novel factor which impedes the exoribonucleolytic activity of PNPase at stem-loop structures. Mol. Microbiol. 14, 731741.
  • [59]
    Mitra, S., Hue, K., Bechhofer, D.H. (1996) In vitro processing activity of Bacillus subtilis polynucleotide phosphorylase. Mol. Microbiol. 19, 329342.
  • [60]
    McLaren, R.S., Newbury, S.F., Dance, G.S., Causton, H.C., Higgins, C.F. (1991) mRNA degradation by processive 3′-5′ exoribonucleases in vitro and the implications for prokaryotic mRNA decay in vivo. J. Mol. Biol. 221, 8195.
  • [61]
    Py, B., Higgins, C.F., Krisch, H.M., Carpousis, A.J. (1996) A DEAD-box RNA helicase in the Escherichia coli RNA degradosome. Nature 381, 169172.
  • [62]
    Vanzo, N.F., Li, Y.S., Py, B., Blum, E., Higgins, C.F., Raynal, L.C., Krisch, H.M., Carpousis, A.J. (1998) Ribonuclease E organizes the protein interactions in the Escherichia coli RNA degradosome. Genes Dev. 12, 27702781.
  • [63]
    Regnier, P., Grunberg-Manago, M., Portier, C. (1987) Nucleotide sequence of the pnp gene of Escherichia coli encoding polynucleotide phosphorylase. Homology of the primary structure of the protein with the RNA-binding domain of ribosomal protein S1. J. Biol. Chem. 262, 6368.
  • [64]
    Higgins, C.F., Causton, H.C., Dance, G.S.C. and Mudd, E.A. (1993) The role of the 3′ end in mRNA stability and decay. In: Control of Messenger RNA Stability (Belasco, J.G. and Brawerman, G., Eds.), pp. 13–52. Academic Press, San Diego, CA.
  • [65]
    Deutscher, M.P., Reuven, N.B. (1991) Enzymatic basis for hydrolytic versus phosphorolytic mRNA degradation in Escherichia coli and Bacillus subtilis. Proc. Natl. Acad. Sci. USA 88, 32773280.
  • [66]
    Zilhao, R., Cairrao, F., Regnier, P., Arraiano, C.M. (1996) PNPase modulates RNase II expression in Escherichia coli: implications for mRNA decay and cell metabolism. Mol. Microbiol. 20, 10331042.
  • [67]
    Donovan, W.P., Kushner, S.R. (1986) Polynucleotide phosphorylase and ribonuclease II are required for cell viability and mRNA turnover in Escherichia coli K-12. Proc. Natl. Acad. Sci. USA 83, 120124.
  • [68]
    Wang, W., Bechhofer, D.H. (1996) Properties of a Bacillus subtilis polynucleotide phosphorylase deletion strain. J. Bacteriol. 178, 23752382.
  • [69]
    Blum, E., Py, B., Carpousis, A.J., Higgins, C.F. (1997) Polyphosphate kinase is a component of the Escherichia coli RNA degradosome. Mol. Microbiol. 26, 387398.
  • [70]
    Kornberg, A. (1995) Inorganic polyphosphate: toward making a forgotten polymer unforgettable. J. Bacteriol. 177, 491496.
  • [71]
    Tzeng, C.M., Kornberg, A. (1998) Polyphosphate kinase is highly conserved in many bacterial pathogens. Mol. Microbiol. 29, 381382.
  • [72]
    Ybarra, J., Horowitz, P.M. (1996) Nucleotides reveal polynucleotide phosphorylase activity from conventionally purified GroEL. J. Biol. Chem. 271, 2506325066.
  • [73]
    Miczak, A., Kaberdin, V.R., Wei, C.L., Lin-Chao, S. (1996) Proteins associated with RNase E in a multicomponent ribonucleolytic complex. Proc. Natl. Acad. Sci. USA 93, 38653869.
  • [74]
    Nicholson, A.W. (1996) Structure, reactivity, and biology of double-stranded RNA. Prog. Nucleic Acid Res. Mol. Biol. 52, 165.
  • [75]
    Luking, A., Stahl, U., Schmidt, U. (1998) The protein family of RNA helicases. Crit. Rev. Biochem. Mol. Biol. 33, 259296.
  • [76]
    Bird, L.E., Subramanya, H.S., Wigley, D.B. (1998) Helicases: a unifying structural theme. Curr. Opin. Struct. Biol. 8, 1418.
  • [77]
    Schmid, S.R., Linder, P. (1992) D-E-A-D protein family of putative RNA helicases. Mol. Microbiol. 6, 283291.
  • [78]
    Anderson, J.S., Parker, R. (1996) RNA turnover: the helicase story unwinds. Curr. Biol. 6, 780782.
  • [79]
    Kido, M., Yamanaka, K., Mitani, T., Niki, H., Ogura, T., Hiraga, S. (1996) RNase E polypeptides lacking a carboxyl-terminal half suppress a mukB mutation in Escherichia coli. J. Bacteriol. 178, 39173925.
  • [80]
    Kaberdin, V.R., Miczak, A., Jakobsen, J.S., Lin-Chao, S., McDowall, K.J., von Gabain, A. (1998) The endoribonucleolytic N-terminal half of Escherichia coli RNase E is evolutionarily conserved in Synechocystis sp. and other bacteria but not the C-terminal half, which is sufficient for degradosome assembly. Proc. Natl. Acad. Sci. USA 95, 1163711642.
  • [81]
    Wang, M., Cohen, S.N. (1994) ard-1: a human gene that reverses the effects of temperature-sensitive and deletion mutations in the Escherichia coli rne gene and encodes an activity producing RNase E-like cleavages. Proc. Natl. Acad. Sci. USA 91, 1059110595.
  • [82]
    Claverie-Martin, F., Wang, M., Cohen, S.N. (1997) ARD-1 cDNA from human cells encodes a site-specific single-strand endoribonuclease that functionally resembles Escherichia coli RNase E. J. Biol. Chem. 272, 1382313828.
  • [83]
    Lisitsky, I., Kotler, A., Schuster, G. (1997) The mechanism of preferential degradation of polyadenylated RNA in the chloroplast. The exoribonuclease 100RNP/polynucleotide phosphorylase displays high binding affinity for poly(A) sequence. J. Biol. Chem. 272, 1764817653.
  • [84]
    Hayes, R., Kudla, J., Schuster, G., Gabay, L., Maliga, P., Gruissem, W. (1996) Chloroplast mRNA 3′-end processing by a high molecular weight protein complex is regulated by nuclear encoded RNA binding proteins. EMBO J. 15, 11321141.
  • [85]
    Min, J., Heuertz, R.M., Zassenhaus, H.P. (1993) Isolation and characterization of an NTP-dependent 3′-exoribonuclease from mitochondria of Saccharomyces cerevisiae. J. Biol. Chem. 268, 73507357.
  • [86]
    Margossian, S.P., Li, H., Zassenhaus, H.P., Butow, R.A. (1996) The DExH box protein Suv3p is a component of a yeast mitochondrial 3′- to-5′ exoribonuclease that suppresses group I intron toxicity. Cell 84, 199209.
  • [87]
    Mitchell, P., Petfalski, E., Shevchenko, A., Mann, M., Tollervey, D. (1997) The exosome: a conserved eukaryotic RNA processing complex containing multiple 3′[RIGHTWARDS ARROW]5′ exoribonucleases. Cell 91, 457466.
  • [88]
    de la Cruz, J., Kressler, D., Tollervey, D., Linder, P. (1998) Dob1p (Mtr4p) is a putative ATP-dependent RNA helicase required for the 3′ end formation of 5.8S rRNA in Saccharomyces cerevisiae. EMBO J. 17, 11281140.
  • [89]
    Anderson, J.S.J., Parker, R.P. (1998) The 3′ to 5′ degradation of yeast mRNAs is a general mechanism for mRNA turnover that requires the SKI2 DEVH box protein and 3′ to 5′ exonucleases of the exosome complex. EMBO J. 17, 14971506.
  • [90]
    Decker, C.J. (1998) The exosome: a versatile RNA processing machine. Curr. Biol. 8, R238240.
  • [91]
    Bessarab, D.A., Kaberdin, V.R., Wei, C.L., Liou, G.G., Lin-Chao, S. (1998) RNA components of Escherichia coli degradosome: evidence for rRNA decay. Proc. Natl. Acad. Sci. USA 95, 31573161.
  • [92]
    Stern, M.J., Ames, G.F., Smith, N.H., Robinson, E.C., Higgins, C.F. (1984) Repetitive extragenic palindromic sequences: a major component of the bacterial genome. Cell 37, 10151026.
  • [93]
    Klug, G., Adams, C.W., Belasco, J., Doerge, B., Cohen, S.N. (1987) Biological consequences of segmental alterations in mRNA stability: effects of deletion of the intercistronic hairpin loop region of the Rhodobacter capsulatus puf operon. EMBO J. 6, 35153520.
  • [94]
    Mackie, G.A. (1987) Posttranscriptional regulation of ribosomal protein S20 and stability of the S20 mRNA species. J. Bacteriol. 169, 26972701.
  • [95]
    Chen, C.Y., Beatty, J.T., Cohen, S.N., Belasco, J.G. (1988) An intercistronic stem-loop structure functions as an mRNA decay terminator necessary but insufficient for puf mRNA stability. Cell 52, 609619.
  • [96]
    Washio, T., Sasayama, J., Tomita, M. (1998) Analysis of complete genomes suggests that many prokaryotes do not rely on hairpin formation in transcription termination. Nucleic Acids Res. 26, 54565463.
  • [97]
    Sarkar, N. (1997) Polyadenylation of mRNA in prokaryotes. Annu. Rev. Biochem. 66, 173197.
  • [98]
    O'Hara, E.B., Chekanova, J.A., Ingle, C.A., Kushner, Z.R., Peters, E., Kushner, S.R. (1995) Polyadenylylation helps regulate mRNA decay in Escherichia coli. Proc. Natl. Acad. Sci. USA 92, 18071811.
  • [99]
    Li, Z., Pandit, S., Deutscher, M.P. (1998) Polyadenylation of stable RNA precursors in vivo. Proc. Natl. Acad. Sci. USA 95, 1215812162.
  • [100]
    Korner, C.G., Wahle, E. (1997) Poly(A) tail shortening by a mammalian poly(A)-specific 3′- exoribonuclease. J. Biol. Chem. 272, 1044810456.
  • [101]
    Cao, G.J., Sarkar, N. (1992) Identification of the gene for an Escherichia coli poly(A) polymerase. Proc. Natl. Acad. Sci. USA 89, 1038010384.
  • [102]
    Kalapos, M.P., Cao, G.J., Kushner, S.R., Sarkar, N. (1994) Identification of a second poly(A) polymerase in Escherichia coli. Biochem. Biophys. Res. Commun. 198, 459465.
  • [103]
    Sarkar, B., Cao, G.J., Sarkar, N. (1997) Identification of two poly(A) polymerases in Bacillus subtilis. Biochem. Mol. Biol. Int. 41, 10451050.
  • [104]
    Coburn, G.A., Mackie, G.A. (1996) Overexpression, purification, and properties of Escherichia coli ribonuclease II. J. Biol. Chem. 271, 10481053.
  • [105]
    Cohen, S.N. (1995) Surprises at the 3′ end of prokaryotic RNA. Cell 80, 829832.
  • [106]
    Cao, G.J., Kalapos, M.P., Sarkar, N. (1997) Polyadenylated mRNA in Escherichia coli: modulation of poly(A) RNA levels by polynucleotide phosphorylase and ribonuclease II. Biochimie 79, 211220.
  • [107]
    Gruissem, W. and Schuster, G. (1993) Control of mRNA degradation in organelles. In: Control of Messenger RNA Stability (Belasco, J.G. and Brawerman, G., Eds.), pp. 329–365. Academic Press, San Diego, CA.
  • [108]
    Li, Q.S., Gupta, J.D., Hunt, A.G. (1998) Polynucleotide phosphorylase is a component of a novel plant poly(A) polymerase. J. Biol. Chem. 273, 1753917543.
  • [109]
    Kudla, J., Hayes, R., Gruissem, W. (1996) Polyadenylation accelerates degradation of chloroplast mRNA. EMBO J. 15, 71377146.
  • [110]
    Hajnsdorf, E., Steier, O., Coscoy, L., Teysset, L., Regnier, P. (1994) Roles of RNase E, RNase II and PNPase in the degradation of the rpsO transcripts of Escherichia coli: stabilizing function of RNase II and evidence for efficient degradation in an ams pnp rnb mutant. EMBO J. 13, 33683377.
  • [111]
    Coburn, G.A., Mackie, G.A. (1998) Reconstitution of the degradation of the mRNA for ribosomal protein S20 with purified enzymes. J. Mol. Biol. 279, 10611074.
  • [112]
    Braun, F., Le Derout, J., Regnier, P. (1998) Ribosomes inhibit an RNase E cleavage which induces the decay of the rpsO mRNA of Escherichia coli. EMBO J. 17, 47904797.
  • [113]
    Coburn, G.A., Mackie, G.A. (1999) Degradation of mRNA in Escherichia coli: an old problem with some new twists. Prog. Nucleic Acid Res. Mol. Biol. 62, 55108.