The RNA Exosomes
RNA exosomes are large multimeric 3′-5′ exo- and endonucleases found in eukaryotes and many archaeal species. They represent the central RNA 3′-end processing factor and are implicated in processing, quality control, and turnover of both coding and noncoding RNAs. RNA exosomes are highly regulated and processive machineries, assembled as large macromolecular cages that channel RNA to the ribonuclease sites. The primordial exosome – found in archaea and related to bacterial and organelle degradosomes – possesses a phosphorolytic active cage that can both degrade and polyadenylate RNA in RNA decay processes. Human and yeast exosomes lost phosphorolytic activities but gained ectopic subunits with hydrolytic activities, while preserving the RNA channeling function.
In all kingdoms of life, the basal physiology, homeostasis, adaptation, and differentiation of cells is regulated at different levels, including transcription, translation, and mRNA decay. A critical quantity is the level of mRNA, which is controlled on one side by transcription by RNA polymerases, and degradation by ribonucleases on the other side.
It is well known that transcription of different RNA species is a highly differential, regulated process and requires the temporal as well as spatial activity of a large number of transcription factors and co-activators as well as other features such as chromatin structure. Although mRNA decay appears much less regulated and fine tuned than transcription, it also has temporal as well as spatial components and it is of course important for cells to be able to either stabilize mRNAs or degrade them in a controlled, regulated manner that is coordinated with the synthesis machinery (Maniatis and Reed 2002). Furthermore, besides the turnover of “normal” RNA, surveillance mechanisms often check the functionality of RNA molecules. For example, mRNAs that lack a stop codon or contain premature stop codons are removed by nucleolytic degradation in processes called nonstop and nonsense-mediated decay, respectively (Belgrader et al. 1994; Frischmeyer et al. 2002; Whitfield et al. 1994).
RNA is degraded by a large variety of nucleases, but the RNA exosome, a large multisubunit complex, emerged over the past decade as the key player in the regulated, controlled RNA processing and degradation in eukaryotes. After its original discovery in yeast and humans, exosome-like complexes have been identified in archaea (Allmang et al. 1999b; Estevez et al. 2001; Evguenieva-Hackenberg et al. 2003; Koonin et al. 2001; Mitchell et al. 1997). It was then established by structural biology results that eukaryotic and archaeal exosomes share many characteristic features and are related to the bacterial degradosome or polynucleotide phosphorylase (PNPase) as well as the phosphorolytic RNase PH (Buttner et al. 2005; Lorentzen et al. 2005; Symmons et al. 2000).
Exosome components in archaea and eukaryotes
PH domain 1
Rrp41, Mtr3, Rrp46
PH domain 2
Rrp42, Rrp43, Rrp45
Rrp4, Rrp40, CsI4
Active hydrolytic RNase
Rrp44(DIS3), Rrp6, DIS3L
Accessory or associated proteins
Nucleus: Rrp47, Nip7, TRAMP complexes, Mpp6, Ndr1-Nab3,Rnt1,AID
Cytoplasm: TUTases, ARE-BP, ZAP, Dom34/Hbs1, Ski7, SKI complex
In archaea, the exosome contains three copies of each of the RNase PH-type subunits aRrp41 and aRrp42. They are assembled in a large macromolecular cage with three self-compartmentalized phosphorolytic active sites. This architectural principle, shielding active sites from unrestricted access, resembles the proteasome for protein degradation. RNA is channeled into the exosome through a pore, a feature that is well suited to provide means for tight regulation, but also to provide highly processive RNA degradation, once RNA is loaded into the exosome.
The eukaryotic exosome contains six different RNase PH-type subunits (Rrp41, Rrp46, Mtr3, Rrp42, Rrp43, Rrp46 in S. cerevisiae) along with S1 and/or KH domain containing subunits Rrp4, Rrp40, and Csl4 (Table 9.1, Fig. 9.1). The RNase-PH type subunits have lost phosphorolytic activity in yeast and humans and the eukaryotic exosome contains the additional hydrolytic subunits Rrp6 and/or Rrp44. Despite this shift in nuclease activities, the overall architecture and assembly of the exosome remains preserved and is clearly important as deduced from the essential nature of the “inactive” RNase PH subunits on yeast viability. In fact, the RNase-PH core still channels RNA Bonneau et al. 2009. From an evolutionary point of view the remarkable switch in activity from phosphorolytic to hydrolytic presumably might be a consequence or prerequisite to the more complexly regulated eukaryotic RNA degradation pathways.
All in all, the emerging architectures reveal exosomes as a fascinating platform for targeted, regulated, and processive degradation of RNA molecules, and this chapter highlights the current understanding in terms of biology and molecular mechanism.
9.2 Biology of Exosomes
9.2.1 Exosomes in Eukaryotes
126.96.36.199 A General Machinery for 3′-5′Degradation
The “exosome” as a multisubunit RNase assembly was first described by David Tollervey and coworkers as large complex with 3′-5′ exonuclease activity in S. cerevisiae (Mitchell et al. 1997). It was then shown that the human equivalent is the PM-Scl complex, a complex previously found to be the component of an antibody-autoantigen system in patients with polymyositis and scleroderma (Allmang et al. 1999b; Reimer et al. 1986). Exosomes were furthermore studied in trypanosomes (Estevez et al. 2001) and plants (Chekanova et al. 2000). The first identified activity of the yeast exosome was an essential role in rRNA biogenesis (Allmang et al. 2000; Mitchell et al. 1997; Zanchin and Goldfarb 1999b), but subsequent studies showed that the exosome is in fact a key player in most, if not all, pathways that require RNA processing. For instance, the exosome together with the Ski (superkiller) complex has been found to be a key enzyme in 3′-5′mRNA decay (Anderson and Parker 1998; Araki et al. 2001; van Hoof et al. 2000b; Wang and Kiledjian 2001). The involvement of the Ski complex provided a first insight that exosomes also need cofactors. The exosome is furthermore important for the biogenesis of the signal recognition particle (Grosshans et al. 2001), the turnover of small nuclear and nucleolar RNAs (van Hoof et al. 2000a), and rapid decay of AU-rich element (ARE) unstable mRNAs (Chen et al. 2001; Haile et al. 2003).
Besides the processing of rRNA and decay or turnover of small RNAs and mRNA, it became rapidly clear that the exosome is also a key player in RNA quality control, for example, for degradation of RNAs that lack a stop codon (nonstop decay) (Frischmeyer et al. 2002; van Hoof et al. 2002) as well as nonsense-mediated decay, which degrades RNA with premature stop codons (Frischmeyer et al. 2002; Lejeune et al. 2003; Mitchell and Tollervey 2003; Takahashi et al. 2003). The exosome furthermore degrades splice-defective RNA in the cytoplasm (Hilleren and Parker 2003) and hypomodified tRNAs, rRNA precursors, and mRNAs with defective polyadenylation in the nucleus (Fang et al. 2005; Kadaba et al. 2004).
The depletion or inactivation of exosome subunits not only identified their role in living cells and which of the RNA processing and degradation pathways requires the exosome, but also uncovered new RNA processing pathways and even new types of RNA molecules that were enriched in the absence of the exosome. This led, for instance, to the discovery of a pathway for nuclear pre-mRNA turnover (Bousquet-Antonelli et al. 2000). The mechanistic basis for this activity is a physical link between the exosome and elongating RNA polymerase in Drosophila (Andrulis et al. 2002).
Furthermore, depletion or inactivation of exosomes in yeast and human cells led to the enrichment and identification of otherwise highly unstable, short RNAs that are normally rapidly degraded by exosomes. In yeast, these RNAs are called cryptic unstable transcripts (CUTs) and arise from transcription of intergenic regions (Wyers et al. 2005). In humans, related RNA molecules result from transcription upstream of active promoters and are called PROMPTs (promoter upstream transcripts) (Preker et al. 2008). Furthermore, RNAs arising from antisense transcription of the noncoding strand in transcriptional gene silencing are degraded by the exosome as well (Camblong et al. 2007). A recent review on these classes or RNA can be found, for example, in Carninci (2010).
Despite its role in degrading coding and noncoding RNAs, the exosome also plays more complex regulatory roles in cells that perhaps go beyond simple RNA degradation. For instance, the exosome plays a role in Neurospora circadian gene expression (Guo et al. 2009), while in immune B cells, the exosome is involved in the generation of antibody diversity via somatic hypermutation and class switch recombination. Here, the exosome recruits the activation-induced cytidine deaminase (AID) to transcription elongation complexes at immunoglobulin loci (Basu et al. 2011).
188.8.131.52 Exosome Isoforms
In eukaryotes, the exosome exists in nuclear and cytoplasmic isoforms. Both isoforms share the nine core subunits as well as the RNase-R like Rrp44/Dis3 subunit. Rrp44/Dis3 carries the main constitutive exonuclease activity of the exosome (Dziembowski et al. 2007). Subsequent studies showed that Rrp44 also possesses endonuclease activity, which was a surprise to the field and created much excitement (Lebreton et al. 2008; Schaeffer et al. 2009; Schneider et al. 2009), because it uncovered a far more complex role of the exosome in degrading RNA substrates and drew exciting parallels to the combination of endo- and exonuclease activities of prokaryotic and organellar degradosomes (for a recent review, see, e.g., (Tomecki and Dziembowski 2010)).
The nuclear isoform of the exosome also contains the Rrp6 protein, an RNase-D like hydrolytic nuclease. The activity of Rrp6 is for instance required to trim rRNA (Allmang et al. 1999a) and small nuclear and nucleolar RNAs (snRNAs and snoRNAs) (van Hoof et al. 2000a). In contrast, Rrp6 is not required for mRNA decay (cytoplasm), but required for nuclear mRNA quality control (Hilleren et al. 2001). In humans, the nuclear exosome associates with PM-Scl100, the homolog of yeast Rrp6. Thus the presence of a nuclear isoform containing an RNase-D activity appears to be evolutionarily preserved.
Besides the specific association of the nuclear exosome with PM-Scl100/Rrp6, human exosomes associate with two different homologs of yeast Dis3p, denoted hDIS3 and hDIS3L (or DIS3-like 1) (Staals et al. 2010; Tomecki et al. 2010). Significantly, hDIS3 and hDIS3L are differently distributed in human cells: hDIS3 is nuclear, while hDIS3L is cytoplasmic. Consistently, hDIS3L is involved in cytoplasmic mRNA decay (Staals et al. 2010). While both proteins are active exonucleases, only hDIS3 has an additional endonuclease activity (Tomecki et al. 2010). In summary, the current data suggest that the human exosome exists with respect to both PM-Scl100 and DIS3 in distinct isoforms in the nucleus and cytoplasm, and hence possesses different endo- and exonuclease properties.
Since the exosome and its isoforms play key roles in numerous RNA processing and degradation pathways, they must be able to target RNA molecules in a manner independent of sequence, secondary structure, and even bound proteins. On the other hand, the activity of the exosome needs to be tightly controlled to avoid unregulated degradation of RNAs. There are several mechanisms that ensure broad but also regulated activity. One mechanism is the structure of the exosome itself. It channels RNA through a narrow pore that restricts entry to single-stranded unfolded RNA molecules (see below). Another aspect is the specific targeting of the exosome, for example, to transcription elongation complexes or stalled ribosomes (via Ski7). Finally, a variety of cofactors help to prepare RNA molecules for degradation by the exosome.
It was not a surprise that Ski2 and Mtr4, two members of the large superfamily 2 helicase/nucleic acid translocases, were found to act as cofactors of the exosome (Anderson and Parker 1998; de la Cruz et al. 1998). Naturally, these enzymes can help unwind structured RNA for degradation and strip proteins from RNA. Ski2 exists in a hetero-trimeric or -tetrameric complex together with Ski3 and Ski8 (Synowsky and Heck 2008; Wang et al. 2005). The Ski complex is found in the cytoplasm and is required for several cytoplasmic activities of the exosome, including mRNA degradation (Anderson and Parker 1998; van Hoof et al. 2000a), degradation of mRNA targeted by the RNA-induced silencing complex (RISC) (Orban and Izaurralde 2005), and degradation of deadenylated mRNA in NMD (Mitchell and Tollervey 2003). Mechanistically, the activity of the Ski complex is not well understood, but the crystal structures of the Mtr4 homolog and more distantly related bacterial Ski-like helicases (which function in DNA replication and repair) as well as of the β-propeller protein Ski8 have been determined (Buttner et al. 2007; Cheng et al. 2004; Jackson et al. 2010; Madrona and Wilson 2004; Weir et al. 2010; Zhang et al. 2008).
Ski2 is closely related to Mtr4, which is found in the nucleus and is involved in processing and degradation of rRNA, snRNAs, snoRNAs, and tRNAs (Allmang et al. 1999a; Cristodero and Clayton 2007; van Hoof et al. 2000a; Wang et al. 2008). Mtr4 has been biochemically characterized as a 3′-5′, ATP-dependent helicase. ATP hydrolysis is stimulated by, for example, tRNA but not by poly(A) RNA (Bernstein et al. 2008), indicating that Mtr4 prefers structured RNA. In this regard, Mtr4 possesses a peculiar arch domain that could help present structured RNA/tRNA to the helicase core (Jackson et al. 2010; Weir et al. 2010).
Mtr4 exists in the nucleus in a complex with two other polypeptides and forms the TRAMP4 (Mtr4-Air2-Trf4) and TRAMP5 (Mtr4-Ari2-Trf5) complexes (Houseley and Tollervey 2006; LaCava et al. 2005). TRAMP oligo/polyadenylates nuclear RNAs that are subsequently degraded by the exosome in nuclear pathways. These results established that adenylation is a common mechanism for RNA stability and turnover through exosome-mediated degradation in both the nucleus and the cytoplasm.
9.2.2 Exosomes in Archaea
While the exosome is a firmly established key player in the 3′ degradation of RNA molecules in eukaryotes, the biology of archaeal exosomes is less clear. In fact, not all archaea contain homologs of exosome subunits, so other activities such as RNase-R (Portnoy and Schuster 2006) can perhaps compensate for a lack of exosome-like complexes. The archaeal exosome was first predicted by bioinformatic analysis of archaeal genomes (Koonin et al. 2001) and shortly after experimentally identified in Sulfolobus solfataricus (Evguenieva-Hackenberg et al. 2003) and Methanococcoides burtonii (Goodchild et al. 2004).
The main biochemical difference between archaeal and eukaryotic exosomes is the presence of phosphorolytic activity in the former, while the latter appears to lack this activity and instead adopted additional subunits with hydrolytic activity. Initial insights into the role of exosomes in archaea came from the analysis of poly(A) tails on archaeal RNA. It was found that RNAs from a halophilic archaeon, which does not encode homologs for exosome subunits in its genome, do not show RNA poly(A) tails. It has been suggested that in the absence of the exosome, RNA degradation could be performed by an RNase R homolog (Portnoy and Schuster 2006). Interestingly, RNase-R is related to Rrp44, a hydrolytic subunit of the eukaryotic exosome. While poly(A)-rich tails are not found in halophilic archaea, heteropolymeric, A-rich tails were found on RNA from S. solfataricus, which possesses an exosome (Portnoy et al. 2005). The presence of these tails in Sulfolobus and methanogens correlates with the presence on an exosome (Portnoy and Schuster 2006). The added tails are poly(A)-rich but heteropolymeric, suggesting that the composition perhaps reflects the nucleoside diphosphate concentrations in the cell. In other words, the exosome is not a specific poly(A) polymerase, but non-discriminatorily adds nucleoside diphosphates in vivo similar to its biochemical properties in vitro. A role in both adding tails and degrading RNA is reminiscent of chloroplast PNPase, which also adds tails to RNA, followed by endonucleolytic cleavage and exonucleolytic degradation (Yehudai-Resheff et al. 2003). It is therefore possible that the archaeal exosome, like PNPases, acts in vivo both by adding tails and degradation of RNA (Slomovic et al. 2008), but it remains to be shown whether these activities are regulated. Recent single molecule studies showed that exosomes can easily switch between the two types of activities and there appears to be a memory effect for polymerization as well as degradation modes under conditions when both reactions are thermodynamically in equilibrium, that is, the exosome stochastically switches between degradation and polymerization but stays in one mode for a while (Lee et al. 2010).
It is yet unclear how archaeal exosomes recognize their specific RNA targets and how the activity of the exosome is regulated in the archaeal cell. The core exosome is found associated with other polypeptides such as the archaeal DnaG homolog (Evguenieva-Hackenberg et al. 2003; Walter et al. 2006); however, the function of this protein in the context of the exosome is still unclear. DnaG is the primase in bacterial DNA replication. Archaea contain in addition to an archaeo-eukaryotic type of primase, which acts in DNA replication, a DnaG-like enzyme with a Toprim fold, and this protein is found in the exosome complex (Evguenieva-Hackenberg et al. 2003; Iyer et al. 2005; Walter et al. 2006). However, Sulfolobous DnaG also has a primase type activity in vitro, and the biochemical connection to exosomes is unclear (Zuo et al. 2010). Other interactors include archaeal Nip7, which inhibits degradation of poly-(A) and poly-(AU) RNA in vitro. Eukaryotic Nip7 also interacts with the eukaryotic exosome and is required for pre-rRNA processing, suggesting an evolutionary conserved targeting or regulation (Luz et al. 2010; Zanchin and Goldfarb 1999a).
Recent exciting results show that the exosome is localized to membranes in archaea (Roppelt et al. 2010), a feature it shares with the degradosome (Khemici et al. 2008). In the case of degradosomes, membrane association is mediated by an amphipathic helix in the RNase-E subunit (Khemici et al. 2008). A potential candidate for membrane association is the DnaG subunit of the archaeal exosome (Evguenieva-Hackenberg et al. 2011). The functional importance of this subcellular localization remains to be established. In principal, it could be an important mechanism to target specific RNA molecules or a consequence of RNA localization to bacterial membranes (Kawamoto et al. 2005; Nevo-Dinur et al. 2011).
In summary, the physiological isoforms of exosomes in archaea and their differential role in RNA processing and degradation require further study, although considerable progress in this respect has been made. For instance, archaeal exosomes can have different RNA binding (aRrp4 and aCsl4) caps, which confer different biochemical properties on the exosome. The role of different cap types in different archaeal RNA degradation or processing pathways remains to be studied. However, the cap proteins could be differentially expressed: While the aRrp4 gene is in the same operon as genes for aRrp41 and aRrp42, the gene for aCsl4 is located elsewhere in the genomes, offering the possibility of production of different exosome isoforms also in archaea by gene regulation.
9.3 Exosome Architecture
9.3.1 The Nine Subunit Core Exosome
The pioneering studies on the archaeal exosome began with the structure of the 6-subunit RNase-PH-like ring that revealed the presence and location of three active sites and established the structural relation with PNPase (Lorentzen et al. 2005). Shortly after, structures of the full 9-subunit exosome in two different isoforms (with aRrp4- and aCsl4-type caps) (Buttner et al. 2005), and the 6-subunit and 9-subunit exosomes in complex with RNA were reported (Lorentzen and Conti 2005; Lorentzen et al. 2007). These studies showed that exosomes are “self-compartmentalized” RNases with an entry and exit pore for RNA and degradation products, respectively. Subsequent studies on different archaeal exosomes, with and without RNA, showed features of degradative processivity, conformational flexibility, and RNase as well as polymerization mechanisms (Hartung et al. 2010; Lu et al. 2010; Navarro et al. 2008).
The (Rrp41:Rrp42)3 ring is bound in total by three copies of Csl4 and/or Rrp4 domains. Csl4 and Rrp4 both contain an S1 domain and the three copies of the S1 domains frame the “entry” pore for RNA into the active site. Although direct binding of the RNA to the S1 domains has not yet been observed, it is likely that these domains help to recruit and channel RNA into the central processing chamber. In fact, RNA could be visualized in the narrow neck between the S1 domains and the processing chamber using RNA with a single strand 3′ tail and a stable hairpin structure at the 5′ end (Lorentzen et al. 2007) and mutations in the neck also interfere with exosome activity (Buttner et al. 2005; Oddone et al. 2007). In summary, the narrow neck is the likely entry point for RNA and presumably functions to limit entry into the RNA processing chamber of only unfolded, protein-stripped RNAs.
Besides the more central S1 domains that may guide RNA through the entry pore, both Csl4 and Rrp4 contain additional peripheral domains. Csl4 contains a ZR (zinc ribbon) domain and Rrp4 a KH domain. KH domains are well-known RNA interaction domains, suggesting that these domains are important elements of RNA targeting. The function of the ZR is less clear and it could mediate interactions either with specific RNA substrates or alternatively with other proteins.
The eukaryotic RNase-PH-like ring is composed of three “aRrp41” homologs (Rrp41, Rrp46, Mtr3) and three “aRrp42” homologs (Rrp42, Rrp43, Rrp45). This core ring binds to Rrp4, Rrp40, and Csl4 and the positioning of S1, KH, and ZR is very similar to the position of equivalent domains on the archaeal exosome. The conservation of S1, KH, and ZR domains on the exosome surface between archaea and eukaryotes is remarkable and suggests that basal mechanism of RNA recognition and recruitment are preserved.
9.3.2 The Eukaryotic Exosome
The 9-subunit eukaryotic core exosome associates with the hydrolytic nucleases Rrp44 and Rrp6, a feature that is very distinct from archaeal exosomes. The interaction of Rrp44 subunits with the core has been visualized by X-ray crystallography and by electron microscopy and additionally analyzed by mass spectrometry (Bonneau et al. 2009; Cristodero et al. 2008; Malet et al. 2010; Taverner et al. 2008). These results showed that Rrp44 binds to several RNase-PH subunits on the opposite site of the surface that is covered by the S1 domains (Figs. 9.1 and 9.2). This led to the idea that RNA exiting the core exosome is directly fed into the exonuclease active site of Rrp44. The PIN domain harboring the endonucleolytic site is located slightly to the outside of the exit pore of the core ring. It remains to be shown how this endoribunuclease acts in the context of the exosome. In contrast to Rrp44, the location of Rrp6 is less well understood. Electron microscopy on trypanosome exosomes suggests it binds to the outside and not near the exit pore (Cristodero et al. 2008). Thus, the mutual orientation of exo- and endoribonuclease active sites on the eukaryotic exosomes needs to be further addressed.
9.4.1 Archaeal Exosomes
Several crystal structures of archaeal exosomes have been solved in complex with different RNA molecules (Hartung and Hopfner 2009; Lorentzen and Conti 2005; Lorentzen et al. 2007; Navarro et al. 2008). These structures showed that the active site binds the 3′-end of RNA through interactions with its backbone and base stacking. This sequence-unspecific recognition of the RNA substrate explains why the exosome can degrade mRNAs with any sequence and leaves the role of substrate selection to the additional RNA-binding domains and/or accessory proteins. Interactions between the 2′-hydroxy groups of two riboses and side chains of the exosome ensure RNA over DNA discrimination. A tyrosine residue in the active site of the exosome is stacked between the fourth and the fifth base of the RNA, thus positioning the RNA at the active site and enabling efficient cleavage (Hartung et al. 2010).
Although the main active side residues are located in the Rrp41 protein, some residues (like the above-mentioned tyrosine) from the Rrp42 protein are important for RNA binding and influence degradation efficiency and processivity. All crystal structures of exosomes with RNA have one important aspect in common: even in the presence of longer RNA substrates, not more than 4–6 bases can be observed directly in the active sites, with clear electron density for an additional nucleotide in the neck region of the exosome ring. The conformation of all nucleotides in between is too flexible to be determined by crystallography. This observation raises the question whether (and if so, how) the three active sites within the processing chamber interact: does the 3′-end of the RNA molecule, after entering the processing chamber, bind to a single active site for degradation of the whole molecule? Or could RNA, while being fixed at the narrow neck, stochastically or in an ordered manner switch between the three active sites? The question still remains unanswered.
The molecular mechanism of RNA polymerization and degradation has been extensively studied for Escherichia coli PNPase. Here, the detection of heteropolymeric tails after inactivation of the poly(A)-polymerase gene led to the identification of PNPase as the responsible enzyme (Mohanty and Kushner 2000). In accordance with these experiments it was also shown for other organisms, including plant chloroplasts, that PNPase in bacteria/eukaryotic organelles and the exosome in archaea are the only enzymes that produce heteropolymeric tails (Rott et al. 2003; Sohlberg et al. 2003; Yehudai-Resheff et al. 2003).
9.4.2 RNA Recognition, S1, ZnR, and KH Domains
Although the architecture of the processing chamber ensures that no RNA molecule with secondary structure can be degraded, there must be an additional regulatory mechanism directing some RNA molecules to the exosome for degradation while protecting others. Obvious candidates for some targeting are the Csl4 and Rrp4-type subunits. It was indeed shown for the archaeal exosome that Rrp4 and Csl4 greatly increase the RNA degradation activity, whereas they are less important for efficient polymerization (Evguenieva-Hackenberg et al. 2008).
Accessory factors like helicases can direct substrates to the exosome, but this is not the only mode of substrate selection. The genetic organization of the four archaeal exosome genes already indicated a possible regulatory function of the cap proteins: the genes rrp41, rrp42, and rrp4 are located in one operon whereas csl4 is under the control of a different promoter. Different promoters allow for specific expression of the Csl4 cap protein independent from the rest of the exosome. The composition of exosome caps can be varied based on expression levels, ranging from three Rrp4 proteins, via mixed complexes to three Csl4 proteins. Archaeal exosomes can be isolated as either Rrp4-only or Csl4-only complexes. Availability of the two different exosomes allows for the determination of different substrate specificity of the cap proteins: detailed kinetic studies on the degradation of poly(A) RNA showed clear differences between the mode and speed of degradation dependant on the cap proteins (Hartung et al. 2010; Niederberger et al. 2011). Additionally, the preference for various RNA substrates was examined and the exosome with Rrp4 was shown to degrade poly(A) RNA much better than heteropolymeric sequences, whereas exosomes with Csl4 prefer RNA with A-poor sequences. This hints to a possible role of Rrp4 in degradation of RNAs that are targeted by eukaryotic poly(A) polymerases such as Trf4/5.
184.108.40.206 Channeling and Processivity
Structural studies on different archaeal exosomes in complex with RNA led to the view that RNA binds to the RNA-binding domains of the cap proteins and the 3′-end is subsequently threaded through the narrow pore into the processing chamber and toward the active site (Fig. 9.2); this mechanism of binding was therefore named “channeling.” Mutagenesis experiments with bacterial PNPase suggest a similar channeling mechanism for RNA (Shi et al. 2008). The ring architecture of the exosome and the channeling mechanism seem to have advantages and disadvantages. Through arginine side chains in the narrow pore region, RNA is fixed to the exosome and processivity is ensured. This mechanism is also true for the eukaryotic exosome that uses a completely different protein as the catalytic component, but still uses the processing chamber as tong to hold on to the RNA (Bonneau et al. 2009). Still this architecture seems to hinder substrate binding: the 3′-end of a long RNA molecule has to thread through this narrow hole to reach into the processing chamber. This mechanism raises the questions:how the exosome can be such an efficient machine and how does the RNA find the correct path?
There are two aspects to these questions. First, time-dependent experiments show that the rate for binding of the 3′-end of RNA to the active site of the exosome is considerably lower than the degradation rate (Hartung et al. 2010). This means that binding is slow, but the cleavage rate is dramatically high, so that altogether the exosome is still a very efficient machine. Second, comparison of various exosome complexes (with or without different caps; crosslinked or forced dimeric Rrp41/Rrp42 complexes) led to the conclusion that the 6- and 9-subunit assembly of the exosome is variable, the “ring is breathing.” Without RNA, the three arginines in the neck region are not fixed, the diameter of ring increases, and a wider opening facilitates RNA threading. Only after the negatively charged RNA is bound does the neck tighten, and the exosome remains bound to its substrate until it is degraded.
9.4.3 Eukaryotic Exosomes
220.127.116.11 Loss of Phosphorolytic Active Sites
Based on structural, biochemical, and genetic work it is clear that the 9-subunit eukaryotic core exosome contains no enzymatically active center in human cells and yeast (Dziembowski et al. 2007; Liu et al. 2006) and that the two main proteins that are present as part of the eukaryotic complex must be responsible for the enzymatic activity in eukaryotes: Rrp44/DIS3 and Rrp6. In contrast to yeast and human exosomes, the Rrp41 protein from plant exosomes possibly retained its catalytic competence because phosphorolytic RNase activity could be detected for purified Rrp41 in vitro (Chekanova et al. 2000). Perhaps plant exosomes are evolutionary intermediates between the archaeal homolog and the yeast and human exosomes. Further research on plant exosomes confirmed this observation: First, Rrp44 could not be found as part of the exosome complex and second, Csl4 (in contrast to yeast) is not essential for viability and deletion of the csl4 gene affected only a very small subset of target RNAs (Chekanova et al. 2007).
18.104.22.168 Two Different Hydrolytic Activities in Rrp44
The yeast complex and its corresponding enzymatically active subunit Rrp44 is from a biochemical standpoint the most intensively characterized eukaryotic exosome. Rrp44/Dis3 is homologous to the bacterial RNase II, a hydrolytic RNA-degrading enzyme (Fig. 9.3), and biochemical characterization of the isolated Rrp44 showed similar activities for the isolated and the exosome-bound enzyme (Dziembowski et al. 2007). This however is only true for single-stranded, unfolded RNA molecules. Rrp44 alone can unwind and degrade substrates with secondary structures, but this activity is strongly inhibited by the presence of the 9-subunit exosome complex (Bonneau et al. 2009; Lorentzen et al. 2008). RNase protection assays showed that isolated Rrp44 binds to a stretch of 9–12 nucleotides, whereas the 10-subunit exosome complex including Rrp44 binds 31–33 nucleotides. This is equivalent to the distance from the active site of Rrp44, through its substrate channel, the processing chamber of the exosome and to the neck region between the cap proteins. Only shortly after those discoveries, a new feature of Rrp44 was detected that was both surprising and intriguing: after the bacterial and the archaeal exosomes as well as Rrp44 were described as exonucleases, an additional endoribonucleolytic activity, mediated by a second domain of Rrp44, the PIN domain, was identified (Lebreton et al. 2008). Only the inactivation of both active sites within Rrp44 results in a synthetic growth phenotype, indicating that both activities have a physiological role. The cleavage of certain natural substrates could be assigned to this endonuclease activity (Schneider et al. 2009). Structural and mutagenesis studies helped to indentify the binding site of Rrp44 on the exosome (Bonneau et al. 2009): matching biochemical data already available, the RNA entry site of Rrp44 co-localizes with the RNA exit site of the processing chamber (Fig. 9.2a).
22.214.171.124 Hydrolytic Activities of Rrp6
The central part of Rrp6 is homologous to bacterial RNase D, which consists of an exonuclease and one or more helicase and RNase D C-terminal (HRDC) domains. Those RNases are characterized by at least four conserved acidic residues (DEDD) in the active site and were shown to require two divalent metal ions for the activation of a water molecule that attacks the last phosphodiester bond, resulting in 3′-to-5′ exonucleolytic cleavage (Steitz and Steitz 1993). The eukaryotic RNaseD enzymes evolved from the bacterial homologs and have almost twice the size. The additional N-terminal domain of yeast Rrp6 covers the catalytic core and creates an interaction surface for the RNA-binding protein Rrp47 (Stead et al. 2007). The atomic structure of yeast Rrp6 shows the position of the two metal ions (Fig. 9.3), the specific ribonucleotide recognition mechanism and parts of the eukaryotic specific N-terminal extension (Midtgaard et al. 2006). The activating effect of the TRAMP complex on exosome activity in the nucleus could be ascribed to the hydrolytic activity from Rrp6, and is independent of Rrp44. Since the activating effect is not ATP dependent, neither the poly(A) polymerase activity of Trf4 nor the helicase activity of Mtr4 seem to be responsible (Callahan and Butler 2010).
But why does the nuclear exosome need two different hydrolytic active sites? It can be speculated that in addition to specificity for certain substrates, accessibility might be a reason. From structural studies on Rrp44 we know that the active site is deeply buried in the protein and the 3′-tail of a structured RNA molecule cannot be degraded completely. No structural information on full-length Rrp6 is available; however, the solved structure of an Rrp6 fragment suggests a more surface-exposed active site, and thus Rrp6 could trim RNA overhangs for maturation in a way that Rrp44 is not capable of.
126.96.36.199 RNA Recognition and Channeling
Purification of stable human exosomes with fewer than 9 subunits has not been successful to date, and biochemical characterization of the cap proteins is therefore challenging (Liu et al. 2006). Recently, yeast exosomes could be purified as 7-subunit complexes containing only one of the cap proteins, and as 8-subunit complexes with two different caps (Malet et al. 2010). This is a very important step toward a better understanding of substrate selection and specificity of those proteins. RNA protection assays with all exosome variants determined a critical role for Rrp4 in the formation of an exosome-RNA complex and identified Rrp40 as the only cap protein that can alone form a stable RNase PH-like exosome. It seems that Rrp40 binds to the RNase PH-like ring in multiple copies, producing a structure more similar to the archaeal exosome.
The channeling of RNA through the eukaryotic exosome could also help to select for RNAs with unfolded 3′-ends over folded RNAs or RNPs, explaining how poly/oligo(A) tails are an important structural feature for unstable RNAs. The RNA is believed to bind to the cap proteins Rrp4/Rrp40/Csl4, followed by a threading through the central channel of the RNase PH-like ring to the active site of Rrp44 (Fig. 9.2) (Malet et al. 2010). This mechanism not only maintains the roles of cap proteins and exosome ring-structure, but also protects the Rrp44 active site from solvent and regulates its enzymatic activity. In summary, RNA channeling seems to be an evolutionarily conserved mechanism for RNA degrading machines in all kingdoms of life.
9.5 Concluding Remarks
9.5.1 Evolutionary Considerations
The conservation of exosomes in eukaryotes and many archaea, and the relationship to degradosomes in bacteria and organelles establish exosome-like complexes as quite ancient RNA degradation and polymerization machineries. Activities associated with eukaryotic exosomes and prokaryotic degradosomes show related overall mechanisms and pathways in RNA decay. These include the addition of oligo/poly-(A) tails to the 3′ end of RNA molecules for stabilization (long tails in eukaryotes) or degradation (short tails in eukaryotes and tails in prokaryotes), endonucleolytic cleavage and processive 3′ degradation. Nevertheless, it is remarkable that the exosome lost phosphorolytic activity in more complex organisms. In terms of cell physiology, a phosphorolytic enzyme may be viewed as a mediator between a free NDP pool and RNA polymers, and phosphorolytic activity is probably quite efficient in terms of energy conservation. Thus, it could be of substantial benefit to degrade RNA phosphorolytically, especially for fast growing organisms such as bacterial, or organisms in scant or extreme environments such as archaea. On the other hand, phosphorolytic degradation and polymerization is readily reversible and the cell might have little means to specifically control degradation versus polymerization although the reversible reaction could also be a simple means of regulation of mRNA stability based on the physiological state of the cell.
It is therefore quite intuitive that in eukaryotes the phosphorolytic degradation and polymerization activities of the core exosome have been inactivated and mechanistically separated into two irreversible, highly specialized, and regulatable activities in the form of hydrolytic nucleases and poly(A) polymerases.
9.5.2 Future Questions
Despite the tremendous progress in the field over the past decade with respect to both pathways and structure, several important aspects still need to be addressed. From a structural and biochemical point of view, it is still unclear how substrates are chosen for degradation, what are the precise functions of Rrp4, Rrp40 and Csl4 caps in RNA recognition or interaction with other protein partners? How is the exosome targeted to the sites of RNA degradation, for instance in surveillance of nonfunctional RNA? How does it structurally or functionally interact with the Ski and Tramp complexes? What is the function of associated archaeal subunits such as DnaG? Is there also endonuclease activity associated with the archaeal exosome, and what are the functional equivalents of Ski2 and Mtr4 helicases in archaea? How are endo- and various exonuclease activities mechanistically coordinated? Finally, are there exosomes with both phosphorolytic and nucleolytic activities? There are of course many aspects to discover in the complex role of exosomes in eukaryotic and archaeal cell biology and we look forward to the next decade of exciting research on exosomes.
Work in KPHs laboratory on exosomes is supported by a grant from the German Research Council (DFG HO2489/3).
- Chekanova JA, Shaw RJ, Wills MA, Belostotsky DA (2000) Poly(A) tail-dependent exonuclease AtRrp41p from Arabidopsis thaliana rescues 5.8 S rRNA processing and mRNA decay defects of the yeast ski6 mutant and is found in an exosome-sized complex in plant and yeast cells. J Biol Chem 275:33158–33166PubMedGoogle Scholar
- Evguenieva-Hackenberg E, Roppelt V, Lassek C, Klug G (2011) Subcellular localization of RNA degrading proteins and protein complexes in prokaryotes. RNA Biol 8:49–54Google Scholar
- Lee G, Hartung S, Hopfner KP, Ha T (2010) Reversible and Controllable Nanolocomotion of an RNA-Processing Machinery. Nano LettGoogle Scholar
- Niederberger T, Hartung S, Hopfner KP, Tresch A (2011) Processive RNA decay by the exosome: merits of a quantitative Bayesian sampling approach. RNA Biol 8:55–60Google Scholar