Cytoplasmic Ribonucleoprotein Foci in Eukaryotes: Hotspots of Bio(chemical)Diversity

The life of an mRNA from transcription to degradation offers multiple control check points that regulate gene expression. Transcription, splicing, and translation have been widely studied for many years; however, in recent years, new layers of posttranscriptional and posttranslational control have been uncovered. They involve the regulation of the metabolism of mRNA in cytoplasmic foci. They are collections of ribonucleoprotein complexes that, in most cases, remain still uncharacterized, except the processing bodies (PBs) and stress granules (SGs), which have been studied (and reviewed) in detail. A challenging prospective is to know how many different classes of foci exist, which functions they support, how are they formed, and how do they relate one to each other. Here, we present an update of the component of the different granules, a possible function, and hypothesis on their in vivo dynamics related to translational control.


Introduction
In recent years, several cytoplasmic foci/granules that contain proteins and RNA have been described. Two of them have been studied in more detail as they are related to mRNA silencing: stress granules (SG) and processing bodies (PB). SG are repressed mRNPs transiently induced in response to cellular stress. They range from 0,5 to 5 μm [1]. PB are discrete RNP cytoplasmic foci of 0,1-2 μm where the machinery of RNA interference, degradation and storage locates. In PB the mRNAs are forming mRNP complexes either repressing translation, in degradation complexes or stored for further use [2,3]. SG and PB have been shown to share a growing number of proteins that are added in a dayto-day basis to the list of their components. SG, PB and other cytoplasmic foci are highly dynamic structures, although PB are quite stable over the time [4]; see also Supplementary Movie 1 available online at doi: 10.1155/2012/504292. They are in a dynamic steady state with other mRNPs, such as polysomes in response to the translational state of the cell [5]. Although we do not intend to extensively review SG and PB, which have been matter of fine reviews in the last years [6][7][8][9][10], we will overview their functions before we address neglected issues and hypothesis.

Stress Granules
Translation initiation is the key regulatory step of translational control. Therefore, it is the most sensitive step to changes in the cellular environment, including stress. A key step in translation initiation inhibition is the phosphorylation of eIF2α, which results in an increase on the affinity of eIF2-GDP for eIF2B, sequestering this factor to prevent new round of translational initiation [11]. During this process, translation is inhibited and polysomes become released from the mRNA leading to the accumulation of inactive mRNPs in SG. The SG are in equilibrium with active polysomes. Protein elongation inhibitors, such as cycloheximide, prevent the assembly of SG by blocking the polysomes in an inactive state, while protein initiation inhibitors promote the formation of SG [8]. Table 1 shows the components of SG described up to now. They can be classified in three main groups as follows. (1) Core components: stalled initiation complexes (polyadenylated mRNAs and translation factors eIF4E, eIF4A, eIF4G, eIF3, eIF2, PABP, and proteins of the small ribosome subunit).
The key concept regarding SG is that they are responsible of protecting the mRNA during cell stress, altering the composition of the mRNPs in a reversible manner. As soon as the cell recovers, the mRNPs regain their translational capacity.

Processing Bodies
These structures have been described many times since 1997, when Bashkirov et al. observed that the exonuclease Xrn1 is located in small granular structures in the cytoplasm of mammalian cells and call them "Xrn1 foci" [17]. Later on, the decapping enzyme Dcp2 was also described to occur in cytoplasmic foci [18]. Contemporary, Eystathioy et al. have described that a protein associated to neuropathy named GW182 occurs in cytoplasmic speckles called GW bodies [19,20]. Other RNA-related protein, the eIF4E-transporter, was also localized in discrete cytoplasmic foci [16,21,22]. Short after, a seminal work of Sheth and Parker established the functional bases of the now called PB that resulted in the same structures described many times before [23]. They Suppressor of deccaping Sc Germin 5 Part of small nuclear RNPs Human Dcs2 Stress-induced regulator Sc APOBEC3G, APOBEC3F Antiviral activity Human demonstrated that PB contains enzymes involved in the degradation of the mRNA. Later one further studies showed that they are also related to miRNA metabolism and can store mRNAs to bring them back to polysomes (reviewed in [3,24]. They include, different than SG, neither ribosomal proteins nor translation factors, except eIF4E. They do not present either the exosome components [25]. eIF4G and PABP were found in yeast PB, although at low level and in stress conditions resulting on glucose deprivation [26]. Proteins and mRNA can reversely go in and out of PB [25]. The relationship of PB and polysomes is demonstrated by the blocking of PB formation by cycloheximide. A summary of the components in different organisms is shown in Table 2. The occurrence of such large and diverse set of proteins (and the list continuously grows up) suggests that PBs are involved in a plethora of posttranslational processes regulating gene expression, such as mRNA degradation and silencing. mRNA degradation starts with the shortening of the poly-A tailthe deadenylation. In eukaryotes, there are several complexes involved in the process: PARN2-PARN3 initiates the process, which continues with the action of the CAF1-CCR4-NOT complex. Later on, mRNA degradation continues by nucleolytic cleavage on both ends. 3 → 5 degradation is catalyzed by the exosome and the SKI complex, while 5 → 3 degradation requires previous decapping by DCP2 and the coactivator DCP1 and the action of the exonuclease XRN1. All these enzyme localize in PB. There are several evidences indicating that mRNA degradation occurs in PB.
(i) The assembly of PB depends on mRNA, as RNase treatment of the cells induces the disappearing of PB [27,28].
(ii) Inhibition or removal of the deadenylase Ccr4 reduces the number and size of PB, while the removal of the downstream-acting enzymes Xrn1 and Dcp1 does not affect the stability of PB [21].
Therefore, one can conclude that mRNA degradation occurs in PB and depends on the existence of degradation enzymes and mRNA degradation intermediates [21,23,25,29]. Many of the PB components are not restricted to the foci and also are present in the soluble cytoplasm and nuclei, suggesting that the different processes might start before the mRNAs entry into PB. PBs are also related to mRNA quality control mechanisms, such as nonsense-mediated decay (NMD). The detection of premature termination in the cells by spotting an mRNA with an abnormal stop codon is mediated by a surveillance complex composed by UPF1, UPF2 y UPF3, additional proteins, namely, SMG1, and SMG5-7 [30][31][32]. As soon as the surveillance complex is assembled, the degradation enzymes (Dcp1, Xrn1) are recruited to the mRNA in PB. Although the degrading enzymes are located in PB, the mechanism of recruitment is unknown. In silencing, there are two types of small mRNAs that regulate posttranscriptional gene expression: siRNAs and miRNAs. Despite the different mechanism of silencing, in both cases participate the protein Argonaute (Ago) and the RISC (RNA-induced silencing complex). In the case of siRNAs, Ago produces an endonucleolytic cleavage of the mRNA to promote degradation by the 3 → 5 and 5 → 3 decay machinery in PB. In the case of the miRNAs, they

The Cycle of an mRNA in the Cell: SG, PB, Polysomes, and the Unknown Intermediates
From the previous analysis, one can establish many unsolved aspects on cytoplasmic foci function. One of them is the dynamic of the mRNP remodeling. The current model suggests an active movement of mRNPs from and to polysomes and from and to SG and PB [33]. However, how does it happen and the factors involved are not known. Translationally active mRNAs can interact, in response to errors in translational initiation or to specific recruitment of regulatory proteins, with translational repressors such as Dhh1, Pat 1, Lsm1-7, eIF4E-T. Those factors would promote the replacement of the translational machinery from the mRNA, promote the cap removal and determine degradation [33] or the accumulation of silenced mRNA in PB. Within PB, mRNPs could undergo further remodeling and define a path to follow, including their return to polysomes. In addition, PBs have been shown to interact and exchange components or their own nature with SG (reviewed in [6,7]) in a process that may result in mRNPs intermediates of unknown nature. Evidence for the diversity of cytoplasmic foci and their components results from immunocytochemistry and colocalization studies. A common factor present in most cytoplasmic mRNPs is the cap-binding protein eIF4E. eIF4E occurs in active polysomes as a translation initiation factor, in SG as part of the stalled initiation complex, and in PB as the only translation factor present there in multicellular eukaryotes. We observed in Drosophila S2 cells that eIF4E colocalizes with different pairs of markers, either for PB (GW182, Lsm1, Me31B-an ortholog of the helicase rck/p54) or SG (TIA-1) and that the colocalization does not occur in all foci in the same way (PVF, CL, and RRP, unpublished data and Figure 1). In some cases, the foci contain one, the other, or both components. In the foci that show colocalization of both factors, the relative amount of each component may vary from foci to foci, as judged by confocal microscopy quantification of the colocalized factors (PVF, CL, RRP, unpublished observation). This implies that there are a diversity of granules. An appealing hypothesis is that eIF4E is a common link among different mRNPs, playing different roles depending on their interactors. One plausible function could be that the accumulation of mRNPs in eIF4E-containing foci is a way to regulate the rate of translation in different physiological states (cell cycle phases, developmental stages, circadian rhythms). Moreover, it has been reported that, in mammalian cells, eIF4E interacts in PB with at least two factors, rck/p54 and eIF4E-T [21]. These are simultaneous interactions within the PB and imply that both proteins could contact different domains of the same eIF4E molecule or that they would represent different populations of mRNPs or different functions within the same PB. In either cases, the complexity of the interactions in vivo is more diverse than it has been expected. A model for the remodeling of active mRNPs to silence and degradation based on Andrei et al. [21] is depicted in Figure 2. This might require several intermediate states that can be the maturation steps of a mRNA in the way of a PB or within a PB. This 6 Comparative and Functional Genomics would correlate with the large diversity of components and interactions within a cytoplasmic foci and the diversity of the foci within a cell. The understanding of the dynamics of mRNP is far from clear and unpredictable paths remain to be discovered. They will need further research and more sophisticated methods for in vivo studies.