Elsevier

Methods in Enzymology

Volume 448, 2008, Pages 521-552
Methods in Enzymology

Chapter 26 Real‐Time and Quantitative Imaging of Mammalian Stress Granules and Processing Bodies

https://doi.org/10.1016/S0076-6879(08)02626-8Get rights and content

Abstract

Nuclear mRNA domains such as nucleoli, speckles, Cajal bodies, and gems demonstrate that RNA function and morphology are inextricably linked; granular mRNA structures are self‐generated in tandem with metabolic activity. Similarly, cytoplasmic compartmentalization of mRNA into mRNP structures such as stress granules (SGs) and processing bodies (PBs) reiterate the link between function and structure; the assembly of SGs and PBs requires mRNA released from disassembling polysomes on translational arrest. SGs contain mRNA still associated with some of the translational machinery, specifically 40S subunits and a subset of translation initiation factors including eIF3, eIF4F, eIF4B, and PABP. PBs also contain mRNA and eIF4E but lack other preinitation factors and contain instead a number of proteins associated with mRNA decay such as DCP1a, DCP2, hedls/GE‐1, p54/RCK. Many other proteins (e.g., argonaute, FAST, RAP‐55, TTP) and microRNAs are present in both SGs and PBs, sometimes shepherding specific mRNA transcripts between the translation and decay machineries. Recently, we described markers and methods to visualize SGs and PBs in fixed cells (Kedersha and Anderson, 2007), but understanding the dynamic nature of SGs and PBs requires live cell imaging. This presents unique challenges, because it requires the overexpression of fluorescently tagged SG/PB marker proteins, which can shift the mRNA equilibrium toward SGs or PBs, thus obscuring the result. We describe stably expressed, fluorescently tagged SG and PB markers that exhibit similar behavior to their endogenous counterparts, thus allowing real‐time imaging of SGs and PBs.

Introduction

Cells reprogram their translation state in response to developmental cues, cell cycle, hormonal or other physiologic signals, or environmental changes. In contrast to genetically programmed developmental changes, environmental changes are unpredictable. Survival requires that these changes trigger a rapid response, in which ongoing translation is arrested, polysomes are disassembled, and translation machinery is reprioritized. Although some level of mRNA turnover occurs continuously, severe stresses activate the eIF2α kinases (e.g., PKR, PERK, HRI, and GCN2), which reduce levels of eIF2‐GTP‐tRNAiMet ternary complex, thereby preventing assembly of the 48S preinitiation complexes (Anderson, 2002, Anderson and Kedersha, 2006). Stalled initiation drives polysome disassembly as ribosomes terminate and run off, leaving mRNA transcripts bound to abortive 48S complexes. The sudden influx of disassembled polysomes overloads the cell's ability to process and remodel these transcripts, so they are temporarily packaged into SGs and PBs. It is important to realize that proteins and mRNA transcripts are not brought “to” SGs or PBs. Instead, mRNA transcripts and proteins participate in the assembly of SGs and PBs, much as water flows into and defines a river. Both SGs and PBs are dynamic structures whose contents are in continuous flux. Although P‐bodies appear continually present in some cell lines such as COS or HeLa, we observe that quiescent cultures (e.g., cells stably arrested in G0 because of serum starvation) exhibit few or no PBs and no SGs. However, when stressed, even quiescent cells respond by forming SGs and PBs, indicating that the basal levels of translation in arrested cells are sufficient to require stress‐induced reprogramming.

SGs are transient structures that are assembled rather hastily as a consequence of interrupted RNA translation. SGs contain several classes of components, any of which may legitimately be used to follow SG dynamics, but these exhibit different kinetics and some may also be present in PBs. SG components include mRNA, SG‐associated mRNA binding proteins (such as argonaute, ataxin‐2, BRF‐1, CPEB, FMRP, FXR1, HuR, PABP1, smaug, staufen, TIA‐1, TIAR, TTP, pumillio, ZPG1), components of the translation initiation machinery itself (such as eIF3, eIF4F, eIF4B, and small ribosomal subunits), and other proteins associated with SGs in less obvious ways (FAST, ORF1p, SMN, SRC3). Live studies that used tagged TIA‐1, TIAR, PABP, G3BP, TTP, FAST (Kedersha, 2000, Kedersha, 2005), GFP‐Ago2 (Leung et al., 2006), and hnRNPA1 (Guil et al., 2006) have revealed that different SG and PB components exhibit different residence time within SGs and PBs.

Transient expression of many SG‐associated proteins results in the spontaneous appearance of seeming SGs in the absence of stress. In some cases, as in the case of enforced expression of phospho‐eIF2α mimetic S51A (Kedersha et al., 1999) or of eIF2α kinases such as PKR or HRI (our unpublished results), the “spontaneous” SGs are logically the result of eIF2α phosphorylation and thus stalled initiation. In other cases, overexpression of translational silencers such as FMRP, FXR1, TIA‐1, TIAR, CPEB, p54/RCK, or argonaute (Khandjian et al., 2004; 2007; Wilczynska et al., 2005) may recruit specific subsets of mRNAs to assemble SGs from their high‐affinity mRNA targets. In other cases, the mechanism whereby overexpression of SG‐associated proteins produces SGs is not clear (notably SMN and FAST [Hua and Zhou, 2004, Kedersha, 2005]). Regardless of the cause, the use of a tagged protein that induces SGs by its overexpression does not allow one to study the SG assembly process itself. Stable cell lines expressing tagged proteins can be obtained that no longer display SGs in the absence of stress, as was recently shown by Leung and Sharp (2006), who stably expressed GFP‐Ago2 in HeLa cells. However, in other cases, overexpression of tagged PB marker proteins (e.g., hedls/GE‐1, DCP1a) can result in giant PBs and increases the expression of other PB components (Fenger‐Gron et al., 2005). In the case of DCP1a, stable cells expressing low levels of FLAG‐tagged DCP1a were obtained, which were used for biochemical studies of PB subcomplexes (Fenger‐Gron et al., 2005). We, therefore, undertook the development of cell lines stably expressing fluorescently tagged SG or PB markers to allow the study of SGs and PBs.

Section snippets

Choice of SG and PB marker proteins

Table 26.1 indicates some of the proteins we have successfully tagged and stably expressed in U2OS osteosarcoma cells. Two different classes of proteins are possible for use as SG markers: RNA‐binding proteins associated with mRNA and components of the translation machinery itself. In our system, we are able to stably express tagged versions of the RNA‐binding proteins TIA‐1, TIAR, PABP1, eIF4E, and G3BP1. The resulting cell lines display normal SG assembly and disassembly, and the tagged

Choice of fluorescent tag

Fluorescent tags are now available in a wide spectrum of colors, stability, and brightness (see Shaner et al. [2007] for a helpful review). Earlier versions of GFP or dsRFP are not optimal for SG or PB studies, because these proteins tend to form dimers and tetramers—obviously not desired properties for tagging protein markers of structures whose assembly is regulated by aggregation. The choice of fluorescent tag should be compatible with the filter sets available on the particular fluorescent

Selection Criteria

Cells expressing tagged SG/PB markers should exhibit normal behavior, defined as follows: (1) cells should exhibit no SGs and few PBs under normal conditions, (2) cells should display rapid SG/PB assembly upon treatment with sodium arsenite or other stresses, and (3) cells should completely disassemble PBs and SGs when treated with emetine or cycloheximide, agents that stabilize polysomes (see Table II of Kedersha and Anderson [2007] for a list of different stresses that induce SGs). To confirm

Transfection

  • We use SuperFect® (Qiagen) as our transfection agent; any reliable agent may be used.

  • Plate cells in a 6‐well tissue culture plate to be nearly 80% confluent. For U2OS cells, a concentration of 5.0 × 105 cells per well is sufficient.

  • Once cells have spread and appear to be growing and dividing, approximately 6 h later, they are ready to be transfected.

  • In a 4‐ml polystyrene tube, aliquot 100 μl of serum‐free DMEM. The use of a serum‐free medium here is imperative, because the presence of serum may

Properties of Representative Stable Lines

Stable overexpression of the classical SG markers TIA‐1 and TIAR is difficult because of the translational silencing activities of these proteins that inhibit cell growth. After drug selection, a very low percentage of cells (<5%) express any fluorescence, and those that do express have a dim signal unsuitable for viewing over time (e.g., as required for time‐lapse imaging). After subcloning, only half of the YFP‐positive clones exhibited normal recruitment to SGs because of rearrangement of

Environmental Control

Environmental control is critical to imaging live cells, especially because heat shock can induce SGs. Furthermore, although photobleaching measurements taken on preformed SGs at room temperature are similar to those obtained at 37° C, SG assembly is dramatically slowed at room temperature, and movies of SG assembly may not recapitulate normal progression. Experimental temperature is often difficult to control. Although it is ideal to run a live cell experiment at 37° C, adding an external heat

Microscope Hardware: Widefield vs Confocal

An ideal live cell video will have high x, y, and z resolution, rapid frame rate for time or temporal resolution, and high signal‐to‐noise ratio (signal:noise), with minimal light exposure. Unfortunately, some aspects must be optimized at the expense of others. High‐resolution images, for example, take longer to acquire, thereby decreasing the frame rate, which in turn increases bleaching of the fluorophores and thus decreases signal:noise and damages the cells with intense light. In some

Collection of protocols and guides

Fluorescence spectra viewers

General microscopy information

Nature microscopy submission guideline

Live cell imaging for tracking processing bodies

The goal is to create a high‐frame‐rate movie that is high enough resolution to track PBs, without excessively photobleaching the specimen.

Day 1:

Day 2:

  • 1

    Replace medium with HBSS in chambers that will not be used immediately. Add cells to 90%

Conclusions

SGs and PBs exhibit several types of dynamic behavior: (1) assembly and disassembly, (2) motility within cells and docking with each other, and (3) turnover of various contents, especially RNA and RNA‐binding proteins. Each of these dynamics can be altered by overexpression of individual SG/PB markers, yet only by overexpression of fluorescently tagged proteins can we view these dynamic processes in real time. Stably expressing cell lines are superior to transient lines for imaging studies and

References (22)

  • K. Irvine

    Rasputin, more promiscuous than ever: a review of G3BP

    Int. J. Dev. Biol.

    (2004)
  • Cited by (0)

    View full text