RNA imaging in living cells – methods and applications

Abstract

Numerous types of transcripts perform multiple functions in cells, and these functions are mainly facilitated by the interactions of the RNA with various proteins and other RNAs. Insight into the dynamics of RNA biosynthesis, processing and cellular activities is highly desirable because this knowledge will deepen our understanding of cell physiology and help explain the mechanisms of RNA-mediated pathologies. In this review, we discuss the live RNA imaging systems that have been developed to date. We highlight information on the design of these systems, briefly discuss their advantages and limitations and provide examples of their numerous applications in various organisms and cell types. We present a detailed examination of one application of RNA imaging systems: this application aims to explain the role of mutant transcripts in human disease pathogenesis caused by triplet repeat expansions. Thus, this review introduces live RNA imaging systems and provides a glimpse into their various applications.

Introduction

RNA molecules are synthesized in cells through highly regulated biogenesis pathways, and when these molecules mature, they participate in and regulate fundamental cellular processes. To gain better insight into the variety of RNA functions, the entire cellular life of transcripts, from their synthesis to their decay, needs to be examined. Molecular methods widely used in RNA biology can identify the length, sequence and structure of an RNA molecule and can be used to determine the mean cellular RNA levels in cell populations. Pull-down assays provide insight into the proteins interacting with RNAs but do not elucidate the spatial and temporal changes that ribonucleoprotein complexes undergo. Fluorescence in situ hybridization (FISH) enables the observation of RNA localization but captures only a single time point in the cellular RNA pathway. The main advantage of live RNA imaging systems is that they allow the study of RNA movement and temporal changes in RNAs and RNPs. These techniques include direct RNA labeling,^1,2 the labeling of endogenous transcripts with molecular beacons,^3-6 multiply labeled tetravalent RNA imaging probes (MTRIPs)⁷ or Pumilio,⁸ and vector-based systems that employ highly specific RNA-protein and RNA-dye interactions. RNA live imaging systems substantially broaden the types of information that can be attained compared with static in situ analyses, by adding a dynamic dimension.

Live imaging systems enable researchers to investigate many cellular processes involving RNAs. Among these processes are the biosynthesis, function and decay of eukaryotic mRNAs, which include multiple steps: i.e., Pol II transcription, primary transcript (pre-mRNA) modifications and splicing, nuclear transport and export of mature mRNA, cytoplasmic mRNA transport to its localization site, mRNA translation and, finally, degradation. With live RNA imaging systems, the entire cellular route of a transcript may be observed in a single experiment, and each step of this pathway may be analyzed separately and in more detail.

In this article, we describe the vector-based systems for RNA imaging in living cells and the adaptations of these systems to various applications in bacterial, fungal, plant and animal cells. We provide details regarding the design of these systems and their important features and critically discuss the advantages and limitations of the individual systems. In the application section, we briefly refer to relevant publications, organizing them according to the cellular processes investigated. We present a detailed discussion of 2 applications of live RNA imaging systems that address the role of RNA nuclear foci in the pathogenesis and treatment of human neurological diseases caused by triplet repeat expansions.

Aptamer-Based Transcript Imaging Systems

Nearly all aptamer-based systems that have been developed for the fluorescence microscopy imaging of single transcripts in living cells require the expression of an exogenous transcript of interest fused to an aptamer sequence (chimeric transcript). The aptamer itself is not fluorescent, but when it binds a specific protein partner fused to an autofluorescent protein (chimeric protein), the entire system becomes fluorescent after excitation at appropriate wavelength. Alternatively, the chimeric transcript may be imaged using a suitable organic dye whose fluorescence can be activated after binding to an aptamer. The rigorous high-affinity and high-specificity criteria are fulfilled by a number of systems that have thus far been devised for use in RNA imaging (Table 1).

Table 1.

Characteristics of vector-based systems for RNA live imaging. Affinity is presented as K_D (dissociation constant) – the propensity of a complex to dissociate into parts. Legend: * - Eukaryota/Prokaryota, ** - threshold not estimated, ‘-’ - not applicable, NA - not analyzed, app. – approximately, ? - not specified by authors. Unless stated otherwise, data were obtained from studies cited in the systems description paragraph

		RNA-dye systems
Name Feature	MS2	λN22	BglG	PP7	U1Ap	HTLV-1 Rex	TAT-TAR	REV-RRE	eiF4A	Spinach	Spinach2	Malachite Green	SRB-2
RNA size	19 nt	15 nt	29 nt	25 nt	21 nt	36 nt	23 nt?38	30 nt?38	58 nt	98 nt	95 nt	38 nt	54 nt
number of hairpins	1–96	1–25	18	1–24	4–16	1	1	1	2	1	1	1	1
number of hairpins required for single molecule imaging	24/96*	≤12**	≤18**	≤24 **	NA	NA	NA	NA	NA	NA	NA	NA	NA
minimal peptide size used (aa)	117	22	58?20	128	94	16	9–86?38	14–116?38	215 +191	—	—	—	—
protein dimerization	YES	NO	NO	YES	NO	NO	NO	NO	—	—	—	—	—
affinity of ligands to the aptamer(KD)	5 nM	22 nM	NA	1 nM	< 1 nM?33	30 nM	0.31–2.1 nM?40	1–3 nM?39	44 nM	450 nM	430 nM	110 nM–1.8 μM	1.4 ± 0.1 μM
BiFC	NO	NO	NO	NO	NO	NO	NO	NO	YES	—	—	—	—
background problem	YES	YES	YES	YES	YES	YES	YES	YES	NO	NO	NO	NO	NO
comment	—	—	BglG is a bacterial protein	—	U1Ap is a human protein	—	Only tested in TriFC experi-ment	Only tested in TriFC experi-ment	eiF4A is a murine protein	thermal instability	+ tRNA sequence	—	+ tRNA sequence

Aptamer-protein systems

The aptamer-protein systems are based on naturally occurring high-affinity interactions between specific RNA structure motifs and their binding proteins. Both the RNA and protein components used in these systems are typically engineered to optimize their sequences for tighter binding. Often, only the RNA binding domain of the protein is retained to minimize the size of the protein-RNA complex. The main drawback of aptamer-protein systems is the strong background signal produced by the constant fluorescence of the chimeric protein not bound to the aptamer, which we address further in the text as a background problem.

MS2 system

The MS2 system is the prototypical RNA imaging system design introduced by Singer and colleagues in the study of ASH1 mRNA localization in yeast cells⁹ and first applied by Bloom and colleagues to study the localization of mRNAs in budding yeast.¹⁰ In the MS2 bacteriophage, the genomic RNA and its coat protein form a viral capsid through protein multimerization and RNA-protein interactions. The MS2 system (Fig. 1a, Supplementary Table 1) consists of the RNA operator of the MS2 bacteriophage genome and an engineered MS2 protein that binds to this RNA as a dimer and has an enhanced RNA-protein affinity.^11,12 The aptamer fused to a transcript of interest forms a 19 nt RNA hairpin with a 4 nt loop and a 7 bp stem that harbors a single adenine bulge. The U>C mutation in the natural loop sequence increased protein binding 50 times.¹³ Because a single MS2 RNA hairpin is not sufficient to image a single molecule of transcript, multiple hairpins are typically used. The advantages of the MS2 system, such as high specificity and sensitivity, led to its position as the most frequently used system to label and image RNA in living cells (Supplementary Table 2). To date, more than one hundred publications have described the use of MS2 in different biological systems imaging a broad range of RNAs.

Figure 1.

Vector-based systems for RNA live imaging. Schematic structures of RNAs with protein partners or fluorescent dyes are presented (a–l). Additionally, examples of genetic constructs used for imaging experiments are depicted (o–r). (a) MS2 systems, (b) λN22 system, (c) BglG systems, (d) PP7 system, (e) U1Ap system, (f) HTLV-1 Rex system, (g) TAT system, (h) REV system, (i) eIF4A system, BiFC with the use of 2 domains, (j) Spinach system, (k) Malachite green system, (l) SRB-2 system, (m) BiFC with the use of 2 systems, (n) TriFC, (o) DNA construct for MS2 system, (p) DNA constructs for BiFC with 2 systems and (r) system for gene locus, mRNA and protein product imaging.

λN22 system

The λN22 system developed by Daigle and Ellenberg, which was first described in mammalian cells,¹⁴ takes advantage of the phage lambda transcription antitermination signal (box B RNA) and its binding N protein (Fig. 1b, Supplementary Table 1). In the λN22 system, only the 22 aa arginine-rich fragment of the N protein binds the aptamer, which is a 15 nt hairpin composed of a 5 nt loop and a 5 bp stem (Table 1). The hairpin loop of the box B RNA contains a GNRA fold (R = purine, N = any nucleotide), which is tightly bound by the N protein.¹⁵ The small size of the fused N protein fragment is a significant advantage of the system because this feature minimizes the risk of influencing the cellular properties of the investigated transcript.¹⁴ The background problem found in the MS2 system, which is partially caused by the proteolysis of the peptide linker between MS2 and GFP, was reduced in the λN22 system by increasing the stability of the chimeric λN-GFP protein.¹⁶ The λN22 system is the second most frequently implemented system after MS2, and it has been used to image transcripts in bacterial, fungal, plant and mammalian cells.

BglG system

The BglG system (Fig. 1c, Supplementary Table 1) developed by Hu and colleagues¹⁷ was used for the first time together with the MS2 system to demonstrate the heterozygosity of HIV-1 virions in human cells. Two genomes were labeled, one with MS2 and one with a BglG aptamer, and they were imaged simultaneously to differentiate the virus variants and to study virion formation. The system is based on the E. coli transcription antitermination protein, which in principle limits its applications in bacterial cells. The BglG protein binds to an imperfectly palindromic 29 nt hairpin containing a 4 nt terminal loop and 2 bulges in the hairpin stem^18,19 (Table 1). In contrast to most other systems, the RNA-protein specific recognition triggers structural changes leading to hairpin destabilization. The minimum peptide size that is able to efficiently bind the RNA structure is 58 aa and spans the N-terminal RNA binding domain of the BglG protein.²⁰ So far, this system has been used only in studies of viruses.²¹

PP7 system

The PP7 system (Fig. 1d, Supplementary Table 1) was introduced in 2011 by Singer and colleagues to study transcription initiation and elongation in yeast.²² It is based on a similar principle as the MS2 system. The dimer of the PP7 bacteriophage coat protein binds to its cognate RNA structure with very high affinity (Table 1). The aptamer is built from an RNA hairpin containing a 6 nt loop and an 8 bp stem harboring a purine bulge on its 5’ side.²³ The PP7 system was used together with the MS2 system to enable bimolecular fluorescence complementation (BiFC) through the reconstruction of the EGFP protein from its fragments (Fig. 1m, p).^24,25

U1Ap system

The U1Ap system (Fig. 1e, Supplementary Table 1) was developed simultaneously by Silver and Brodsky²⁶ and by Takizawa and Vale,²⁷ and has been used thus far in only 4 studies carried out in yeast cells.^28-31 The system is based on a fragment of the human splicing protein U1Ap and the RNA signal it recognizes. To reduce the size of the protein, a fragment containing the first 94 aa, which comprises the RRM domain (the domain that binds RNA), was fused to a fluorescent protein (Table 1). This polypeptide shows a high affinity for a 21 nt hairpin structure containing a 10 nt loop.^32,33 U1Ap was shown to be advantageous in the investigation of the nuclear export of RNA.²⁶

HTLV-1 Rex system

The HTLV-1 Rex system (Fig. 1f, Supplementary Table 1), proposed by Broude and colleagues,³⁴ was used together with λN22 in a BiFC experiment performed in E. coli. The viral peptide used in this system is only 16 aa, which makes it the smallest peptide used in an imaging system (Table 1). The peptide binds specifically to a 36 nt hairpin based on the HTLV-1 Rex responsive element (RxRE), which contains a 4 nt loop and 2 bulges in the stem.³⁵ Physiologically, the RxRE present in the viral RNA is bound by the HTLV-1 Rex protein, which enhances its export from the nucleus.³⁶ It remains to be established whether a chimeric HTLV-1 Rex protein also affects mRNA export.

TAT-TAR and REV-RRE systems

Two novel systems based on sequences present in HIV were introduced recently by Cui and colleagues.³⁷ The natural affinities of the TAT (Fig. 1g, Supplementary Table 1) and REV (Fig. 1h, Supplementary Table 1) proteins to the TAR and RRE RNAs,^38-40 respectively, were used to investigate mRNA-protein interactions. Both systems use unmodified RNA sequences, which means that their RNA components are relatively large. The systems were invented as alternatives to earlier designs, taking advantage of fluorescence complementation, and applied to the study of the nuclear export of viral RNAs.³⁷

eIF4A system

The eIF4A system was first described in 2007 by Broude and colleagues and evaluated for the imaging of mRNA and rRNA.⁴¹ The system is based on the mouse version of the eukaryotic initiation factor 4A peptide (Table 1), which contains 2 RNA binding domains and binds with a high affinity to a single relatively large 58 nt aptamer.⁴² This feature can be exploited using fluorescence complementation, in which the 2 domains are dissected and each is fused with half of a fluorescent protein (Fig. 1i, Supplementary Table 1).⁴³

Aptamer-Dye systems

An alternative to the protein-aptamer RNA imaging methods is the replacement of the protein component with a small organic dye. Fluorescence is generated after excitation only when the dye is captured by a specific RNA aptamer structure because the dye can no longer dissipate its energy through intramolecular motions. In this design, the undesired fluorescence background problem was not observed.⁴⁴ The replacement of proteins with small organic dyes and shorter sequences fused to the investigated transcripts diminishes concerns that proteins or polypeptides bound to aptamer sequences influence transcript properties. The high potential of aptamer-dye systems to image RNAs in various biological systems has been highlighted by several authors.^45-47 There are many aptamer-dye pairs described for in vitro RNA labeling, but so far, they have not been implemented in biological studies. Among them are the II-mini3–4 system,⁴⁸ dimethyl indole red,⁴⁹ ASR7⁵⁰ and DCF-MPP.⁵¹ The systems described below are the first examples of aptamer-fluorophore techniques, which were recently developed and used in living cells.

Spinach and Spinach2 systems

The Spinach system was introduced for RNA imaging by Jaffrey and colleagues in 2011.⁴⁴ The organic dye, 4-hydroxybenzylidene imidazolinone (HBI), corresponds to the fragment of GFP responsible for its fluorescence. The fluorescent signal is generated when the fluorophore is captured by an 80 nt aptamer fused to the mRNA of interest (Fig. 1j, Supplementary Table 1). Several derivatives of HBI provide a range of different fluorescence wavelengths (different colors). So far, DFHBI (3,5-difluoro-4-hydroxybenzylidene imidazolinone) has been shown to be the most efficient (Table 1); however, the dye can be further improved.⁵² More fluorophores could be synthesized based on a range of fluorescent proteins (GFP derivatives) because the chemical structure that determines its fluorescence is known.⁵³ These dyes have reduced photobleaching due to fast kinetics of exchange of DFHBI dye in Spinach aptamer, which imposes some limitations on the application of the system. Recently, Spinach2 was introduced due to problems with the thermal stability of the Spinach aptamer in the cellular environment. The aptamer was mutated and stabilized using the tRNA structure to facilitate proper folding. With these improvements, the RNA containing the CGG repeats was imaged.⁵⁴

Malachite green system

The malachite green system (Fig. 1k, Supplementary Table 1) with 2-photon excitation (TPE) signal-enhancement was introduced by Nicoud and colleagues.⁵⁵ The malachite green-binding aptamer (38 nt) is considerably shorter than the Spinach aptamer; therefore, it is even less likely to alter transcript properties (Table 1).⁵⁵ The interactions of malachite green (MG) with the RNA aptamer were previously studied^56-58 but without implementation in the field of RNA imaging. The malachite green system was used in the investigation of RNA degradation in mammalian cells. In this experiment, the malachite green aptamer was stabilized by the pRNA 3WJ sequence.⁵⁹

SRB-2 system

The most recently introduced system is the SRB-2 system (Fig. 1l, Supplementary Table 1), which was developed by Sunbul and Jaschke.⁶⁰ It uses a well-described 54 nt SRB-2 aptamer and Sulforhodamine B (SR) as the dye (Table 1). Because of the constant fluorescence of SR, a dinitroaniline (DN) quencher was added to create the small molecule marker SR-DN for the SRB-2 aptamer. The optimal length of the linker placed between the fluorophore and the quencher was shown to be 2–3 ethylene glycol units. After the aptamer binds the small dye, the labeled RNA is observed as a red signal. The performance of the system has been tested only in E. coli cells so far.

Critical issues in aptamer-based systems design

A crucial issue in the experimental design of RNA imaging systems is the site of the aptamer insertion within the transcript. The most common practice in mRNA trafficking studies is to place the aptamer sequence between the ORF and the 3’ UTR (Supplementary Table 2). This location is considered to be the safest and the most likely to preserve the properties of the endogenous transcript because mRNA localization signals are not altered and mRNA translation is expected to occur normally. For some research purposes, the experimental system may be simplified: e.g., for localization analysis only, the 3’ UTR of the mRNA may be used.

Some applications may require aptamers to be placed at positions other than between the ORF and the 3’ UTR (Table 2). However, it has been shown that the site of aptamer insertion may affect the level of protein translated from the chimeric mRNA,⁶¹ influence transcript localization and interactions with endogenous proteins⁶² and disturb pre-mRNA splicing⁶³ (Table 2). It is therefore hard to provide general guidelines regarding the optimal placing of the aptamer sequence within the investigated transcript. To ensure that the cellular pathway and the final localization of the mRNA are not affected by the aptamers, FISH experiments are typically performed.

Table 2.

Localization of hairpins within transcripts: opportunities and concerns

localization		research purposes	disadvantages
in the 5’ UTR		protein binding	nonsense peptide production, decreased protein level
in the ORF	intron	splicing localization	splicing alteration
	exon	alternative splicing	protein alteration
between the ORF and the 3’ UTR		mRNA localization	decreased protein level
in the 3’ UTR		protein binding	protein binding prevention, transport alteration

A second important issue is the background problem, which occurs when autofluorescent proteins are used. To solve this problem and increase the signal-to-noise ratio, several options have been considered. One possible solution is to increase the number of GFP units attached to a single aptamer-binding protein⁶⁴; another is to multiply the aptamer sequences in the transcript. For a system based on the GFP-MS2-NLS protein, 24 MS2 hairpins enabled the imaging of a single mRNA molecule in mammalian cells.⁶⁵ Concerns that the attachment of large RNA-protein complexes can limit the use of these systems to only large RNAs were dispelled by the successful application of live imaging techniques to small RNAs.^66,67 In studies of cytoplasmic events, a nuclear localization signal (NLS) may be added to a chimeric protein to direct the unbound protein to the nucleus^9,14 However, it must be verified that the addition of an NLS does not alter the localization of the imaged RNA. Yet another way to resolve the background problem is the use of fluorescence complementation systems^43,68,69 as reviewed by Tyagi⁷⁰ (Fig. 1i, m, p).

A transcript with an aptamer sequence and a chimeric protein is typically delivered to cells using genetic vectors. Two standard plasmids harboring relevant expression cassettes are used as depicted in Figure 1o. Using differently designed cassettes, the mRNA, protein and gene locus can be imaged in one experiment (Fig. 1r).⁷¹ In the case of aptamer-dye systems, only one vector is needed to image a single mRNA. Small fluorophores easily enter the cell; however, some of them: e.g., MG, cause considerable cytotoxicity.⁴⁴ The use of 2 different systems allows researchers to image transcripts from both alleles of the same gene⁷² or to demonstrate the co-localization of distinct mRNAs⁶² in a single cell.

Examples of Applications of Live RNA Imaging Systems

Numerous types of RNAs have been imaged in both prokaryotic and eukaryotic cells (Supplementary Table 2), including mRNAs, non-coding RNAs and RNA viruses. However, the most frequently investigated topic is the life cycle of mRNA in mammalian cells. In the following section, we describe which aspects of the cellular mRNA journey can be and already have been explored using live RNA imaging systems. We discuss in more detail the results of RNA imaging experiments focused on the roles of mutant RNAs in neurodegenerative diseases caused by triplet repeat expansions.

Nuclear journey of mRNA

Monitoring transcript synthesis begins with the measurement of gene activity. This analysis includes studying transcription initiation and single transcription events.⁷³ The number of newly synthesized transcripts can be counted in a time-lapse manner, and temporal changes in transcriptional activity can be measured under changing conditions. Variations in the transcription dynamics of the investigated promoter have been analyzed during the cell cycle^74,75 and in response to various transcription inducers or inhibitors.^76-78 Additionally, the time period elapsed from the recruitment of initiation factors to the start of transcription has been be measured.⁷⁹

Live imaging systems allow the determination of the transcript elongation rate and a detailed description of RNA polymerase II activity.²² To distinguish between newly synthesized transcripts and transcripts already released from the transcription site, the Fluorescence Recovery After Photobleaching (FRAP) technique was used.^78,80 The transcript elongation rate was measured with dually labeled RNA, in which PP7 hairpins were inserted into the 5’ UTR and MS2 hairpins were inserted into the 3’ UTR. The time between capturing the signals from these tags corresponds to the time required for the elongation of the sequence between the PP7 and MS2 hairpins.⁷²

The nuclear steps accompanying and following transcription include splicing and 3’ UTR formation. Both steps can influence the retention of pre-mRNA at a transcription site. In contrast to transcripts devoid of introns, transcripts containing introns stay at the transcription site longer than the polymerase, which implies that splicing stops the pre-mRNA at the transcription site.⁸¹ In studies of splicing dynamics, the aptamer sequence was placed inside the intron, and the duration of its signal was measured.⁶³ Using RNA imaging systems, it was also shown that mRNA is released from the transcription site after 3’end formation.⁸²

After the nuclear processing of the transcript is completed, the mature mRNA is transported to the nuclear borders to be exported from the nucleus. By imaging the mobility of RNA molecules, it was shown that transcript movement is not directional but is influenced by energy.⁸³ Transcripts leave the nucleus in different ways,⁸⁴ typically through nuclear pores. Export was shown to be faster than the nucleoplasmic diffusion rates of mRNA.⁸⁵ The retention of mRNAs was observed in cells with decreased ATP levels,⁸³ disturbed splicing, 3’end processing²⁶ and triplet repeat expansion mutations.^54,86

Cytoplasmic journey of mRNA

Transcripts accompanied by various proteins are transported in the cytoplasm in many ways, including along actin filaments, along microtubules or through simple diffusion. The transport of numerous transcripts has been investigated with live imaging systems to answer questions regarding the velocity,⁸³ efficiency,⁸⁷ direction^28,88 and continuity⁸⁸ of transcript movement. These features have been shown in many cases to differ substantially between the analyzed transcripts. Therefore, instead of attempting to provide the reader with specific answers and numbers, which is beyond the scope of this review, we only provide references to the relevant papers.

Numerous analyses were performed to study the proteins involved in transcript transport.^9,89 Interactions between RNAs and proteins are typically analyzed using double-labeling co-localization experiments. The influence of specific proteins on RNA transport can also be investigated in cells with silenced expression of the implicated protein.⁹⁰ Additionally, live imaging systems allow precise analyses of the mRNA sequences responsible for transport and localization. Two approaches for such analyses were proposed: the use of a construct containing only the transport signal, and mutating this signal to disturb transcript transport.^17,91,92

The use of the Trimolecular Fluorescence Complementation (TriFC) technique has shown that some protein-protein interactions are observed only in the presence of specific RNA molecules.⁹³ A fluorescent protein is split into halves; one half is fused with the protein required for RNA imaging, and the other is fused with a protein indirectly bound to the RNA. The fluorescent signal is restored only when all components co-localize (Fig. 1n).

The retention of mRNAs in specific cytoplasmic bodies occurs when the transcript is stored for delayed translation or degradation.^94,95 The exchange of mRNAs between the motile fraction and stable granules can be analyzed using the FRAP technique.⁹⁶ Disturbances in transcript localization caused by various environmental factors, e.g., amino acid starvation or temperature changes, were also monitored with RNA imaging systems.^97,98

The labeling of both mRNAs and their protein products has allowed researchers to answer several questions about translation. It has been shown that transcripts are not translationally active during transport in the cytoplasm.⁹⁹ Addressing the question of whether transcript localization depends on protein translation resulted in the finding that the localization of some mRNAs defines proper protein localization⁹⁹ and in the converse direction, translation activity can define transcript localization in an SRP-dependent manner.¹⁰⁰ Further analyses demonstrated that several proteins and microRNAs regulate RNA stability¹⁰¹ and influence the translation^{90 102,103} of the investigated transcripts.

The cellular life of RNA is terminated in several ways. The main mechanism of mRNA decay usually starts with deadenylation, continues through 5’ decapping and ends with 5’→3’ exonucleolytic degradation.¹⁰⁴ The analysis of mRNA degradation using live imaging systems was conducted by analyzing transcript co-localization with the cellular compartments responsible for degradation: e.g., P bodies.^31,105,106 The site where transcript degradation occurs was observed after inhibiting its nucleolytic decay,¹⁰⁷ and the dynamics of RNA degradation were monitored in cells after the global inhibition of transcription.¹⁰⁸

RNA imaging in diseases caused by simple repeat expansions

The pathology of a group of genetic neurodegenerative diseases associated with a simple repeat expansion may be caused by toxic effects induced by the mutant transcript, mutant protein or both. Another mechanism of pathogenesis is the decrease in the amount of functional protein product translated from the mutant gene. The earliest symptom is often neuronal cell dysfunction; therefore, studies examining the pathogenesis of these diseases should be performed in relevant experimental models. Many experiments taking advantage of RNA imaging in living cells have been conducted thus far in neuronal cell lines (see Table 3). In this section, we summarize the results of imaging experiments that explore the aberrant life of transcripts from mutant alleles of the genes involved in triplet repeat expansion diseases.

Table 3.

Examples of RNA live imaging systems used in studies on neuronal cell line. The other studies are presented in Supplementary Table 2

mRNA function examined	specific research purpose	system used	imaged mRNA	Ref.
transcriptional activity	β-actin mRNA transcription and transport	MS2	β-actin mRNA	135
RNA localization	dendritic targeting signals	MS2	Kv4.2 mRNA	136
	localization signals of nos mRNA	MS2	nos mRNA	137
	CaMKIIα mRNA localization	MS2	CaMKIIα mRNA	94
	region responsible for dendritic transport	MS2	ApoE mRNAs	92
	localization of Arc mRNA	MS2	Arc mRNA	88
	5’ UTR and 3’ UTR transport signals	MS2	kor, SV40 mRNAs	138
	altered localization in memory	MS2	CaMKIIα mRNA	139
RNA movement	MMP-9 mRNA movement	MS2	MMP-9 mRNA	140
	mobility of kor mRNAs	MS2	kor mRNAs	141
RNA transport mechanisms	dynein-dependent transport	MS2	nos, osk mRNA	142
	Stau2 role in mRNA distribution	MS2	Map1b, Map2 mRNA	143
	Kv4.2 transport in dendrites	MS2	Kv4.2 mRNA	144
	Htt role in BDNF mRNA transport	MS2	BDNF mRNA	145
	Htt role in mRNAs transport	MS2, λN22	β-actin mRNA	146
	role of FMRP in mRNA transport	MS2	CaMKIIα, Fmr1 mRNAs	147
	Htt role in dendritic transport	MS2, λN22	IP3R1, β-actin mRNA	148
	FMRP role in mRNAs transport	MS2	CG9293, chic mRNAs	96
	FMRP role in mRNAs transport	MS2	CaMKIIα mRNA	149
RNA-Protein interactions	RNG105 colocalization with NKA mRNAs	MS2	NKA mRNAs	150
	FMRP interaction with MMP-9 mRNA	MS2	MMP-9 mRNA	151
	Copb1 function	MS2	kor, SV40 mRNAs	152
	SYNCRIP role in mRNA granules	MS2	IP3R1 mRNA	153
RNA translation	synapse-specific mRNA translation	MS2	Arc mRNA	154
	FMRP and hnRNP C competitive translation control	MS2	APP mRNA	90
RNA stability	DLK-1 function	MS2	CEBP-1, UNC-54 mRNA	101

RNA toxicity manifests mainly through the formation of nuclear foci. Ribonuclear inclusions are observed in the cells of patients with neurological diseases caused by trinucleotide (CAG in SCA3 in HD,^109,110 CTG in DM1^111-113 and HDL2,¹¹⁴ CGG in FXTAS^115,116), tetranucleotide (CCTG in DM2^117-119), pentanucleotide (TGGAA in SCA31,¹²⁰ ATTCT in SCA10¹²¹) and hexanucleotide (GGGGCC in ALS/FTD^122,123) repeat expansions as reviewed by Wojciechowska and Krzyzosiak.¹²⁴ Transcripts containing expanded repeat tracts form aggregates in which they sequester various proteins. The application of FISH and immunofluorescence techniques allowed for the determination of the sizes and shapes of the inclusions as well as the identification of several sequestered proteins.^125,126 FISH imaging was used to estimate the number of foci in a single nucleus¹²⁷ and to establish whether the foci harbored full-length mutant transcript or only fragments containing expanded repeats.¹¹⁹ The influence of the mRNA expression level on foci formation^113,128,129 and the toxicity of cytoplasmic RNA foci were also described.¹³⁰ However, not all features of RNA aggregates can be described using static methods such as FISH, immunodetection and fluorescent protein labeling.

To gain better insight into the nature of RNA toxicity and its role in pathogenesis, it is crucial to learn more about the differences in the cellular behavior of normal and mutant transcripts. The cellular fates of both transcripts can be followed with the use of one imaging system in different experiments^54,86 or using 2 different imaging systems in one experiment.⁷² The toxicity of the mutant transcripts, understood as abnormal processing or functioning, can be followed at every step of its life.^131,,132 In polyglutamine diseases, mutant transcripts temporarily retained in the nucleus can reach nuclear pores and be exported to the cytoplasm. The effectiveness of their translation may differ from that of normal transcripts.

It is not well understood whether aggregation is fully controlled or stochastic. The important issues concerning nuclear foci include foci dynamics and stability, changes in composition, factors influencing stability and therapeutic molecules targeting RNA foci. All aspects concerning the dynamics of foci formation, mobility, stability and interactions with proteins can be investigated using live RNA imaging systems. Using these methods, 2 types of nuclear inclusions containing transcripts with expanded triplet repeats were analyzed. The studies were performed using the MS2 and Spinach2 systems, which were used to label fragments of the DMPK and FMR1 transcripts, containing mutant CUG and CGG repeats, respectively^54,86 (Table 4).

Table 4.

Details of RNA imaging experiment design used in studies of CUG and CGG repeat toxicity in DM1 and FXTAS

	CUG-repeat transcript imaging	CGG-repeat transcript imaging
Goals	• movement of mutant transcripts • dynamics of foci formation • RNA co-localization with Mbnl1 • Mbnl1 role in foci formation	• RNA foci formation and stability • foci dynamics during the cell cycle • effect of drugs on foci formation • RNA co-localization with Sam68
System	MS2	Spinach2
Number of aptamers	24	1
Localization of aptamers	Upstream of 3’UTR	Upstream of the polyA signal
Fluorescent protein/dye	GFP/mCherry	DFHBI
Delivery method	Retroviral vectors	Plasmid vectors
Promoter	TRE (inducible)	CMV (strong, non-inducible)
Additional techniques	FRAP, FLIP, imaged simultaneously with protein	imaged simultaneously with protein
Cell type	C2C12 myoblasts	COS-7
Temporal resolution	Every 333 msec for 20 sec or every 30 sec for up to 23 min	Every 20 min for 6 h.
Microscopy	Spinning disk confocal microscopy	Epifluorescence microscopy
Software	MetaMorph software	NIS-Elements software
Comment	• Higher temporal resolution is needed for the analysis of transcript movement. • Standard microscopic filters can be used. • Protein-RNA systems enable photobleaching. • Single RNA granule tracking is possible with 24 MS2 hairpins	• Longer observation is needed for changes during the cell cycle. • Fast dye exchange reduces photobleaching. • DFHBI exhibits better fluorescence signal stability than GFP. • DFHBI can be replaced with DFHBI-1T for standard microscopic filters
Reference	Querido et al. 2011	Strack et al. 2013

The movements of the unsequestered mutant transcripts into foci were followed in the nucleoplasm. The mobility of the transcript with expanded CUG repeats was decreased to half that of the normal transcript.⁸⁶ The RNA retardation may contribute to the compromised export of the mutant transcript observed in DM1 cells.¹¹¹ The co-localization of the transcript containing expanded CUG repeats with the MBNL1 protein was detected soon after transcription. This result suggests that the MBNL1 protein actively participates in foci formation rather than being passively sequestered by already-formed foci. In support of this conclusion, the number of foci in the cells decreased after silencing the MBNL1 protein. The interaction of the protein with transcripts present outside of aggregates is consistent with alterations in alternative splicing, which are characteristic of DM1 and also observed in cells devoid of nuclear foci.¹³³

In imaging experiments of transcripts containing CGG repeats, the foci are formed within a few hours from the start of transcription.⁵⁴ The fast transcript aggregation can be explained by the use of a strong promoter in the RNA imaging construct, which is in accord with the results of the FISH experiments.¹¹⁴ The combination of RNA imaging with the FRAP and Fluorescence Loss in Photobleaching (FLIP) techniques enabled the authors to examine the dynamics of CUG foci formation.⁸⁶ Inclusions are formed randomly and have a stochastic nature. With the use of Spinach2, the influence of the cell cycle on foci formation was demonstrated for the first time.⁵⁴ During cell division, all CGG-containing foci combined into one large inclusion, which was then divided between daughter cells. After cell division, the aggregates temporarily dissociated, and the signal from the labeled transcripts was observed across the cytoplasm before rapid foci reconstitution.

The instability of the structure and the composition of the nuclear foci raise hope that drugs targeting nuclear foci could efficiently trigger the disintegration of pathogenic inclusions. With RNA imaging systems, it has been possible to observe how small molecular weight drugs affect foci formation and stability.⁵⁴ The tested drugs either decreased de novo foci formation or induced their disaggregation.

RNA imaging systems have allowed to observe mutant triplet repeat transcripts in one very important phase of their cellular life that leads to pathology. A detailed comparison of normal and mutant transcripts in their entire cellular pathways (Fig. 2) will likely reveal new RNA-mediated pathogenic mechanisms in this group of neurodegenerative diseases.

Figure 2.

RNA cellular journey in neurons. The figure presents RNA-related processes, which were investigated in neuronal cell lines using live imaging systems. Followed RNA and its aggregates are depicted in green, co-transported RNAs are in yellow and interacting proteins are in black. Protein products are illustrated in red, and gene loci are pink.

Final Remarks and Future Perspectives

Insight into dynamics of transcript birth, maturation, adult life and death is one of the major objectives in cell biology. This very ambitious goal cannot be achieved in the short term considering the multitude of coding and noncoding transcripts and their involvement in countless cellular functions. With the advent of RNA live imaging systems described in the first part of this article, many important questions regarding transcripts synthesis, processing, trafficking and interactions were answered, which is evident from the second part of this review. The field of RNA live imaging is no longer in its infancy, mainly because of the widespread applications of the MS2 system, which was invented by Robert Singer and colleagues in the nineties.

There are numerous RNA-related cellular processes current understanding of which would benefit from the application of RNA live imaging systems. These processes include the complex and strictly regulated process of transcription and the multistep process of transcript maturation, which both engage a variety of cellular proteins. An example of a research area that has not yet strongly benefited from RNA imaging systems is the regulation of protein-coding gene expression by numerous non-coding RNAs. Among the extensively investigated but still unresolved issues are the following: How are the mRNAs regulated by microRNAs? By what mechanism do the small RNAs find their cytoplasmic targets? What is the role of the proteins involved in the RNA interference machinery in the cell nucleus¹³⁴? Even more vague is our knowledge of the role of long non-coding RNAs in cells and the functions of multiple antisense transcripts. Biomedical research would benefit from an improved understanding of alterations in RNA dynamics in human diseases so that this knowledge could be used in planning therapeutic interventions.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Funding

This work was supported by Ministry of Science and Higher Education (grant numbers N N301 569340, N N302 633240), National Science Center (2011/03/B/NZ1/03259, 2012/06/A/NZ1/00094), and European Regional Development Fund within Innovative Economy Program (POIG.01.03.01–30–098/08).