Localization and Dynamics of Nuclear Speckles in Plants

The generation of mature mRNAs from most genes (about 80%–90%) in autotrophic eukaryotes requires the removal of noncoding sequences (introns) and splicing of the coding regions (exons; Labadorf et al., 2010). During splicing in some precursor messenger RNAs (pre-mRNAs), the same splice sites are always used, referred to as constitutive splicing (CS), resulting in a single transcript from a gene. However, from many pre-mRNAs, multiple mature mRNAs are generated from a single gene by alternative splicing (AS), where different combinations of splice sites are used. Both CS and AS are critical to the proper expression of intron-containing genes. Recent transcriptome-wide analysis of AS using high-throughput sequencing indicates that pre-mRNAs from up to 42% of intron-containing genes in Arabidopsis (Arabidopsis thaliana; Filichkin et al., 2010) and about 48% in rice (Oryza sativa; Lu et al., 2010) are alternatively spliced, whereas about 95% of human genes are alternatively spliced (Pan et al., 2008). In addition to pre-mRNAs, some primary microRNAs (pri-miRNAs) are also subject to CS and AS (Hirsch et al., 2006; Szarzynska et al., 2009; Mica et al., 2010). AS increases the protein-coding capacity of a genome and generates functionally different proteins from the same gene (Reddy, 2007). AS results in protein isoforms with loss or gain of function and altered subcellular localization, protein stability, and/or posttranslational modifications. Furthermore, AS plays an important role in gene regulation through regulated unproductive splicing and translation, leading to RNA degradation by mRNA surveillance mechanisms, differential recruitment of mRNAs to ribosomes, or translatability of splice variants (Kurihara et al., 2009; Licatalosi and Darnell, 2010; Palusa and Reddy, 2010). Numerous spliceosomal proteins either promote or suppress splicing by interacting with splicing regulatory elements on the pre-mRNAs. In recent years, the localization and dynamics of some splicing regulators in plants have been analyzed using a variety of approaches. This review summarizes the current status of research in this area, with emphasis on RNA-binding proteins (RBPs) that are involved in splicing regulation, discusses open questions, and presents some approaches to address these questions.

SPLICING REGULATORS

pre-mRNAs splicing takes place cotranscriptionally in the spliceosome, a large multicomponent complex composed of small nuclear RNAs (snRNAs) and about 170 proteins, many of which are involved in the regulation of splicing (Wahl et al., 2009; Valadkhan and Jaladat, 2010). The composition of the human and yeast spliceosome has been analyzed in great detail (Wahl et al., 2009); however, information on plant splicing has been scarce, as no in vitro assembly of a functional plant spliceosome has been possible to date. A detailed search for Arabidopsis orthologs of the human spliceosome proteome revealed that most splicing factors are present in the Arabidopsis genome, indicating a similar complexity (Barta et al., 2011). Interestingly, many of the plant orthologs have close or related homologs evolving from several genome duplication events with possible plant-specific functions. This is best exemplified by the Ser/Arg-rich (SR) protein subfamilies (see below). In Arabidopsis, there are over 200 RBPs, including many unique to plants (Lorković, 2009), suggesting plant-specific functions. Some of them are components of small nuclear ribonucleoprotein particles (snRNPs), while many others are not but are present in the spliceosome and hence are called non-snRNP proteins. A survey of RBPs including RNA recognition motif (RRM)-containing proteins, such as SR proteins, SR-like proteins, heterogeneous ribonucleoprotein particles (hnRNPs), and RNA-binding KH domain-containing proteins, revealed that many of them play key roles in regulating CS and AS. Some of these RBPs are relatively well studied in plants (Lorković, 2009; Reddy and Ali, 2011).

SnRNP PROTEINS

There are two types of spliceosomes in higher eukaryotes, the major (U2-type) and minor (U12-type) spliceosomes. Plants, like other eukaryotes, contain U1, U2, U4, U5, and U6 snRNAs in the major spliceosome, while the minor spliceosome contains U11, U12, U4atac, U5, and U6atac (Lorkovic et al., 2005; Ru et al., 2008). These snRNAs form complexes with proteins, including some proteins specific to each U-RNA, to form snRNPs. In animals, where the snRNP complexes have been extensively studied, all the major snRNPs except U6 snRNP have a core of seven Sm proteins and additional UsnRNP-specific proteins (Valadkhan and Jaladat, 2010). The U6snRNP has a different set of seven proteins termed LSm plus only one U6snRNP-specific protein. The minor spliceosome snRNPs also have Sm proteins and snRNP-specific proteins but are not as well studied as the major spliceosome proteins (Ru et al., 2008). Only one Sm protein (SAD1/LSm5) that is necessary for abscisic acid sensitivity and drought tolerance has been characterized in Arabidopsis (Xiong et al., 2001), but analysis of the Arabidopsis predicted proteome revealed that all the major snRNP proteins are conserved in Arabidopsis (Wang and Brendel, 2004). The counterparts of three U1snRNP-specific proteins (U1-A, U1-C, and U1-70K) identified in animals and yeast have been characterized in plants. U1-A in Arabidopsis has been shown to bind U1snRNA (Simpson et al., 1995), while Arabidopsis U1-70K has an RNA-binding domain and an Arg/Ser-rich (RS) domain responsible for interaction with SR proteins (see below; Golovkin and Reddy, 1996, 1998). An Arabidopsis U11 snRNP-specific 35K protein identified as a binding partner of an SR protein has been shown to incorporate into monomeric U11 snRNPs and U11/U12 di-snRNP complexes. These interactions, together with its sequence similarity to U1 70K in the U1snRNP, suggest that the U11-35K protein functions in the splicing of minor AT:AC introns (Lorkovic et al., 2005). In addition, other genes coding for minor snRNP proteins have been identified in both monocot and dicot genomes, indicating a high evolutionary conservation of minor spliceosome components and splicing function (Lorkovic et al., 2005).

NON-SnRNP PROTEINS

The non-snRNP proteins include splicing factors involved in splice site selection, U-snRNP-associated proteins, SR proteins, hnRNPs, and other splicing regulators. U2AF, composed of two subunits, U2AF⁶⁵ and U2AF³⁵, is a non-snRNP protein involved in 3′ splice site selection. Both subunits, each with two isoforms, have been identified in plants (Wang and Brendel, 2006). U2AF³⁵ has been shown to interact with SR-like proteins SR45 and SR45a, which also interact with U1-70K, suggesting a role for these proteins in bridging of the 5′ splice site and 3′ splice site selection complexes (Tanabe et al., 2009; Reddy et al., 2011). Two subunits of the Cap Binding Complex (CBC20 and CBC80) and SERRATE, a zinc finger protein, regulate pre-mRNA splicing and pri-miRNA processing (Laubinger et al., 2008; Raczynska et al., 2010). MOS4, a protein involved in plant immunity, is thought to be a component of the spliceosome (Palma et al., 2007). An interesting group of splicing-related proteins is the RS domain-containing cyclophilins, termed atCypRS64 and atCypRS92, which were found in a yeast two-hybrid screen with Arabidopsis SR proteins. These proteins contain a prolyl-peptidyl isomerase domain important for protein structure and modification and have been shown to interact with SR proteins in vitro and in vivo, suggesting that they might regulate spliceosome assembly (Lorković et al., 2004). Another multidomain cyclophilin, atCyp59 (prolyl-peptidyl isomerase, RRM zinc knuckle, and RS/RD domain), has been found to interact with SR proteins as well as with the C-terminal domain of RNA polymerase II and was suggested to connect splicing and transcription (Gullerova et al., 2006).

SR PROTEINS

SR proteins have been redefined recently in plants (Barta et al., 2010) based on sequence features, including the number and location of RRMs and the RS domain. In plants, proteins with one or two N-terminal RRMs followed by a downstream RS domain of at least 50 amino acids and a minimum of 20% RS or SR dipeptides are considered as SR proteins (Barta et al., 2010). The RRM domain recognizes and binds to regulatory elements in pre-mRNAs, and the RS domain is involved in interactions with other proteins. SR proteins are important splicing regulators in both animals and plants. In animals, SRs are known to perform many other diverse functions in the nucleus as well as in the cytoplasm, including mRNA transport, localization, translation, and decay, as well as in genome stability and microRNA biogenesis (Twyffels et al., 2011). Interestingly, some SRs in animals are present at the cell surface and exhibit carbohydrate-binding activity (Hatakeyama et al., 2009). The SR proteins are major regulators of CS and AS (Barta et al., 2008; Long and Caceres, 2009; Reddy and Ali, 2011). Plants have more SR proteins (18 in Arabidopsis and 22 in rice) as compared with mammals (12 in humans; Barta et al., 2010). Plant SR proteins are grouped into six subfamilies: three subfamilies are orthologous to animal SRs, whereas 10 AtSRs in three subfamilies are plant specific, suggesting some differences in splicing regulation between plants and animals (Barta et al., 2010). The SR subfamily, which has two RRMs (the second RRM having a conserved SWQDLKD motif) followed by the RS domain, is orthologous to the mammalian SRSF1/SF2/ASF family. The RSZ subfamily members, with one RRM and one RS domain separated by a zinc (Zn) knuckle, are orthologous to the mammalian SRSF7/9G8 subfamily. The SC subfamily has one RRM and one RS domain and is orthologous to the SRSF2/SC35 subfamily. The SCL subfamily, while similar to the SRSF2 subfamily, has an RRM with an N-terminal charged extension. The RS2Z subfamily members have two Zn knuckles and an additional SP-rich region following the RS domain. Finally, the RS subfamily has two RRMs, but without the conserved SWQDLKD motif, followed by the RS domain. Based on the new definition, two plant proteins (SR45 and SR45a) previously classified as SR proteins are no longer included in the SR family. SR45 has two RS domains, one preceding and one following the RRM domain, with sequence similarity to an exon junction complex protein, RNPS1. SR45a is not related to SR45 at the sequence level, as the name implies, and is a homolog of metazoan tra-2. It is interesting that 14 of the 18 Arabidopsis SR proteins are alternatively spliced, as are SR45 and SR45a (Palusa et al., 2007; Tanabe et al., 2009).

OTHER PLANT RBPS INVOLVED IN SPLICING

An hnRNP-like protein that binds to U-rich or AU-rich sequences, called UBP1, with three RRMs enhances the splicing of suboptimal introns and enhances the steady-state levels of mRNA by preventing mRNA degradation (Lambermon et al., 2000). Two plant-specific proteins (UBA1 and UBA2) that interact with UBP1 also contain an RRM and increase mRNA levels but do not enhance splicing (Lambermon et al., 2002). RRM-containing RBPs with a C-terminal Gly-rich region (GRPs) are also involved in different aspects of RNA metabolism, including mRNA export and AS. In Arabidopsis, there are eight GRPs, and some of them (GRP7 and GRP8) are involved in regulating circadian rhythm by autoregulation and cross-regulation of AS of their pre-mRNAs coupled to nonsense-mediated decay (Schöning et al., 2008). Similarly, three polypyrimidine tract-binding proteins that belong to the hnRNP family are involved in autoregulation and cross-regulation of AS of their pre-mRNAs coupled to nonsense-mediated decay (Stauffer et al., 2010).

EXPERIMENTAL APPROACHES USED TO STUDY THE LOCALIZATION AND DYNAMICS OF SPLICING REGULATORS

Early approaches to localize splicing regulators used fixed cells or tissues with antibodies against mammalian splicing-related proteins using indirect immunofluorescence microscopy and electron microscopy (Ali and Reddy, 2008b), which provided static images of a protein at a particular point in time. The discovery of fluorescent proteins (FPs) has opened many new innovative methods to detect the localization and movement of individual proteins and RNA molecules and to monitor protein-protein, protein-RNA, and RNA-RNA interactions in living cells (Chudakov et al., 2010; Urbinati and Long, 2011). These methods are providing an unprecedented view of protein/RNA localization and their movement spatially and temporally (Larson et al., 2011; Urbinati and Long, 2011). Translational fusions of the protein of interest with GFP, its derivatives red (RFP), yellow (YFP), and cyan (CFP) FPs, or other FPs with different excitation and emission spectra are well suited for investigating the temporal and spatial distribution of proteins and have permitted the visualization of multiple proteins simultaneously (for review, see Chudakov et al., 2010).

Fluorescence recovery after photobleaching (FRAP) is an imaging method where fluorophores from a FP-tagged protein are inactivated through bleaching of a defined area with intense laser pulses. Since bleaching permanently destroys fluorescence, the recovery of fluorescence to the bleached area depends on the mobility of the labeled protein from the unbleached area and is monitored over time for estimation of the dynamics of proteins (i.e. diffusion parameters or trafficking). FRAP provides information about how much of the total available protein is free to diffuse in live cells versus how much is restricted due to its association with complexes or partition into compartments or microdomains. Fluorescence loss in photobleaching (FLIP), a technique complementary to FRAP, also permits the analysis of the intracellular mobility of proteins. In FLIP analysis, an area of fluorescence is repeatedly bleached, and the progressive loss of fluorescence intensity in the unbleached region indicates the mobility of the labeled protein, providing the opportunity to monitor the direction of movement (for review, see Wang et al., 2008). Wild-type GFP exists in two forms, which lead to a minor and a major absorbance peak (395 and 475 nm, respectively). By activating the protein with intense illumination of 390 to 415 nm, the absorbance of the minor peak will increase dramatically and can be monitored. Photoactivatable GFP has only a negligible minor excitation peak at 475 nm; consequently, upon activation, this peak has a considerably greater proportional increase when compared with wild-type GFP, leading to a higher contrast (Patterson and Lippincott-Schwartz, 2002). This feature can be used for following activated proteins over time using time-lapse imaging and provides a simple and more direct alternative to standard molecular tracking procedures like FRAP and FLIP. A reversible photoswitchable fluorescent reporter protein, which allows repeated switching between a fluorescent and a nonfluorescent state, has been used to monitor intracellular protein trafficking. For instance, a variant of the photoswitchable fluorescent protein DRONPA, designed for use in transgenic Arabidopsis plants, was fused to the Arabidopsis RBP AtGRP7 under the control of the endogenous AtGRP7 promoter (Lummer et al., 2011). Fluorescence microscopy showed that AtGRP7 is a nucleocytoplasmic shuttling protein.

To study the in vivo interaction/association of proteins, two methods involving a two-component system are feasible: bimolecular fluorescence complementation (BiFC)- and fluorescence resonance energy transfer (FRET)-based reporters. In BiFC, two halves of an FP are separately fused to two putative interacting proteins (Kerppola, 2009). When the two proteins interact, the two halves of the FP come close enough for fluorescence to be established. Because of the irreversible formation of the complex, weak or transient interactions can also be monitored. Reconstitution of fluorescence is also possible if the two proteins are in close proximity in a complex and do not interact directly. Interestingly, this method can also be adapted to monitor protein-RNA interactions (Rackham and Brown, 2004). FRET can be used to measure the proximity of two fluorophores, usually coupled to the proteins of interest. The fluorophore of one protein has an emission corresponding to the excitation of the other. Within 10 nm, a transfer of energy from the excited fluorophore (the donor) to the other fluorophore (the acceptor) can occur. The acceptor molecule enters an excited state from which it emits a longer wavelength light. By monitoring the donor and acceptor emissions, the efficiency of the resonance energy transfer and consequently the relative distance between the molecules can be calculated, with even higher accuracy than the optical resolution of a standard light microscope (for review, see Wang et al., 2008). RNA-protein interactions were also visualized in live cells using BiFC and FRET (Urbinati and Long, 2011).

Fluorescence correlation spectroscopy monitors fluctuations of the fluorescent signal in a noninvasive fashion. It relies on changes in the specific signal provided by fluorescent particles to analyze their motions and interactions. In fluorescence correlation spectroscopy, a laser beam is focused on an area of fluorescence. Fluctuations in fluorescent signals are then recorded over time to determine the diffusion coefficients and binding constants of the protein (Krichevsky and Bonnet, 2002; Malchus and Weiss, 2010). During the last decade, FPs tagged to splicing regulatory proteins have been widely used in plants to monitor their location and dynamics to address issues about the spatiotemporal dynamics of splicing, which is discussed below.

LOCALIZATION OF SPLICING REGULATORS

In plant cells, only a limited number of splicing regulators have been investigated regarding their cellular localization, and most of them belong to the SR proteins, snRNP proteins, or RS-containing cyclophilins (Lorković and Barta, 2004; Ali and Reddy, 2008b). The nuclear space is compartmentalized by nuclear bodies, which are dynamic but relatively stable structures composed of proteins and RNAs without delineating membranes. Their morphology is dependent on cell differentiation and metabolic status. In animals, these include structures like the nucleoli, nuclear speckles, Cajal bodies, histone locus bodies, paraspeckles, and others (Misteli and Spector, 2011). Splicing regulators are detected in nucleoplasm, nuclear speckles, Cajal bodies, and sometimes the nucleolus. In plants, the best-investigated nuclear bodies are the nucleolus, photobodies, which contain phytochrome and proteins involved in photomorphogenesis, and nuclear speckles. As the nucleolus and photobodies are described in other Update articles in this Focus issue, we will focus on the occurrence and the dynamic aspects of nuclear speckles. Nuclear speckles are located in the interchromatin space and were detected as a storage place for splicing factors (Fang et al., 2004; Spector and Lamond, 2011). In animal tissues, nuclear speckles have also been termed SC35 bodies, as many SR proteins accumulate in these nuclear bodies. However, nuclear speckles also contain snRNPs, non-snRNP splicing proteins, transcription factors, and 3′ end processing factors. Speckles are often observed near active transcription sites, and pre-mRNAs were detected outside the speckles in fibrillar structures (for review, see Spector and Lamond, 2011). Hence, nuclear speckles are viewed as storage and assembly areas that procure splicing factors to active transcription sites. In line with this hypothesis, actively transcribing cells have smaller and more speckles, whereas inhibition of transcription or splicing leads to an accumulation of splicing factors and larger speckles. Like speckles, paraspeckles are subnuclear bodies found in interchromatin spaces and contain long non-protein-coding RNAs and novel proteins (Spector and Lamond, 2011); however, they have not been described yet in plant cells.

The complex compartmentalization of plant nuclei is now an active field of research. Genomic and proteomic approaches are being used to investigate the nucleolus and the upcoming new nuclear dicing bodies. In plants, the RNA processing machinery was first described in experiments using antibodies to U2snRNP-specific proteins, which were detected diffusely in the nucleoplasm and in Cajal bodies (Beven et al., 1995). SRs representing each group of SR proteins in Arabidopsis have been shown to display a characteristic distribution pattern of high concentration in speckles and diffuse distribution in the nucleoplasm (Fang et al., 2004; Lorković and Barta, 2004; Lorković et al., 2004, 2008; Tillemans et al., 2005, 2006; Ali and Reddy, 2006). The speckle localization of proteins involved in pre-mRNA splicing is so characteristic that it is used for a diagnostic tool for splicing proteins (Spector, 2001; Lorković and Barta, 2004). Recently, a protein similar to transportin-SR (also called MOS14 in Arabidopsis) that is required for plant immunity was shown to be necessary for transporting SR proteins into the nucleus (Xu et al., 2011). Studies using GFP-labeled RRM and RS domains of plant proteins suggest that RS domains of the plant SR proteins seem to contain the nuclear localization and speckle targeting/retention signals, as the GFP-RRM fusions of SR45, AtRS31, and AtRSZ22 were localized to the nucleoplasm or cytoplasm (Tillemans et al., 2005; Ali and Reddy, 2006). Using fluorescently labeled Arabidopsis SR proteins, several groups have demonstrated that plant nuclei have nuclear speckles whose size and shape are dependent on cell type, metabolic state, and transcriptional activity (Fang et al., 2004; Lorković and Barta, 2004; Lorković et al., 2004, 2008; Tillemans et al., 2005, 2006; Ali and Reddy, 2006). An interesting phenomenon was observed when colocalization studies of Arabidopsis SR proteins were performed in plant protoplasts (Lorković et al., 2008). While representatives of all six subfamilies of plant SRs localize to speckles, it could be demonstrated that several of the SR proteins do not or only partly colocalize to the same nuclear area. Interestingly, the two families containing a Zn knuckle motif did not colocalize with AtSR34 (an SF2/ASF ortholog). Figure 1 shows an example of RS2Z33-GFP cotransfected with SR34-RFP, where the projection image almost does not have overlapping speckles (Lorković et al., 2008). Furthermore, AtSR34 only partly colocalized with AtSC35, although both proteins are used as genuine markers for nuclear speckles. In some cases, colocalization results were found to differ compared with other studies using different plants and conditions (Tillemans et al., 2005), suggesting that colocalization of SR proteins may depend on cell type and particular conditions. In addition, colocalization did not always correlate with in vitro protein-protein interaction studies (Lorković et al., 2008); therefore, these data are not straightforward to interpret. Other plant spliceosomal proteins (U1-70K, U2AF^35a, U2AF^35b) are also localized to speckles in the nucleus (Wang and Brendel, 2006; Ali et al., 2008). Speckle components have been shown to shuttle continuously between speckles and nucleoplasm (Ali and Reddy, 2008b; Rausin et al., 2010). Their biogenesis is proposed to occur through self-assembly processes whereby a transient binding interaction to macromolecules will determine the size and shape of the speckles (Dundr and Misteli, 2010). Recently, in animals, large scaffolding proteins have been implicated in the organization of nuclear speckles (Sharma et al., 2010). The colocalization data might indicate that binding interactions between some of the SR proteins are less strong than their self-assembly ability, resulting in different speckled areas in the nucleus. These data also indicate that nuclear speckles might exhibit different splicing factor compositions dependent on the different states of the cell and the needs for transcription/splicing of particular genes.

Figure 1.

Plant SR proteins do not always colocalize in the same speckles. Maximum intensity projection images show RS2Z33-GFP and SR34-RFP transiently coexpressed in tobacco protoplasts. The merged image shows no or very little overlap in the speckled areas (Lorković et al., 2008).

DYNAMICS OF SPLICING REGULATORS

Speckles and their components are dynamic structures. They change in size, in shape, and in number. These properties of a given nuclear body can be different in different cell types or at different times in the same cell type. At the same time, they are structurally stable, individual nuclear bodies persist during the entire interphase (entry to G1 to exit from G2), but there is an exchange of a large number of the major nuclear body components with the surrounding nucleoplasm (Dundr and Misteli, 2001). The morphological appearance of a nuclear body is determined by the components in the body and their interactions. Individual speckle components are able to shuttle in and out of the speckles. It was proposed that the size, and possibly the shape, of nuclear bodies is likely determined by the balance of the on rate relative to the off rate of its components, with an increase in on rate or a drop in off rate leading to an increase in size and a decrease in on rate or an increase in off rate leading to shrinkage (Dundr and Misteli, 2001). The intranuclear mobility of RSZ22-GFP was investigated using FRAP and FLIP (Tillemans et al., 2006). These experiments showed that this SR splicing factor shuttles rapidly between subnuclear compartments, including the nucleolus. A sensitive shuttling assay using FLIP and nuclear export inhibitors with cells transiently expressing FP fusions strongly suggested that RSZ22 is a nucleocytoplasmic shuttling SR splicing factor (Tillemans et al., 2006). The dynamics of RSZ22 was recently studied in stable transformants enabling tissue-specific expression of the transgene (Rausin et al., 2010). The results demonstrated that RSZ22 is a bona fide nucleocytoplasmic shuttling SR protein. RSZ22 mutants with mutated RNP1 or Zn knuckle domains were able to shuttle and localize to speckles, suggesting that neither RNP1 nor Zn knuckle motifs are key to nuclear localization and speckle-like organization. However, FRET studies established that molecular interactions between splicing factors were strongly destabilized, although not entirely inhibited, with Zn knuckle and RNP1 mutants, suggesting that both RNA-binding domains might be needed either for direct protein-protein interactions or for mRNA-mediated interactions independent of splicing factor preassembly.

Studies using GFP-SR45 showed its localization to speckles and nucleoplasm, and the speckles exhibited intranuclear movements and changes in morphology (Ali et al., 2003; Zhang and Mount, 2009). Later, using FRAP, the effective diffusion coefficient (D_eff), a measure of the apparent mobility of a protein, for GFP-SR45 (Fig. 2) was determined to be 1.01 μm² s⁻¹ in the nucleoplasm and 0.38 μm² s⁻¹ in the speckles (Ali and Reddy, 2006). Using GFP-labeled U1-70K, FRAP analyses were performed on speckles and the nucleoplasmic regions of the nucleus (Ali and Reddy, 2006). It was shown that U1-70K moves in the nucleus very rapidly, on a millisecond scale. The expected D_eff value of uncomplexed GFP/U1-70K based on its size is approximately 12 (in nucleoplasm) to 31 (in speckles) times greater than the experimental values, suggesting that U1-70K interacts with other nuclear components or is in a larger snRNP complex (Ali and Reddy, 2006).

Figure 2.

Mobility analysis of SR45 and SRp34 in control and ATP-depleted cells using FRAP. A speckle or a defined nucleoplasmic area in control or ATP-depleted cells expressing either GFP-SR45 (A) or SRp34-YFP (B) was bleached, and the recovery of fluorescence was quantified for 80 s. The mobility of these splicing factors was different in nucleoplasm as compared with speckles, and the mobility required ATP, as ATP depletion reduced the movement in both nucleoplasm and speckles (Ali and Reddy, 2006). D_eff values (in μm² s⁻¹) are as follows: for GFP-SR45, in nucleoplasm, 1.01, −ATP 0.06; in speckle, 0.38, −ATP 0.05; for SR1/SRp34-YFP, in nucleoplasm, 1.34; −ATP 0.07; in speckle, 0.88, −ATP 0.04.

Time-lapse microscopy studies of the mobility of speckles in different plant cells and under different physiological conditions showed that SR speckles in plant nuclei display limited movements in a constrained area (Ali and Reddy, 2008b, and refs. therein). The speckles are possibly loosely anchored to less mobile nuclear components or restrained from mobility by physical barriers such as chromatin. Besides these restricted movements, plant SR speckles also fuse with each other and bud off from speckles (Ali et al., 2003; Ali and Reddy, 2008b). Studies with animal cells have shown that the distribution pattern of SR proteins and other splicing factors changes with the cell cycle (for review, see Dundr and Misteli, 2010; Spector and Lamond, 2011). Changes in the distribution pattern of SR proteins during the cell cycle were also observed in plant cells (Fang et al., 2004). SR proteins became more diffuse, so that during metaphase they were distributed throughout the cytoplasm, with speckles reforming in late telophase.

REGULATION OF THE LOCALIZATION AND DYNAMICS OF SPLICING REGULATORS

The localization and dynamics of these splicing regulators are regulated by phosphorylation status, transcriptional activity of the cell, stages of the cell life cycle, such as M phase versus interphase, cell type, developmental stage, stresses imposed on plants, and hormone signals. A study of SR-FP fusion proteins found that meristematic cells or otherwise rapidly growing cells had the highest number of small speckles, with most of the FP fusion proteins found in a diffuse nucleoplasmic pool, while highly differentiated cells (e.g. leaf mesophyll or epidermal cells) showed larger speckles with fewer FP fusion proteins in the nucleoplasm (Lorković et al., 2008).

Phosphorylation and Dephosphorylation

SR proteins are phosphoproteins, with the RS domains being extensively phosphorylated in vivo (Reddy, 2007; Barta et al., 2008). In plants Clk/LAMMER type, SRPKs, and mitogen-activated protein kinases phosphorylate one or more SR proteins. Phosphorylation is thought to influence subcellular localization, protein-protein interactions, RNA-protein interactions, and splicing activities. Inside speckles, many SR proteins exist in a hypophosphorylated state; when phosphorylated by speckle-resident SR-protein kinases, the hyperphosphorylated SR proteins dissociate from speckles and are competent to participate in the splicing reaction (Dundr and Misteli, 2010). The SR proteins are dephosphorylated in the splicing process, and their affinity for speckles is reestablished. Early studies have shown that inhibition of phosphorylation in plant cells sequestered SR and SR-like proteins to large speckles and other altered localization patterns (Fig. 3; Ali et al., 2003; Tillemans et al., 2005). Not only is the localization pattern different but also the mobility is altered. FRAP and FLIP analyses with AtSR45 and AtSR34 demonstrated that their mobility is reduced by the inhibition of protein phosphorylation (Ali and Reddy, 2006). The use of staurosporine to inhibit protein kinases resulted in a significant decrease in the mobility of AtSR45 and AtSR34. The D_eff values of AtSR45 were reduced four times, whereas the D_eff values of AtSR34 were reduced 14 times. Their mobile fractions were also significantly reduced. The enlargement of speckles in the presence of staurosporine suggests that the release of SR proteins from speckles requires their phosphorylation (Ali et al., 2003; Tillemans et al., 2005). Studies by Tillemans et al. (2006) suggested that the motility of AtRSZ22 is mainly phosphorylation dependent and that the phosphorylation/dephosphorylation cycle of the RS domain may influence the subnuclear distribution and dynamics of RSZp22 into or out of the nucleolus. AtRSZ22-GFP was shown to concentrate in the nucleolus upon phosphorylation inhibition, while other tested AtSRs did not (Tillemans et al., 2006). Later studies in root and pollen cells of stably transformed plants confirmed the accumulation of AtRSZ22-GFP within nucleoli upon phosphorylation inhibition (Rausin et al., 2010). In transiently expressing root cells, AtRS31-GFP, unlike other tested SRs, accumulated into concave cap-like structures surrounding regions devoid of fluorescence corresponding to the nucleoli following treatment with staurosporine. This occurred in all cell types of the primary root, from meristematic cells to trichoblasts, and there was a similar redistribution in transiently expressing Arabidopsis leaf cells. The perinucleolar localization of AtRS31 proteins was shown to be reversible, because the AtRS31-GFP relocalized into nucleoplasmic speckles upon the return to physiological conditions. A cyclin-dependent kinase, CDKC2, colocalizes with AtSR34 and other spliceosomal components, and this association is dependent upon the CDKC2 kinase activity and transcriptional status of the cells (Kitsios et al., 2008), suggesting a role for this kinase in the dynamics of spliceosomal proteins. Overexpression of a LAMMER kinase in Arabidopsis, which colocalizes with AtSRp34, altered the splicing pattern of several genes and the plants showed developmental abnormalities, suggesting a role of phosphorylation in modulating AS (Savaldi-Goldstein et al., 2003).

Figure 3.

Redistribution of an SR-related splicing regulator (SR45) in the nucleus in response to temperature stress and inhibition of transcription or phosphorylation. Arabidopsis plants expressing GFP-SR45 show its localization to characteristic speckles and nucleoplasm. Heat treatment (42°C for 24 h) resulted in the accumulation of SR45 into large irregularly shaped speckles, whereas in the cold (4°C for 12 h), speckles disappeared. Inhibition of transcription (actinomycin D for 1 h) resulted in the accumulation of SR45 into a few very large round speckles. Inhibition of protein phosphorylation (staurosporine for 1 h) also resulted in the appearance of large irregularly shaped speckles similar to heat (Ali et al., 2003).

ATP

FRAP and FLIP analyses with AtSR45 and AtSR34 in ATP-depleted cells showed a dramatic reduction in the mobile fraction of both proteins in the nucleoplasm as well as in speckles (Fig. 2), and this reduction in mobility was restored by ATP (Ali and Reddy, 2006), suggesting that the association/dissociation of these SRs with other SRs/spliceosomal proteins is dependent on ATP. The retarded mobility in ATP-depleted cells is not likely due to global nuclear rearrangement, as ATP depletion had no substantial effect on the mobility of GFP alone and of a nucleus-localized NLS-GFP-GUS fusion that is twice the size of SR45. In further support that ATP depletion does not result from nonspecific general perturbation of the cellular environment, FRAP analysis with the deletion mutants of SR45 exhibited different localization and kinetic properties in response to ATP depletion (Ali and Reddy, 2006). In other studies, ATP depletion resulted in an accumulation of AtRSZ22-GFP within nucleoli, as had been demonstrated in phosphorylation inhibition, suggesting that the absence of RSZp22 mobility upon ATP depletion is linked to its phosphorylation state (Tillemans et al., 2006; Rausin et al., 2010). In animals, ATP depletion did not change the mobility of SR proteins (Phair and Misteli, 2000).

Transcription

Several studies have shown that transcription sites are often found at the periphery of speckles and that the disassembly of speckles affects the coordination between transcription and pre-mRNA splicing (for review, see Lorković et al., 2008; Spector and Lamond, 2011). Generally, inhibition of transcription causes a decrease in the number of nuclear speckles and the redistribution of splicing factors into larger storage bodies (Ali et al., 2003; Tillemans et al., 2005; Kitsios et al., 2008). After release of the inhibition of transcription, nuclear speckles can form de novo, speckles expand from a condensed state and take on their typical irregular shape, and new speckles form (for review, see Dundr and Misteli, 2010). Inhibition of transcription sequestered AtSR45 to large speckles and other altered localization patterns (Fig. 3; Ali et al., 2003; Tillemans et al., 2005). Studies of AtSR34, AtSR30, and AtSCL33 also showed increased accumulation of splicing factors in speckles, eventually leading to increased speckle area and the cessation of all kinds of speckle movements, including budding off and peripheral movements (Ali and Reddy, 2008b, and refs. therein). The mobility of the SR and SR-like proteins is also affected by transcription. FRAP analysis revealed that the inhibition of transcription by actinomycin D resulted in a significant decrease in the mobility of AtSR45 and AtSR34. The mobile fraction of AtSR45 was also significantly reduced by actinomycin D treatment but not that of AtSR34. Treatment of plants expressing AtRS31 with okadaic acid and α-amanitin did not result in AtRS31 accumulation around the nucleolus, as with staurosporine treatment, but instead resulted in accumulation into static large speckles (Tillemans et al., 2006).

Stresses

Stresses have been shown to affect the splicing of pre-mRNAs from genes involved in stress responses, including the splicing of many SR genes (Iida et al., 2004; Reddy, 2007; Ali and Reddy, 2008a). The AS patterns of SR proteins have been shown to be affected by heat, cold, salt, drought, and osmotic stress (Palusa et al., 2007; Ali and Reddy, 2008a). Heat has also been shown to affect the distribution of AtSR45. Heat treatment redistributed AtSR45 into enlarged, irregularly shaped compartments, and cold relocalized AtSR45 mostly to the nucleoplasmic pool (Fig. 3; Ali et al., 2003). Studies have shown that heat shock leads to an accumulation of various RBPs and several SR proteins to form de novo a morphologically distinct nuclear stress body (for review, see Dundr and Misteli, 2010). Interestingly, the enlargement of speckles due to heat shock was inhibited by a phosphatase inhibitor, implying that heat shock-induced relocalization involves the dephosphorylation of SR proteins (Ali et al., 2003). As heat and cold also changed the AS pattern of pre-mRNAs of several SR genes, it is possible that regulation of the AS pattern of these genes is related to a change in the subnuclear reorganization of SR splicing factors (Ali and Reddy, 2008b). As with SR45 speckles, the size of nuclear bodies containing one of the protein kinases (CDKC2) implicated in regulating the dynamics of spliceosomal proteins is increased upon heat treatment, whereas cold treatment results in the disappearance of CDKC2-containing speckles (Kitsios et al., 2008).

mRNA LOCALIZATION

Splicing regulators often bind to their target RNA directly or are part of splicing complexes. Following the life cycle of an individual RNA molecule from transcription to processing to localization on both the RNA and protein levels would give insights into the actual parameters of RNA splicing and its regulation. Furthermore, it is of particular importance to determine the fate of alternatively spliced transcripts, as AS is a major regulator of gene regulation in plants (Reddy, 2007, Barta et al., 2011). The distribution and spatiotemporal dynamics of mRNA molecules have been an active field of research in the past few years, including experiments performed using neuronal cells, fibroblasts, Drosophila melanogaster oocytes, and Saccharomyces cerevisiae cells (Park et al., 2010). However, data on compartmentalized RNA localization in planta are scarce. Using in situ hybridization, the distribution and, hence, the intercellular transport of transcripts through plasmodesmata has been investigated in plant cells (Lucas et al., 1995; Sambade et al., 2008; Hyun et al., 2011). However, this technique gives a stagnant picture of the subcellular mRNA distribution. Recent technical advances allow the exact localization and also the tracking of RNA molecules in real time in their natural environment, the living cell (Christensen et al., 2010). An indirect approach for monitoring RNA targets in living cells is the tracing of a fluorescently labeled RBP. The disadvantage of this procedure is that it is not certain whether the correct RNA species is monitored, as the fluorescent protein might bind other RNAs as well. Other approaches alter the RNA target itself by introducing binding motifs for fluorescently labeled RBPs (e.g. MS2-FP system, or λN₂₂) or by inserting dye-binding aptamers into the sequence (for discussion, see Christensen et al., 2010). For localization and tracing purposes, these approaches are not recommended, since the dynamics of the RNA molecules could be changed due to the fluorescent foreign proteins hindering the trafficking or due to alterations in the naturally occurring secondary and tertiary structures of the RNA molecule. In addition, in this case, the injection of directly labeled RNA transcripts is problematic, because these molecules would not proceed via their natural processing steps (e.g. cotranscriptional splicing, polyadenylation, capping, export, etc.) and hence localization and complex formation might be different. Nonetheless, the aforementioned in vivo techniques for monitoring RNA targets are powerful tools that have been used successfully for specialized purposes (for review, see Christensen et al., 2010). Therefore, to investigate RNA in its natural context, it is important to image endogenous, and therefore not altered, RNA transcripts. To accomplish this, hybridization probes in various forms have been used. The most promising ones are termed molecular beacons, which carry a fluorophore and a quencher group at the 3′ and 5′ end, respectively (Tyagi and Kramer, 1996). These probes have a special design allowing them to assume a secondary structure, which blocks fluorescence emission in the closed conformation. Upon hybridization to the target, the probes open and the fluorescence intensity rises up to 100 times above the background level. Since plant cells are known for high autofluorescence originating from plastids, the low background levels of molecular beacons and also the free choice of fluorophoric groups allow the usage of them in living plant cells. One main concern in using hybridization probes in living cells is their delivery method. Microinjection has been used before, but this technique is time consuming and very laborious in plants. One laboratory reports the transfection of molecular beacons into plant suspension cells via electroporation. Treated cells show a sufficient survival rate and transfection efficiencies that are adequate for confocal fluorescence microscopy and single particle tracing (J. Göhring and A. Barta, unpublished data).

CONCLUSION AND PERSPECTIVE

The organization of plant nuclei, once thought to be static, has been shown to be highly dynamic, with the components of splicing in a constant flux between various subnuclear domains. The mechanisms that control the mobility of splicing regulators and the morphology of speckles in plants are emerging (Ali and Reddy, 2008b). Plant SR proteins and other splicing regulators constantly exchange between speckles and the nucleoplasm under a steady-state equilibrium state. A balance of the influx and efflux of SR proteins into and out of the speckles determines the morphology and size of speckles. An enhanced influx and/or decreased efflux increases the size of the SR speckles, and a decrease in the influx and/or increased efflux decreases the size or dismantles the speckles. Inhibition of transcription leads to a decline of pre-mRNAs and, hence, splicing. The SR protein molecules are recycled for the next round of splicing, and in the absence of any pre-mRNA substrate, SR proteins would tend to stay in the speckles, eventually leading to an increased speckle size. Phosphorylated SR proteins leave the speckle and participate in splicing, and when dephosphorylated, they enter the speckles. In the absence of phosphorylation, they are not competent for recruitment to the splicing sites and remain in the speckles. The dynamics of only a few plant splicing regulators have been studied to date. However, with the increased use of in vivo imaging methods combined with high-throughput analyses, a rapid expansion in our understanding of how plant pre-mRNA splicing is spatially and temporally organized and how the interactions of various splicing components are regulated should be forthcoming.

Many fundamental questions pertinent to speckles in plants have yet to be answered. What is the composition of speckles? What is the meaning of different types of speckles, and when do they occur? Are there subpopulations of speckles that differ in their global composition? Does the composition of speckles differ depending on the physiological status of the cells or cell type? To address these questions, technological advances are needed to isolate speckles either biochemically or microscopically, which can then be analyzed using the available highly sensitive proteomic methods. Recently, single lipid droplets from cytoplasm were isolated by combining visualization and microphase separation to analyze their composition (Horn et al., 2011). A similar approach might be useful in analyzing the composition of speckles. Also, we do not have a clear understanding of the relationship between speckle-like patterns observed with various other proteins, including transcription factors (Tao et al., 2005; Chen, 2008), light receptors (Chen, 2008), and microRNA processing proteins (Fang and Spector, 2007), in plant cells. The biogenesis of speckles in plants is also not known. What initiates the formation of a speckle? Is there a small set of proteins that function as scaffolding proteins and initiate speckle formation? Although some plant SRs are found to be shuttling proteins, their cytosolic functions are not known.

AS is a prominent feature of plant gene regulation and provokes questions about the fate of alternative transcripts, their localization, and their dynamics. One promising approach to follow RNA-protein particles is to label specific RNAs and determine their fate during the maturation steps. Future hopes lie specifically on single particle tracing with molecular beacons combined with high-resolution techniques, which should provide insight into the diffusion constants of different subpopulations of mRNA molecules within the nucleus and the cytoplasm (Christensen et al., 2010; J. Göhring and A. Barta, unpublished data). Further studies that address these questions should lead to a better understanding of splicing and gene regulation.