The signal recognition particle (SRP) is a ribonucleoprotein machine that controls the translation and intracellular sorting of membrane and secreted proteins. The SRP contains a core RNA subunit with which six proteins are assembled. Recent work in both yeast and mammalian cells has identified the nucleolus as a possible initial site of SRP assembly. In the present study, SRP RNA and protein components were identified in the extrachromosomal, amplified nucleoli of Xenopus laevis oocytes. Fluorescent SRP RNA microinjected into the oocyte nucleus became specifically localized in the nucleoli, and endogenous SRP RNA was also detected in oocyte nucleoli by RNA in situ hybridization. An initial step in the assembly of SRP involves the binding of the SRP19 protein to SRP RNA. When green fluorescent protein (GFP)-tagged SRP19 protein was injected into the oocyte cytoplasm it was imported into the nucleus and became concentrated in the amplified nucleoli. After visiting the amplified nucleoli, GFP-tagged SRP19 protein was detected in the cytoplasm in a ribonucleoprotein complex, having a sedimentation coefficient characteristic of the SRP. These results suggest that the amplified nucleoli of Xenopus oocytes produce maternal stores not only of ribosomes, the classical product of nucleoli, but also of SRP, presumably as a global developmental strategy for stockpiling translational machinery for early embryogenesis.
The signal recognition particle (SRP) is a ribonucleoprotein complex that directs the spatially correct synthesis and traffic of membrane and secreted proteins (Walter and Johnson, 1994; Walter et al., 2000). In vertebrates, the SRP consists of a ∼300 nucleotide-long RNA and six proteins (Walter and Johnson, 1994). The structure and function of the SRP is currently understood in considerable detail (Keenan et al., 1998; Batey et al., 2000; Walter et al., 2000; Weichenrieder et al., 2000; Batey et al., 2001; Beckmann et al., 2001; Keenan et al., 2001; Weichenrieder et al., 2001; Wild et al., 2001; Hainzl et al., 2002; Oubridge et al., 2002; Pool et al., 2002; Nagai et al., 2003), but less is known about how this critically important ribonucleoprotein particle is assembled in the cell. Studies in both mammalian cells (Jacobson and Pederson, 1998; Politz et al., 2000; Politz et al., 2002; Alavian et al., 2004) and yeast (Ciufo and Brown, 2000; Grosshans et al., 2001) have implicated the nucleolus as a possible initial site in the assembly of the SRP. The presence of SRP components in yeast and mammalian nucleoli constitutes one of several lines of evidence that the nucleolus has functions beyond its classically established role in ribosome synthesis (Pederson, 1998; Olson et al., 2000; Olson et al., 2002; Gerbi et al., 2003).
The oocytes of amphibians and certain other eukaryotic organisms contain a special class of multiple nucleoli (Montgomery, 1898; Painter and Taylor, 1942). These multiple nucleoli are amplified from the genomic ribosomal RNA genes into hundreds or thousands of extrachromosomal nucleoli that produce the massive amounts of maternal ribosomes needed to support early embryonic development (Brown and Dawid, 1968; Gall, 1968; Macgregor, 1972; Gall, 1978). Because these specialized nucleoli have such a clear developmental purpose, we were interested to know whether SRP components might also be present within them. If the biological rationale of the amplified nucleoli is to produce a maternal stockpile not only of ribosomes per se, but of translational machinery altogether, it might be anticipated that they also are committed to SRP biosynthesis. The findings we now report strongly support this hypothesis and suggest that at least the initial steps in SRP ribonucleoprotein assembly, involving SRP RNA and SRP19 protein, occur in these nucleoli.
Materials and Methods
Oocyte microinjection experiments
Female Xenopus laevis at 8-9 cm length were obtained from Blades Biologicals (Kent, UK) and maintained in an aquarium under optimal conditions for further growth. Ovary was removed from mature animals killed under Schedule 1 (Home Office, London, Animals Scientific Procedures Act 1986) and individual oocytes were released by stirring the tissue in a solution of 0.2% collagenase (Sigma, Type I) in OR2 medium minus Ca2+ (Evans and Kay, 1991). Mid-vitellogenic oocytes at stage IV/V (Dumont, 1977) were selected and maintained in OR2 medium (including 1 mM CaCl2) at 19-20°C. Fluorescent RNA or protein was microinjected into either the cytoplasm or germinal vesicle (nucleus) as previously described in detail (Ryan et al., 1999). The injection volume was typically 20 nl for cytoplasmic and 10 nl for nuclear injection. Oocytes were dissected under oil into germinal vesicles (GVs) and cytoplasm at various times after microinjection, and RNA or protein distribution was analysed as detailed below. GV spreads were prepared as previously described (Callan et al., 1987).
Human SRP RNA, which shares ∼87% sequence homology with Xenopus SRP RNA (Ullu and Tschudi, 1984), was used in this study. Plasmid phR (Zwieb, 1991; Jacobson and Pederson, 1998) was linearized by DraI digestion and transcribed with T7 RNA polymerase in the presence of all four ribonucleoside triphosphates and 5-Alexa 488-UTP (Molecular Probes) present at the same concentration as UTP, using a Megascript kit (Ambion). A transcript with a sequence complementary to canine SRP RNA was transcribed with SP6 RNA polymerase from EcoR1 linearized plasmid pSP7SL (Strub et al., 1991) with ribonucleoside triphosphates and 5-Alexa 488-UTP as above. Transcripts were purified using Biogel P-30 spin columns (BioRad), ethanol precipitated and dissolved at a concentration of approximately 1 μg/μl in water. Injected specific (sense) and control (antisense) probes had comparable levels of fluorescence.
Immunoblots and immunocytochemistry
GVs and cytoplasms were isolated and extracted as described previously (Smillie and Sommerville, 2002). Extracts equivalent to four GVs and one cytoplasm were denatured in an equal volume of Laemmli sample buffer (Sigma) and subjected to electrophoresis on 12% SDS-polyacrylamide gels and then electro-transferred onto nitrocellulose membrane (`Protran', Schleicher and Schull). Transfers were blocked overnight at 4°C in 10% dried nonfat milk dissolved in phosphate-buffered saline containing 0.1% Tween 20 (PBST) followed by incubation with a primary antibody diluted in PBST for 1 hour at 20°C, washing through five changes of PBST, and then incubated with a peroxidase-conjugated secondary antibody diluted in PBST. Bands were developed using the ECL (Amersham Biosciences) procedure.
GV spreads were blocked in 10% fetal calf serum in PBS (FCS/PBS), incubated with a primary antibody diluted in FCS/PBS for 1 hour at 20°C, washed through five changes of PBS and then incubated with a fluorochrome-conjugated secondary antibody diluted in FCS/PBS). After further washing with PBS, preparations were mounted in 50% glycerol, 1 mg/ml p-phenylenediamine, pH 8.5, and viewed in a Leitz Ortholux fluorescence microscope. Controls omitting a primary antibody incubation showed no secondary antibody labelling.
GFP-labelled SRP19 protein
The insert from a fusion plasmid described previously (Politz et al., 2000) containing the coding sequence for EGFP-SRP19 was cloned into a bacterial expression plasmid (pET vector, Novagen) containing a [His]6 in-frame coding sequence (Henry et al., 1997) to create a recombinant plasmid that would express H2N-[His]6-EGFP-SRP19-COOH. Escherichia coli was transformed with this plasmid and the expressed GFP-SRP19 protein was recovered and purified by selection on a Ni NTA agarose column (Qiagen) essentially as described previously (Henry et al., 1997), except that samples were concentrated using Biomax-100K filters (Millipore). The purified protein was diluted in water at a concentration of 1 μg/μl for microinjection. The localization of GFP-SRP19 protein in the nucleus after microinjection was analysed by fluorescence microscopy of spread GV contents. The level of GFP-SRP19 protein in the nuclear or cytoplasmic fractions was also determined by immunoblotting using a rabbit antibody to GFP (kindly provided by David Russell, Washington University School of Medicine) at a dilution of 1/8000 and peroxidase-conjugated anti-rabbit antibody (Sigma) at a dilution of 1/10,000. In some experiments, oocytes were incubated with leptomycin B (kindly provided by Minori Yoshida, University of Tokyo) at a concentration of 50 ng/ml. In other experiments, oocytes were co-injected in the cytoplasm with both GFP-SRP19 and an equimolar amount of recombinant human importin-α2 (Calbiochem). Assembly of GFP-SRP19 into a cytoplasmic particle was investigated by sedimentation of cytoplasmic extracts, prepared as previously described (Smillie and Sommerville, 2002) on 15-30% linear glycerol gradients. Some samples were digested before gradient analysis with ribonuclease A at 1 μg/ml for 30 minutes in the absence of Mg2+ (Walter and Blobel, 1983). Fluorescence in each fraction was read first in a fluorimeter (VersaFluor™, Biorad) with filters appropriate for EGFP. Protein was then precipitated from the gradient fractions with 3 volumes of acetone at –20°C and pelleted by centrifugation at 10,000 g for 20 minutes. Dried pellets were dissolved in Laemmli sample buffer (Sigma) and subjected to immunoblot analysis with GFP antibody.
In all, 900, 300, 120, 50, 30 and 25 oocytes at stage I, II, III, IV, V and VI, respectively, were selected to produce equal volumes of clarified supernatant after extraction with trichlorotrifluoroethane (Sigma) and centrifugation at 10,000 g for 2 minutes, followed by RNA extraction as described previously (Evans and Kay, 1991). RT-PCR (Ready-to-Go beads, Amersham Biosciences) was run using primers specific for Xenopus SRP RNA (forward: 5′GCTGTGGCGTGTGCCTGTAATCCAG and reverse: 5′GGGTTTTGACCTGCTCCGTTTCCGAC) and SRP54 mRNA (forward: 5′CTGGAGGAAATGGCATCTGGCTTGA and reverse: 5′TAGAAGGGTATTCTGGCTTTTGTGGC). RT-PCR products amplified after 15, 20 and 25 cycles at 95°C, 48°C and 72°C, respectively, in a MiniCycler (MJ Research), were separated on 2% agarose gels and 5S RNA was separated directly from the total RNA sample. Intensity of staining with ethidium bromide was captured using a GeneSnap system and measured using GeneTools (Synoptics Ltd).
RNA in situ hybridization
GV spread preparations were incubated with a tetramethylrhodamine-6-isothiocyanate-labelled peptide nucleic acid probe synthesized by Applied Biosystems (Politz et al., 2002), complementary to nucleotides 231-245 of Xenopus SRP RNA. Hybridization was carried out with 0.3 ng/μl of probe in 40% formamide, 4× SSC, 0.1 M phosphate, pH 7.2, 0.2 mg/ml tRNA and 0.2 mg/ml denatured sonicated DNA, for 12 hours at 42°C. Washes were as described (Politz et al., 2002). As a control, a peptide nucleic acid probe complementary to nucleotides 192-206 of Schizosaccharomyces pombe SRP RNA was used.
Detection of endogenous SRP19 protein
To detect endogenous SRP19 protein, extracts were electrophoresed on 16% polyacrylamide gels and transferred to nitrocellulose as above. Immunoblotting was with a chicken antibody against the human SRP19 protein (see below) diluted 1/5000, followed by peroxidase-conjugated rabbit anti-chicken IgG (Sigma) diluted 1/10,000.
GV spreads were immunostained with the SRP19 antibody diluted 1/200, followed by FITC-conjugated rabbit anti-chicken IgG (Sigma) diluted 1/2000.
The antibody to human SRP19 protein was generated in collaboration with Jos Raats and Walther van Venrooij, University of Nijmegen, Holland. Human SRP19 protein was expressed in E. coli transformed with a (His)6-tagged human SRP19 plasmid and purified as detailed previously (Henry et al., 1997) and then conjugated to keyhole limpet haemocyanin. Chickens (Gallus domesticus) were immunized and sera were screened by ELISA against the immunogen. An animal displaying a strong immune response (threshold detection at a serum dilution of 1:12,800) was boosted with another injection. Immunoglobulin Y was isolated from eggs laid by this animal by subjecting pooled yolks to lipoprotein depletion and ammonium sulphate precipitation (Eggcellent IgY Purification Kit, Pierce Biotechnology), followed by affinity purification on a Sepharose column to which the recombinant SRP19 protein had been coupled.
Preparations were examined using a Leitz Ortholux Fluorescence Microscope (100× oil-immersion objective) and photographed on Kodak Ektachrome P1600 colour reversal film (Kodak Eastman). Frames were scanned with a Microtek FilmScan 35 using CyberView interface software (Microtek Europe B.V.) and imported into Adobe Photoshop. Any recolouring was carried out using the Photoshop application and merged constrained images were created and flattened.
Synthesis of oocyte SRP components coincides with activity of amplified nucleoli
To establish the occurrence and timing of SRP biosynthesis, equal masses of oocytes from Xenopus stages I to VI (Dumont, 1972) were examined for the presence of SRP RNA, mRNA encoding the 54 kDa SRP protein (SRP54, which is the final protein to be added to the SRP in a reaction that appears to occur in the cytoplasm) and the 19 kDa SRP protein (SRP19) itself.
RT-PCR was set up with primers specific for Xenopus SRP RNA and SRP54 mRNA, using total RNA from each oogenic stage as templates, and reaction products were recorded from agarose gels after 15, 20 and 25 cycles. Products were detected at all oogenic stages (Fig. 1A). To compare the levels of SRP RNA with another RNA polymerase III (pol III) transcript, the amount of endogenous 5S RNA was recorded from the same samples (Fig. 1A). Conversion of band densities to amounts of RNA per oocyte (Fig. 1D) showed that whereas accumulation of 5S RNA occurs during previtellogenesis (stages I to II), accumulation of SRP RNA, which is also transcribed by pol III, occurs throughout oogenesis (including vitellogenic stages II to VI), although at a slower rate after stage IV. Thus, the kinetics of SRP RNA synthesis in oocytes appear to be more akin to those of ribosomal RNA (Scheer, 1973; Scheer et al., 1976) than those of the major pol III species, 5S RNA and tRNA (Ford, 1971; Mairy and Denis, 1971). On estimating the amounts per oocyte of SRP54 mRNA from RT-PCR (Fig. 1A) and the amounts of SRP19 protein contained within particles sedimenting at ∼11S (the sedimentation rate of mature mammalian SRP) (Walter and Blobel, 1983) by immunoblotting (Fig. 1B,C), it appears that the kinetics of synthesis of SRP proteins are similar to those of SRP RNA (Fig. 1D). Given that accumulation of ribosomes in oocytes is dependent on the development of active, amplified nucleoli (Miller and Beatty, 1969), we questioned whether the observed accumulation of SRP components through mid-vitellogenesis likewise involves active, amplified nucleoli.
SRP RNA localizes in the amplified nucleoli after microinjection into the GV
Fluorescent human SRP RNA was injected into the germinal vesicles (GVs) of stage IV/V oocytes and its distribution among nuclear components was examined in spread preparations of isolated GVs (Fig. 2). At 18 hours after microinjection, distinct, strong fluorescence was observed in the nucleoli (Fig. 2C). No significant signal was observed in either the chromosomes or Cajal bodies (Fig. 2A,B), nor in snurposomes (present throughout the field). All of the amplified nucleoli examined were labelled. When GVs were analysed at shorter times after microinjection of SRP RNA, there was no indication of fluorescence in nuclear structures other than the nucleoli, suggesting that SRP RNA directly localizes in the nucleoli after microinjection, rather than first passing through some other nuclear structures. The nucleolar localization of SRP RNA was sequence-specific, as shown by the lack of detectable nucleolar localization of a control RNA with the antisense sequence to canine SRP RNA (Fig. 2D-F).
Endogenous SRP RNA is present in the amplified nucleoli
The localization of fluorescent SRP RNA in the amplified nucleoli following its microinjection into the nucleus does not necessarily mean that endogenous SRP RNA is present there as well. To address this point, GV spreads were subjected to RNA hybridization in situ with a peptide nucleic acid (PNA) probe specific for Xenopus SRP RNA (see Materials and Methods). As shown in Fig. 3C, hybridization was detected specifically in the nucleoli and not appreciably in the chromosomes or other nuclear structures (Fig. 3A). The SRP RNA hybridization was specific, as shown by the lack of hybridization when a control probe for yeast SRP RNA was used (Fig. 3E-G). Fields containing contaminating autofluorescent yolk platelets are shown to allow comparison with specific hybridization signals (Fig. 3C,G).
Notably, the localization of SRP RNA signal within the nucleoli did not appear to be uniform. Nucleoli typically have a tripartite structure, consisting of small foci called fibrillar centres (FCs; the sites of the ribosomal genes) situated within regions called the dense fibrillar component (DFC); these regions in turn are surrounded by a granular component (GC), and specific stages of the ribosome biosynthesis pathway have been assigned to these three domains (Goessens, 1984; Hernandez-Verdun, 1991; Spector, 1993; Shaw and Jordan, 1995; Scheer and Hock, 1999; Olson et al., 2000). This classical tripartite organization is also observed in the amplified nucleoli of amphibian oocytes (Mais and Scheer, 2001).
In the nucleolar preparations examined in the current study, the FCs were identified as the DAPI-staining structures and the DFC was defined by immunostaining with an antibody to the Xenopus nucleolar protein, xNop180 (Cairns and McStay, 1995; Schmidt-Zachmann, et al., 1984). In Fig. 3D, the DNA signal has been computationally coloured green and merged with the in situ hybridization signal (red). In all four of the nucleoli, the DNA-rich regions and the SRP RNA appear to be highly segregated, with the probe located largely around the FCs and extending through much of the DFC and into the GC.
The SRP19 protein is imported into the GV
The findings that microinjected SRP RNA targets the nucleoli and that the endogenous SRP RNA is localized in them raise the possibility that the amplified nucleoli are sites of SRP RNA assembly into nascent SRP particles. The binding of the SRP19 protein to SRP RNA is the first step in the biochemically determined assembly pathway of the signal recognition particle. Therefore, it would be expected that SRP19 would colocalize with SRP RNA in the nucleoli if SRP assembly in fact occurs at these sites. To follow SRP19 protein in the oocyte, we used a recombinant GFP-SRP19 fusion protein (see Materials and Methods). We first investigated whether the GFP-tagged SRP19 protein was capable of being imported into the nucleus. It was injected into the cytoplasm of oocytes and its distribution in nuclear and cytoplasmic fractions was analyzed by immunoblots with a GFP antibody (Fig. 4). The GFP-SRP19 protein appeared in the nuclear fraction in increasing amounts over a 10 hour period (Fig. 4A). The apparent molecular weight of ∼47 kDa is close to that anticipated, as the fusion protein consists of SRP19 and the ∼26 kDa GFP, and the persistence of this band indicates that the nucleus-imported GFP-SRP19 remains intact. Quantitation of the immunoblot results revealed a near-linear nuclear import over the 10 hour period of these experiments (Fig. 4B). By immunoblotting with the SRP19 antibody (see Fig. 1B,C), it was estimated that GFP-SRP19, recovered from injected oocytes after 10 hours, amounted to 10-20% of endogenous SRP19. In parallel experiments, the nuclear import factor importin-α was co-injected into the cytoplasm with GFP-SRP19 protein. This resulted in a pronounced initial increase in the kinetics of GFP-SRP19 import (Fig. 4B), adding further evidence that the GFP version of SRP19 was behaving in a physiologically relevant manner.
The SRP19 protein is exported from the GV and is incorporated into a cytoplasmic RNP particle
To assess further the behaviour of GFP-SRP19 protein in oocytes, its longer-term distribution between the nucleus and the cytoplasm was investigated after treatment of oocytes with leptomycin B, an inhibitor of the Crm1 nuclear export machinery. Experiments in yeast have established that the nuclear export of SRP particles is Crm1-mediated (Ciufo and Brown, 2000; Grosshans et al., 2001), and there is evidence that this is also the case in mammalian cells (Alavian et al., 2004). As shown in Fig. 4C, GFP-SRP19 normally began to leave the nucleus after 10 hours, but treatment with leptomycin B resulted in an increased level of GFP-SRP19 in the nuclear fraction at 10 hours of treatment, this effect becoming more pronounced at 20 hours, suggesting that GFP-SRP19 export from the oocyte nucleus is Crm1-dependent.
We next asked if GFP-SRP19 protein is assembled into a ribonucleoprotein particle in the oocyte. GFP-SRP19 was injected into the nucleus and cytoplasmic extracts were prepared after 48 hours. Glycerol gradient analysis revealed that the majority of GFP fluorescence sedimented at ∼10S (Fig. 5A), which is close to the sedimentation coefficient (∼11S) of native SRP (Fig. 1B). Treatment of the cytoplasmic fraction with ribonuclease resulted in a substantial decrease in the sedimentation velocity of the SRP19 protein to less than 4S (Fig. 5A). In cytoplasms isolated immediately after injection, the fluorescent signal sedimented close to that expected for soluble protein monomers (Fig. 5A). The gradient fractions were further analysed by immunoblotting with GFP antibody (Fig. 5B), confirming that the fluorescent signal corresponded to the distribution of intact GFP-SRP19. Thus, by all these criteria (nuclear import, effect of importin-α on nuclear import, leptomycin B inhibition of nuclear export and assembly the exported protein into a ∼10S cytoplasmic particle, the GFP-SRP19 protein behaved in a manner entirely consistent with the pathway of SRP biosynthesis described in other systems.
The SRP19 protein localizes in the amplified nucleoli
We investigated the nuclear localization of GFP-SRP19 after import from the cytoplasm. As shown in Fig. 6C, at 1 hour after microinjection into the cytoplasm there was a faint signal in the nucleoli. After 2-5 hours, the nucleolar signals became progressively stronger, spreading out from around the DAPI-staining FCs to encompass the whole nucleolar space (Fig. 6F,I,L). Of the GFP-SRP19 contained in the nucleus at 5 hours postinjection, more than 80% could be pelleted by low-speed centifugation (3000 g for 5 minutes), indicating that only a minor component was present in the nucleoplasm. This observation was confirmed by microscopic examination of GV contents isolated in the presence of 1 mM Mg2+; the nucleoplasm, gelled under this condition, exhibiting a much lower concentration of fluorescence than the embedded nucleoli (not shown). There was no indication of GFP-SRP19 protein localizing on the chromosomes or Cajal bodies (compare fluorescence in F,I,L with the phase-contrast images in D,G,J). All other fields examined failed to show labelling on structures other than nucleoli. A negative control for fluorescence is provided by the nuclear behaviour of GFP itself. As we have shown previously, GFP expressed from microinjected plasmids equilibrates throughout the oocyte but does not bind significantly to any nuclear structures (Smillie and Sommerville, 2002).
When spread GV contents are exposed to low-salt medium lacking MgCl2, the nucleoli become distended into a necklace-like array of beaded substructure (Wu and Gall, 1997) due to a loosening and dissociation of a cortical structure that extends over the nucleolar surface (Kneissel et al., 2001). When GFP-SRP19 protein was injected into the cytoplasm and nuclear spreads were then prepared in a low-salt medium lacking MgCl2, the signal was observed to be extended in a spatial configuration that coincided with the distended nucleolar substructure (Fig. 7A-C). Moreover, the GFP-SRP19 protein remained associated with this nucleolar substructure after storage of preparations in 70% ethanol (Fig. 7D-F), a condition under which GFP alone is solubilized (J.S., unpublished). This result establishes that the GFP-SRP19 protein is tightly associated with nucleolar components.
Endogenous SRP19 is present in the amplified nucleoli
GVs were again spread in low-salt medium lacking Mg2+ to reveal the substructures of the nucleoli at greater spatial resolution. As shown in Fig. 8C,E, the nucleoli showed distinct staining with an antibody to SRP19 protein, with no signal evident in Cajal bodies (Fig. 8A). Very fine foci lacking immunoreactivity were observed within the nucleolar beads and these corresponded to the DAPI-stained FCs (Fig. 8B,D). This suggests that the highest concentration of SRP19 protein is circumferentially arrayed around the FCs, as was observed for SRP RNA (Fig. 3). In the merged image (Fig. 8F), the FCs mostly appeared red, with some immediately adjacent merged signal (yellow), in turn surrounded by SRP19 alone (green). This suggests that SRP19 is distributed throughout the nucleoli except for the FCs, and that some of the protein occupies a location immediately adjacent to FCs. (It is probable that most of the merged signal in Fig. 8F represents spatial overlap in the z-axis.) These results show that the amplified nucleoli are the major intranuclear sites of the endogenous SRP19 protein in the oocyte nucleus. Together with the GFP-SRP19 results and the nucleolar localization of both microinjected and endogenous SRP RNA, our results strongly support the hypothesis that the amplified nucleoli of the amphibian oocyte are involved in SRP biosynthesis.
The initial idea that the nucleolus may have functions other than ribosome assembly (Pederson, 1998; Pederson, 1999; Pederson and Politz, 2000) has continued to be borne out by numerous studies that have revealed the presence in yeast and mammalian cell nucleoli of various proteins and RNA species that have no obvious connection with ribosome biosynthesis. In the present investigation we have identified both the RNA and a protein component of the signal recognition particle in the amplified nucleoli of Xenopus oocytes, specifically at a developmental stage, mid-vitellogenesis, when nucleolar activity is at a maximum. Previous studies had also identified SRP RNA and SRP proteins in the nucleoli of somatic mammalian cells (Jacobson and Pederson, 1998; Politz et al., 2000; Politz et al., 2002). One interpretation of those findings was that SRP assembly, at least its initial stages, takes place in nucleoli. However, another interpretation was that SRP RNA and SRP proteins remain uncomplexed in nucleoli and serve some unknown functions unrelated to SRP biosynthesis. This latter hypothesis can now be viewed as quite implausible as we show here that SRP RNA and the SRP19 protein are also present in the oocyte-amplified nucleoli. These nucleoli are specialized for (at least) ribosome assembly and reside in a cell-cycle-arrested (meiotic) nucleus. Yet they contain SRP RNA and an SRP protein, SRP19, which is involved in an early stage of SRP assembly. Therefore, it seems quite likely that the presence of the SRP components in oocyte nucleoli reflects an actual role of these nucleoli in producing a maternal stockpile of SRP. This interpretation is strongly supported by our observation that when GFP-SRP19 was injected into the cytoplasm and localized in the amplified nucleoli, it was subsequently observed to be exported back to the cytoplasm where it was identified in a ribonucleoprotein particle having a sedimentation coefficient in glycerol gradients close to that of the mammalian SRP particle. The present investigation thus moves the plurifunctional nucleolus concept ahead with respect to a role in the biosynthesis of the SRP.
The results of this study point to a high mobility of fluorescent SRP RNA injected into the nucleus of Xenopus oocytes, since the introduced RNA was observed to traffic to all of the ∼1000 extrachromosomal nucleoli. Similar observations have been previously made in studies of other species of RNA injected into the germinal vesicle (Gerbi et al., 2003; Handwerger et al., 2003). The fact that endogenous SRP RNA was also found in all of the nucleoli implies that these molecules traffic extensively from their transcription site throughout the nucleoplasm to reach all of the amplified nucleoli. This wide-ranging mobility of RNA in the germinal vesicle is similar to the diffusive transport of RNA in the mammalian cell nucleus (Jacobson and Pederson, 1998; Politz et al., 1998; Politz et al., 1999; Politz et al., 2003).
We were struck by the unique concentration of SRP RNA in only the amplified nucleoli, and not in other RNA traffic centres of the Xenopus oocyte nucleus. Similarly, in previous studies on mammalian cells there was no evidence of SRP RNA in either of two RNA-rich structures: interchromatin granule clusters (also known as `speckles') or Cajal bodies (Jacobson and Pederson, 1998; Wang et al., 2003), and this was confirmed in the present study in which no association of SRP RNA with either B-snurposomes (homologous with speckles in mammalian somatic nuclei) or Cajal bodies was observed in the Xenopus oocyte nucleus. SRP RNA seems to be distinctive as one of the few nucleolus-trafficking small RNAs that does not also have a Cajal body resident stage (Gall, 2000; Gall, 2003; Pederson, 2004).
The present study also provides new information on the nuclear import of SRP19 protein and its export, studied here for the first time in amphibian oocytes. Our finding that the nuclear import of GFP-SRP19 is accelerated by co-microinjection of importin-α implicates this member of the importin family in the nuclear import of SRP19 in this developmentally specialized cell. A previous study of SRP19 protein nuclear import in a detergent-extracted mammalian cell in vitro system implicated two importins, importin 8 and transportin, both of which are members of the β-importin family (Dean et al., 2001). How SRP proteins are imported into the nucleus of various eukaryotic cells clearly will require further work, but our results draw attention to importin-α in Xenopus oocytes. Whether or not SRP19 is actively recruited to the extrachromosomal nucleoli is not known: a search of the human SRP19 amino acid sequence did not reveal any regions corresponding to nucleolus-localizing elements that have been defined in other nucleolar proteins (S. Yarovoi and T.P., unpublished). Our finding that leptomycin B causes an accumulation of SRP19 protein in the amphibian oocyte nucleus implicates the Crm1 pathway of nuclear export, which is the export machinery identified for SRP nuclear export in yeast (Ciufo and Brown, 2000; Grosshans et al., 2001). We have found no indication of a nuclear export signal (NES) in SRP19; indeed, our results with mammalian cells (Alavian et al., 2004) support an indirect, rather than a direct, export of nascent SRP by the Crm1 pathway.
Our present findings that both SRP RNA and an SRP protein are present in the amplified nucleoli of a cell, the amphibian oocyte, that is designed to stockpile the machinery of translation, adds considerable weight to the hypothesis that the nucleolus plays a role in SRP biosynthesis. Moreover, the present investigation shows that the oocyte is likely to be a valuable system for investigating further the biosynthesis of the SRP, both with respect to its nucleolar phase and its overall developmental timetable. In keeping with numerous studies of gene expression in this specialized cell, in which features of transcription, RNA processing and ribonucleoprotein assembly are quantitatively and cytologically intensified, we are optimistic that the Xenopus oocyte will be a key system in continuing studies of both SRP biosynthesis and the plurifunctional nucleolus concept.
This work was supported by a grant from the Wellcome Trust to J.S. and U.S. National Institutes of Health grant GM-21595 to T.P. We are indebted to Jos Raats (University of Nijmegen, The Netherlands) for his key role in generating the SRP19 antibody. We thank Howard Fried (University of North Carolina, Chapel Hill, NC) for providing the SRP19 bacterial expression plasmid, and we acknowledge Serge Yarovoi in the Pederson laboratory for constructing the GFP fusion protein. We thank Marion Schmidt-Zachmann (German Cancer Research Center, Heidelberg) for monoclonal antibody No-114 to detect xNopp180, David Russell (Washington University School of Medicine) for the GFP antibody and Minori Yoshida (University of Tokyo) for leptomycin B. Annemarie Mullin, Katherine Gilchrist and Ceri Jackson in the Sommerville laboratory are thanked for their help with the immunoblotting, in situ hybridization and RT-PCR experiments, and Laura Lewandowski in the Pederson laboratory is thanked for her skilfull transcription and fluorescent labelling of RNAs.
↵‡ Present address: Department of Neuroscience, University of Connecticut Health Center, Farmington, CT 06032, USA
- Accepted January 13, 2005.
- © The Company of Biologists Limited 2005