|Home | About | Journals | Submit | Contact Us | Français|
A large share of mRNA processing and packaging events occurs cotranscriptionally. To explore the hypothesis that transcription defects may affect mRNA fate, we analyzed poly(A)+ RNA distribution in Saccharomyces cerevisiae strains harboring mutations in Rpb1p, the largest subunit of RNA polymerase II. In certain rpb1 mutants, a poly(A)+ RNA granule, distinct from any known structure, strongly accumulated in a confined space of the cytoplasm. RNA and protein expressed from Ty1 retrovirus-like elements colocalized with this new granule, which we have consequently named the T body. A visual screen revealed that the deletion of most genes with proposed functions in Ty1 biology unexpectedly does not alter T-body levels. In contrast, the deletion of genes encoding the Mediator transcription initiation factor subunits Srb2p and Srb5p as well as the Ty1 transcriptional regulator Spt21p greatly enhances T-body formation. Our data disclose a new cellular body putatively involved in the assembly of Ty1 particles and suggest that the cytoplasmic fate of mRNA can be affected by transcription initiation events.
In eukaryotic cells, mRNA-protein particles (mRNPs) are assembled in the nucleus and are normally transported to cytoplasmic ribosomes for translation. The protein composition of mRNPs is highly dynamic and changes in both a temporal and a spatial fashion, consistent with a changing need for associated factors at different stages of the mRNP life cycle. The initial mRNA-protein association occurs as soon as the 5′ end of the nascent RNA extrudes from the exit channel of the RNA polymerase II (RNAPII) complex (4, 63). A number of factors coat the mRNA cotranscriptionally in order to effectively process the molecule into a mature length as well as to make the new mRNP competent for nuclear export (12). Importantly, RNAPII itself is an active player in these early mRNP assembly processes. This is perhaps best illustrated by the physical interactions of mRNA processing factors with the carboxy-terminal domain (CTD) of Rpb1p (50). However, additional levels of RNAPII/mRNP intimacy exist, exemplified by the suggestion that the RNAPII subunits Rpb4p and Rpb7p, which both affect the cytoplasmic decay rate of certain mRNAs, might be integral mRNP components (35).
Although mRNA ultimately functions as a template for translation, the characteristics of mRNPs made from different genes are highly variable at the levels of cellular localization, mRNA stability, and ribosome availability. Combined with a high level of mRNP maturation control, this allows for ample plasticity of gene expression at posttranscriptional stages. For example, controlling the spatial distribution of mRNPs regulates developmental processes in eukaryotes, e.g., the location of Vg1 transcripts at the vegetal pole of Xenopus oocytes (38, 64) and the asymmetric distribution of ASH1 RNA to the daughter cell during budding in Saccharomyces cerevisiae (34, 59). Variability between mRNPs produced from the same gene also occurs. This may be caused by transcription or mRNA processing inaccuracies or by faulty mRNP packaging. In such cases, a series of nuclear and cytoplasmic quality checkpoints are normally in place to ensure that inappropriate mRNA translation does not take place (16, 53).
Once the productive life of a transcript comes to an end, the mRNP disassembles and the mRNA is degraded. This occurs either in a delocalized manner or within specialized cytoplasmic aggregates or bodies. In S. cerevisiae, the decrease in translation efficiency by severe stress induces mRNA relocalization from ribosomes into so-called processing bodies (P bodies). Apart from mRNAs tagged for degradation, S. cerevisiae P bodies contain decay enzymes such as the decapping factors Dcp1p and Dcp2p as well as the 5′-3′ exoribonuclease Xrn1p (58). Interestingly, P bodies not only are sites of degradation but also can serve to store mRNAs, which may later remerge and engage in translation (6).
Some mRNAs experience a different fate in the cytoplasm. This is the case for transcripts arising from the S. cerevisiae Ty retroelements. Ty genes fall into five classes (Ty1 to Ty5), of which the Ty1 retroelements comprise the largest class, present in 30 to 35 copies per genome and producing approximately 5 to 10% of the steady-state cellular mRNA pool in haploid cells (18). Ty1 RNAs are used as translation templates for synthesis of the virus-like capsid proteins Tya/p1 and Tya-Tyb/p3 (reviewed in reference 51). Ty1 transcripts also serve as reverse transcription templates after they have become packaged along with Tya and Tya-Tyb into virus-like particles (VLPs) (5, 19, 37). After further polypeptide maturation and reverse transcription inside the VLP, preintegration complexes containing Ty1 cDNA and the integrase protein are imported into the nucleus, where the elements are inserted in the genome (28, 39).
In recent years, significant functional cross talk between cotranscriptional mRNP formation and the cytoplasmic fate of transcripts has become evident. For instance, the pre-mRNA splicing process decorates mRNA with specific proteins, which in turn affect downstream processes like translation, cytoplasmic localization, and mRNA surveillance (21, 43). Similarly, binding of hnRNPI and Vg1RPB/vera to Vg1 RNA initiates in the nucleus and is required for the correct RNA localization (13, 30). Additional examples are provided by the exclusively nuclear RNA binding protein Loc1p, which influences the cytoplasmic localization of ASH1 RNA (34), and by the predominantly nuclear zipcode binding protein 2 (ZBP2), which regulates the cytoplasmic localization of β-actin RNA by stimulating the cotranscriptional loading of the shuttling ZBP1 protein (20, 46).
In this paper, we report a hitherto-uncharacterized cytoplasmic granule, the T body, containing Ty1 RNA and Tya protein. Despite unchanged Ty1 RNA levels, the number of T bodies is significantly increased in certain S. cerevisiae strains carrying mutations or deletions affecting transcription. Our data provide support to the concept that alteration of early mRNP definition can have downstream consequences.
Media, growth conditions, and yeast manipulations were performed as previously described (36). Experiments were done with cells grown at 25°C exponentially in rich media, unless indicated otherwise. The strains used (those listed in Table Table11 and BY deletions from the yeast knockout collection) are isogenic (S288C background), with the exception of the rpb1-1, pap1-1, and mex67-5 strains, which are congenic to S288C (retrocrossed three times or more into the S288C background). GRY3020 to GRY3029, THJ2208, and THJ2209 were employed in the experiments shown in Fig. Fig.1.1. Data shown in the remaining figures were obtained using mutant strains with both kinds of mating types, a and α. The mating type did not affect the experimental outcome. Results presented in Tables Tables22 and and33 were produced using deletion strains of the yeast knockout collection affirmed of their expected phenotypes, i.e., drug sensitivity, complementation by the wild-type (WT) gene, tetrad segregation, and/or Southern analysis.
RNA fluorescence in situ hybridization (RNA-FISH), immunolocalization, and combined RNA-FISH/immunolocalization analyses were done as described previously (61, 62). Poly(A)+ RNA was visualized using 100 ng of Cy3-labeled THJ790 probe (62). Ty1 RNA was visualized using 60 ng of Cy3-labeled Ty1-INT probe (5′-GAGGTTCTAAACTACGCATATTCTTAGTATTCCATGTGTCTCGTGATACC-3′; T represents amino-modified C-6 deoxyribosylthymine [dT] residues amenable to Cy3 conjugation]). Similar results were obtained with the Ty1-GAG probe (5′-TCTGTTTTGGAAGCTGAAATGTCTAACGGATCTTGAGTTGTTTGGACTTC-3′). Colocalization experiments with Ty1 and poly(A)+ RNA were performed using a combination of Cy3-labeled THJ790 and Cy5-labeled Ty1-INT probes (similar results were obtained with inverse labeling of probes). The putative detection of Ty1 cDNA was excluded for the following reasons: (i) when using Ty1-GAGanti probe (5′-GAAGTCCAAACAACTCAAGATCCGTTAGACATTTCAGCTTCCAAAACAGA-3′), antisense to the Ty1 RNA, no signal was detected; (ii) under conditions where Ty1 cDNA levels are dramatically decreased (i.e., at 30°C ), T-body levels were unaltered; and (iii) the FISH protocol used is optimized for the specific detection of RNA.
Nop1p, Nsp1p, GFP (green fluorescent protein), and Tya proteins were visualized using the following mouse monoclonal antibodies: mAb28F2 (EnCor), mAb32D6 (EnCor), GFP (B-2; Santa Cruz Biotechnology), and Tya (BB2, serum; ). These antibodies were used in dilutions of 1/1,000, 1/500, 1/250, and 1/100, respectively. As a secondary antibody, anti-mouse immunoglobulin G (fluorescein isothiocyanate conjugate, F-0257; Sigma) was used at a 1/200 dilution. To reveal P bodies, the exogenously expressed YCplac111-DCP2-GFP (58) centromeric plasmid, containing the DCP2 gene under the control of its own promoter and C-terminally fused to GFP, was used.
Images were acquired at room temperature on a BX51 microscope equipped with a cooled DP50 charge-coupled-device camera, using a 100× UPIanFI objective and analySIS software (all items purchased from Olympus). Cy3, DAPI (4′,6′-diamidino-2-phenylindole), and GFP signals were visualized with U-MWIG2 (Olympus), U-MWU2 (Olympus), and U-N41001-HQ (Chroma) filters, respectively. Cy3/Cy5 double-labeling experiments were visualized on an Axiovert 200M (Zeiss) microscope equipped with a Coolsnap HQ camera (Ropers Scientific), using a heated stage and a 100× plan-Apochromat objective (Zeiss) and MetaMorph software (Universal Imaging Corp.). Cy5 was visualized using a 41008-Cy5 (Chroma) filter. Image handling was done in Adobe Photoshop. All images in a single panel were modified in similar ways. Cell counting was done by selecting random fields under the DAPI filter followed by image acquisition using the different filters. A minimum of 100 cells per experiment and per strain were always counted.
RNA extractions were done as described previously (36) from cells growing exponentially in rich media at 25°C. Primer extension analysis was performed with SuperScript II reverse transcriptase (Invitrogen) following the recommendations of the manufacturer. Northern blot analysis was done as described previously (36). Ty1 probes were 5′ end labeled with γ-ATP using T4 polynucleotide kinase (New England Biolabs). Hybridization with oligonucleotides was done using Ultrahyb-Oligo hybridization buffer (Ambion). Poly(A)+ purification was done using a Micropoly(A) purist kit (Ambion).
To learn more about the functional coupling between nuclear and cytoplasmic mRNP biology, we reasoned that defects in the mRNA production machinery might modulate mRNP characteristics and affect their cellular fates. We therefore analyzed the cellular distribution of bulk poly(A)+ RNA in 10 different rpb1 mutant strains, including nine single site rpb1 mutations previously isolated as being either thermosensitive or sensitive to the nucleotide-depleting drug 6-azauracil (36) (named here according to the position of the altered amino acid) as well as the rpb1-1 allele, widely used previously to turn off transcription at high temperatures (41). When cells were cultured at 25°C, we observed a dramatic accumulation of poly(A)+ RNA in a single cellular body in 25 to 40% of exponentially growing rpb1-1, rpb1-67, rpb1-70, and rpb1-80 mutant cells (Fig. 1A and B and data not shown). This phenomenon also occurred in the rpb1-488, rpb1-261, and rpb1-1230 mutants, albeit at lower levels (Fig. (Fig.1B1B).
Remarkably, the localization of these poly(A)+ RNA-containing bodies with respect to the chromatin-rich DAPI-stained area suggested that they did not reside in the nuclei of cells (Fig. (Fig.1A1A and data not shown). Costaining of cells with an antibody against either the nucleolar protein Nop1p or the nuclear envelope marker Nsp1p confirmed this finding and demonstrated a strict cytoplasmic localization of the poly(A)+ RNA body (Fig. (Fig.1C).1C). Additional analyses also revealed no overlap with mitochondria or the vacuole and furthermore demonstrated a random positioning of the poly(A)+ signal relative to these yeast organelles (data not shown). We conclude that a number of rpb1 mutants trigger a localized cytoplasmic accumulation of poly(A)+ RNA.
Ty1 transcripts represent 5 to 10% of the bulk poly(A)+ RNA in haploid S. cerevisiae cells, while Ty1 RNA levels are >95% lower in diploid cells (18). During the course of our investigations, we noticed that rpb1-induced poly(A)+ RNA bodies were not present in homozygous rpb1-67/rpb1-67 diploid cells (data not shown; further analyzed below). This fact, along with the high abundance of Ty1 RNAs, encouraged us to assay the cellular distribution of Ty1 transcripts relative to the poly(A)+ RNA bodies in rpb1-67 and rpb1-1 cells. Double-labeling experiments using the Cy3-labeled dT20 probe together with a Cy5-labeled Ty1-INT probe, targeting a highly conserved region in the Ty1 RNA encoding the integrase protein, demonstrated that the approximately 30% of rpb1-67 and rpb1-1 cells which stained positive for the cytoplasmic poly(A)+ RNA body all harbored an overlapping Ty1 RNA signal (Fig. (Fig.2A).2A). In addition, we observed a high proportion (~30%) of rpb1 mutant cells, which contained only the Ty1 and not the poly(A)+ RNA signal. Likewise, a similar proportion of WT cells also contained localized Ty1 RNAs, which were not associated with poly(A)+ material (Fig. (Fig.2A,2A, top). As signal intensities of the dT20 probe were slightly higher than those of the Ty1-INT probe, we conclude that two qualitatively different Ty1 RNA bodies were detected. Both contain Ty1 RNA, and accordingly we designate these new cytoplasmic structures “T bodies.” Moreover, we classify them T(A)+ or T(A)− bodies depending on whether or not they also stain positive for poly(A)+ RNA.
To address the increase in the number of T bodies and the curious appearance of T(A)+ bodies in rpb1-67 cells compared to those in WT cells, we analyzed total Ty1 RNA levels in the two strains by Northern blotting. Similar levels of Ty1 RNA were detected (Fig. (Fig.2B).2B). This result was also obtained when Ty1 RNA levels were assayed by primer extension analysis (see Fig. Fig.3B).3B). Next, we asked whether the proportion of polyadenylated Ty1 RNA was altered in rpb1-67 cells. We found that comparable amounts of Ty1 RNA were captured on oligo(dT)-Sepharose beads whether RNA was harvested from WT or rpb1-67 cells (Fig. (Fig.2B).2B). Thus, rpb1-67 cell-induced T(A)+ bodies are unlikely to be a consequence of altered Ty1 RNA polyadenylation. Attempts to identify other RNAPII-dependent polyadenylated RNAs associated with T bodies were unsuccessful; i.e., none of the abundant ACT1, PGK1, RPL30, or RPS5 transcripts showed any sign of localized distribution in the rpb1-67 strain (data not shown). Hence, the exact nature of the poly(A)+ signal in T(A)+ bodies remains a puzzle (see Discussion).
Ty1 transcription is dependent on mating-type locus control (18) and on the Spt3p transcription factor (67). To address the impact of Ty1 transcription on T-body formation and further validate the specificity of the Ty1-INT probe, we analyzed the effect of MATa/α expression or deletion of SPT3 on T-body numbers. Both T(A)+- and T(A)−-body signals decreased dramatically in rpb1-67 spt3Δ haploid double mutants and in homozygous rpb1-67/rpb1-67 diploid cells (Fig. (Fig.2C).2C). Taken together, these results demonstrate that T-body formation requires transcription of the retroelements and strongly suggest that the poly(A)+ RNA signal observed in T(A)+ bodies, independent of its origin, also requires Ty1 expression.
To characterize T bodies further, we first asked whether their formation was dependent on Pap1p, the major yeast poly(A)+ polymerase (33). To this end, we employed the pap1-1 allele (48) and quantified the number of cells harboring T bodies when grown at either 25°C or 30°C. As can be seen in Fig. Fig.2D,2D, pap1-1 cells grown at 25°C exhibited 40% the number of T bodies in WT cells. Further inactivation of Pap1p by increasing the temperature to 30°C virtually eliminated the formation of T bodies. A similar trend was observed in the rpb1-67 background; i.e., the introduction of the pap1-1 mutation decreased the levels of both T(A)− and T(A)+ bodies in the rpb1-67 strain at 25°C, an effect which was exacerbated at 30°C (Fig. (Fig.2D).2D). Thus, the conventional polyadenylation pathway is required for T-body formation, even for T(A)− bodies, where RNA poly(A) tails are not readily detected by the dT20 probe.
Finally, we asked whether T-body formation requires Mex67p, the major mRNA export receptor in yeast (54, 56). As shown Fig. Fig.2E,2E, thermosensitive mex67-5 cells contained WT levels of T bodies when grown at 20°C. However, upon a 30-min temperature increase to 30°C, Ty1 RNA started to accumulate in the nuclei of mex67-5 cells, concomitant with a reduced level of cytoplasmic Ty1 RNA signal (Fig. (Fig.2E).2E). We conclude that Ty1 RNA nuclear export and consequently T-body formation are mediated by the Mex67p export pathway.
We further assayed T-body contents by determining the cellular location of the Tya protein, the main structural component of Ty1 VLPs. For this purpose, we used an antibody raised against the Tya N-terminal amino acids 27 to 32 (8). In some cells devoid of T bodies, this reagent revealed a punctuated cytoplasmic Tya distribution, in agreement with the reported speckled cytoplasmic VLP distribution (9). However, in T-body-containing cells, a single strong Tya signal appeared, concomitant with the disappearance of the Tya speckled signal (Fig. (Fig.3A).3A). The Tya signal exhibited a behavior similar to that of the signal visualized with the Ty1 RNA-specific probe: (i) it was present in ~30% of WT and ~60% of rpb1-67 mutant cells, and (ii) it was observed in neither diploid nor spt3Δ cells (Fig. (Fig.3A3A and data not shown). Moreover, it colocalized with Ty1 RNA in 100% of the cases.
A previous report showed that N-terminal truncations of Tya form cytoplasmic aggregates (7). In contrast, T bodies harbor Tya proteins with intact N termini; i.e., the antibody used in our study recognizes the very N terminus of Tya and would not detect such truncated forms (8). Consistently, primer extension analysis of Ty1 RNA from rpb1-67 cells ruled out the existence of alternative transcription initiation sites which potentially could give rise to such truncated Tya species (Fig. (Fig.3B).3B). We conclude that Tya, most likely in its full-length form, is present in the T body.
P bodies in S. cerevisiae have previously been described as cytoplasmic structures containing mRNAs as well as enzymes involved in decapping and 5′-3′ exonucleolytic decay (58). P-body foci accumulate during stress conditions, and their formation is increased in the absence of the 5′-3′ exonuclease Xrn1p. Contrary to P bodies, the T(A)+ RNA bodies are readily detectable in exponentially growing cells. This suggests either that P bodies behave differently in the rpb1 mutants tested or that T(A)+ RNA bodies represent hitherto-undescribed cytoplasmic structures. In order to differentiate between these possibilities, we analyzed the distribution of the decapping enzyme and P-body marker Dcp2p in RPB1 and rpb1-67 strains either during exponential growth or after stressing the cells by glucose deprivation as previously described (6). Upon stress, a Dcp2-GFP fusion protein concentrated in P bodies both in WT and in rpb1-67 cells, concomitant with the disappearance of the T(A)+ body (Fig. (Fig.4A).4A). Moreover, enhancing P-body formation by means of an xrn1 deletion did not induce T(A)+ body formation either under exponential growth or under stress conditions.
Interestingly, when we performed a similar analysis with the Ty1 RNA-specific probe, a small fraction of cells presented T and P bodies simultaneously (Fig. (Fig.4B).4B). Importantly, these signals did not overlap. Moreover, T bodies were completely eliminated in spt3Δ cells, whereas P-body levels were unaffected (Fig. (Fig.4B,4B, bottom right). Finally, similarly to the poly(A)+ RNA signal, Ty1 RNAs and Tya proteins are redistributed upon glucose depletion, greatly decreasing the level of T bodies (Fig. (Fig.4B4B and data not shown). We conclude that T bodies and P bodies are different entities.
In order to identify cellular factors with a role in T-body formation, we monitored the numbers of total T bodies as well as T(A)+ bodies in a series of viable strains from the MATa yeast knockout collection. First, we focused our attention on factors that regulate Ty1 expression at the transcriptional and/or the posttranscriptional level (see references 45, 55, and 69 and references therein). Analyzed strains included those with deletions of genes identified as “regulators of Ty1 transposition” (RTT genes) and “suppressors of Ty” (SPT genes). The result of this scrutiny is summarized in Table Table2.2. As expected on the basis of the reported impaired Ty1 transcription in spt8 cells (66), the spt8Δ mutant, like the spt3Δ mutant, displayed a marked decrease in T-body levels. This argument is also applicable to other deletions shown in Tables Tables22 and and3,3, like the tec1Δ and snf2Δ mutants (22, 31), and strengthens the notion that T-body formation requires Ty1 transcription. Interestingly, Spt3p-independent expression of Ty1 RNA from the galactose-inducible pGTy1-H3mHis3AI construct (15) did not lead to T-body formation in the spt3Δ background (data not shown). Therefore, additional roles of these factors in T-body formation cannot be excluded. Surprisingly, however, of the tested deletion strains with altered Ty1 biology, all except one showed no detectable increase in T-body levels (Table (Table2).2). Specifically, neither increases in Ty1 RNA amounts (asf1Δ ) nor up- or downregulation of Ty1 cDNA and transposition levels (rtt109Δ and dbr1Δ [11, 27, 55]) had any effect.
Of the examined gene deletions, only the spt21Δ strain exhibited significantly increased levels of T bodies comparable to those in the rpb1-67 positive-control strain. Spt21p is a transcription factor that reportedly regulates histone gene expression levels (57). Interestingly, spt10Δ cells, which lack a protein that physically associates with Spt21p in vivo (23), do not display altered T-body counts. This suggests, in agreement with a previous report (24), that the Spt21p protein can function independently of Spt10p.
To assess the functional connection of RPB1 and SPT21 in relation to T-body characteristics, we analyzed total T- and T(A)+-body counts in rpb1-67 spt21Δ mutant cells. In this double mutant background, the levels of T bodies and the proportions of T(A)+ bodies increased significantly (Fig. (Fig.5A).5A). This additive effect of the rpb1-67 and spt21Δ phenotypes suggests that the Rpb1p and Spt21p proteins are involved in two different steps or pathways that normally control T-body levels. In addition, in the rpb1-67 spt21Δ double mutant background, increased T-body signal intensities, which correlated with increased T-body counts, could be observed (Fig. (Fig.5A).5A). Thus, once formed, the T body may continue to accumulate resident RNAs. However, since T-body signals are always found to be quite robust, a certain threshold level of RNA material is presumably required to form the granule.
Given the widespread examples in the literature of nuclear mRNP arrangements affecting cytoplasmic events, we reasoned that cotranscriptional Ty1 RNP formation might be challenged in at least some of the rpb1 mutants we had tested. In such a scenario, the mutation of other factors affecting RNAPII transcription should mimic the rpb1-induced T-body phenotype. Consequently, we extended the screen of knockout strains to also include deletions of genes with well-established roles in transcription initiation and/or elongation.
Surprisingly, none of the strains deleted for factors affecting transcription elongation, including subunits of the PAF and DSIF complexes, as well as TFIIS, Elf1p, Spt2p, and Chd1p (2, 44, 49), had adversely changed T- or T(A)+-body counts (Table (Table3).3). In contrast, certain strains deleted of specific subunits of the Mediator complex showed increased levels of T bodies similar to those for the rpb1 mutants (Table (Table33 and Fig. Fig.5A).5A). Like for the rpb1-67 mutation, this change occurred without a concomitant change in the level of Ty1 RNA (Fig. (Fig.5B).5B). Mediator, together with general transcription factors and RNAPII, forms the basic transcription initiation machinery (reviewed in reference 29). The most dramatic T-body effects were observed for deletion of the SRB2 and SRB5 genes, both of which encode subunits of the head module of Mediator (10). Deletion of the Mediator genes MED1 and MED9 also increased T-body counts but to a lower extent.
Finally, we also constructed the srb2Δ spt21Δ double mutant strain and analyzed its total T- and T(A)+-body levels. This double deletion background closely resembles that of the rpb1-67 spt21Δ strain (Fig. (Fig.5A).5A). Thus, the deletion of some Mediator subunits mimics the effect on T-body characteristics of the most severe rpb1 mutants.
Through the analysis of a collection of rpb1 mutants, we have identified a cytoplasmic Ty1 RNA-containing granule that we named the T body. Similarly to results for rpb1 mutants, defects in the transcription initiation Mediator complex and deletion of the Ty1 transcription regulator Spt21p significantly affect T-body characteristics. Thus, altered transcription in S. cerevisiae notably influences RNA localization. Since the effect of Mediator mutations is specific, i.e., it is not visible when other transcription factor genes are deleted, our data more generally suggest that events during transcription initiation can control the cytoplasmic localization of mRNP.
For eukaryotic cells, cytoplasmic RNA granules with different functions in posttranscriptional gene regulation have been described (1). Among these, P bodies have been characterized extensively for both S. cerevisiae and mammalian cells due to their role in storage and/or decay of mRNA. RNA stress granules and the confined localization of specific transcripts such as ASH1 and Vg1 provide additional examples. Given this active research, it is puzzling that the presence of Ty1 RNA inside T bodies has previously gone unnoticed. A limited use of Ty1 RNA-specific FISH probes probably accounts for the failure to visualize this phenomenon in WT cells. In addition, the fact that T(A)+ bodies are reliably detectable only in mutant backgrounds, coupled with the improved sensitivity of the RNA-FISH assay using short (dT20) DNA oligonucleotide probes spiked with locked nucleic acid residues (62), may explain why T bodies have not previously been revealed.
T bodies are bigger in size than P bodies, and as opposed to P bodies, they appear as one body per cell and are visible only in exponentially growing cells. Colocalization experiments of T versus P bodies confirmed that these structures are distinct. T bodies clearly contain RNA and protein produced from Ty1 retroelements. Interestingly, this is in contrast to yeast Ty3, which is expressed at lower copy numbers and accumulates viral RNAs and proteins in P bodies (3).
By use of Tya immunolabeling, the presence of multiple cytoplasmic aggregates or “clumps,” suggested to be VLPs, has also been reported for the NCYC74 yeast strain (9). In our S288C experimental strain background, we likewise observed a subpopulation of cells with dispersed Tya cytoplasmic foci. Interestingly, we were unable to detect Ty1 RNA in these foci; i.e., in cells containing both a T body and multiple Tya clumps, Ty1 RNAs were detected only inside the T body (Fig. (Fig.3A3A and data not shown). Therefore, in analogy to the suggested role for P bodies as Ty3 “VLP factories” in S. cerevisiae (3) and to the localized assembly of viruses in higher organisms (25, 26, 65), we speculate that T bodies are Ty1 VLP assembly sites. The high abundance of Ty1 retroelements may have favored the creation of a specific cytoplasmic body for this task, whereas the low abundance of the Ty3 element may follow a different strategy by taking advantage of a host structure to facilitate assembly. In such a scenario, the Tya clumps may be either VLP precursors or demolition sites for castaway VLP components.
The increase in T-body number in the more severe rpb1 mutants compared to that in WT cells was not accompanied by a significant variation in level or polyadenylation status of Ty1 RNA. This suggests that alterations in cotranscriptional Ty1 RNP assembly might underlie the observed phenotype. A fraction of improperly assembled Ty1 RNP could affect T-body dynamics and/or disturb the balance between the soluble and translationally active pools of Ty1 RNP. A phenomenon reminiscent of this suggestion has been demonstrated for human immunodeficiency virus type 1 (HIV-1), where depletion of the host RNP factor hnRNPA2 triggers an increased accumulation of HIV-1 RNA in cytoplasmic foci in the vicinity of the microtubule-organizing center, without affecting total levels of HIV-1 RNA (32).
It is not clear whether the increased association of poly(A)+ material with T bodies in, e.g., the rpb1-67 strain reflects alterations of Ty1 or host RNP. Unusual mRNPs could weaken the ability of Tya proteins to discriminate between Ty1 and other RNPs. Such a presence of host RNA inside VLPs has previously been reported (52, 68). Alternatively, a transcription-induced alteration of the Ty1 RNP might stall or slow VLP maturation, leading to inefficient shielding of Ty1 RNA poly(A) tails.
Of the genes tested that impact the Ty1 life cycle, only the deletion of genes known to be important for Ty1 transcription, like SPT3, SNF2, and TEC1, eliminates T-body formation (Table (Table2).2). Interestingly, conditions where Ty1 cDNA levels are dramatically decreased (i.e., at 30°C ) (Fig. (Fig.2D2D and data not shown) and mutations that up- or downregulate Ty1 cDNA and transposition levels (rtt109Δ and dbr1Δ [11, 27, 55]) have no effect. This suggests that the T body represents an intermediate stage in the Ty1 life cycle, previous to reverse transcription inside mature VLPs.
The only Ty1-related gene product whose deletion increased the number of T-body-positive cells was Spt21p. SPT21 was originally isolated, along with SPT10, via its ability to prevent transcription initiation from the 3′ long terminal repeat of Ty1 elements (40). Spt21p physically interacts with Spt10p, and both proteins are called hir (histone regulation) factors because they are required for proper transcription of histone genes (57). However, neither deletion of SPT10 or other hir genes, like SPT1, nor a series of histone gene mutants altered T-body levels (Table (Table22 and data not shown). This suggests an independent effect of Spt21p on T-body biology. Moreover, the additive effect on T-body counts achieved by combining the rpb1-67 or the srb2Δ mutation with the spt21Δ mutation suggests that Spt21p control of the T-body appearance may not completely overlap mechanistically with that of Rpb1p.
Unlike results for the spt21Δ strain, deletion of the SRB2 and SRB5 Mediator subunits affects T-body characteristics in a fashion very similar to that of the strongest of the rpb1 mutants. Again, this phenotype is triggered without altered levels of Ty1 RNA (Fig. (Fig.5B).5B). Thus, it is possible that the underlying defects in these mutants are similar. Given the pivotal role of the Rpb1p CTD in coupling transcription with mRNP assembly, alterations of the Rpb1p CTD seem a possible reason. However, the mutation of SRB genes suppresses the defects of the CTD (42, 60). Moreover, the zinc-binding domain of Rpb1p, in which the rpb1-67, rpb1-70, and rpb1-80 mutations map, is located in a mobile module of RNAPII, which is better appreciated for its role in transcription initiation (14, 29). Mutations in the Rpb1p zinc-binding domain also exhibit reduced affinity for the Rpb4/Rpb7 heterodimer (17), which has a reported effect on mRNP metabolism in the cytoplasm. However, neither deletion of RPB4 nor mutation of RPB7 impacts T-body counts (Table (Table33 and data not shown).
The mechanism by which Mediator affects Ty1 RNA localization has yet to be delineated. Given its large volume and the high mobility of its modules during stages of RNAPII recruitment and initiation, Mediator offers a potential of interaction surfaces way beyond that of the CTD or even the complete RNAPII (29). Further investigations on the biology of T bodies may therefore not only aid in dissecting different steps of the Ty1 viral assembly pathway but also help us to better understand the molecular cross talk between transcription processes and the downstream functions of mRNP.
We thank David J. Garfinkel, Domenico Libri, Antonin Morillon, Dwight V. Nissley, Finn Skou Pedersen, and Manfred Schmid for critical reading of the manuscript; Roy Parker, David J. Garfinkel, Jayne L. Brookman, Maureen Mee, and Andres Aguilera for providing reagents; and Rikke Jespersen for excellent technical assistance.
This work was supported by the Danish National Research Foundation (Grundforskningsfonden) and the Novo Nordisk Foundation. F.M. is the recipient of an Alfred Benzon Investigator Fellowship.
Published ahead of print on 4 August 2008.