|Home | About | Journals | Submit | Contact Us | Français|
The RNA-processing exosome is a complex of riboexonucleases required for 3′-end formation of some noncoding RNAs and for the degradation of mRNAs in eukaryotes. The nuclear form of the exosome functions in an mRNA surveillance pathway that retains and degrades improperly processed precursor mRNAs within the nucleus. We report here that the nuclear exosome controls the level of NAB2 mRNA, encoding the nuclear poly(A)+-RNA-binding protein Nab2p. Mutations affecting the activity of the nuclear, but not the cytoplasmic, exosome cause an increase in the amount of NAB2 mRNA. Cis- and trans-acting mutations that inhibit degradation by the nuclear-exosome subunit Rrp6p result in elevated levels of NAB2 mRNA. Control of NAB2 mRNA levels occurs posttranscriptionally and requires a sequence of 26 consecutive adenosines (A26) in the NAB2 3′ untranslated region, which represses NAB2 3′-end formation and sensitizes the transcript to degradation by Rrp6p. Analysis of NAB2 mRNA levels in a nab2-1 mutant and in the presence of excess Nab2p indicates that Nab2p activity negatively controls NAB2 mRNA levels in an A26- and Rrp6p-dependent manner. These findings suggest a novel regulatory circuit in which the nuclear exosome controls the level of NAB2 mRNA in response to changes in the activity of Nab2 protein.
The synthesis of mRNAs in eukaryotes requires processing of precursor mRNAs (pre-mRNAs) prior to transport of the mature RNAs to the cytoplasm, where they participate in protein synthesis and undergo regulated degradation. The RNA-processing reactions required to produce a functional mRNA include cotranscriptional 5′-end capping, splicing to remove introns, and cleavage and polyadenylation to produce the mature 3′ ends. These processing steps also play a critical role in the transport of mRNAs to the cytoplasm, as demonstrated by the fact that defects in any of these steps decrease the efficiency of mRNA export (31, 33). Defects in splicing and 3′-end formation result in retention of transcripts in nuclear foci that appear to represent the sites of transcription (10, 20). Defects in mRNA transport in Saccharomyces cerevisiae cause the formation of similar nuclear foci, suggesting a functional dependence between mRNA processing and transport, as well as between the completion of transcription and/or the release of transcripts from their sites of synthesis (22, 23). The failure of these processing and transport events also leads to degradation of nuclear transcripts by a recently discovered nuclear mRNA surveillance pathway that features a conserved complex of riboexonucleases called the exosome (7).
The exosome consists of a core complex containing nine 3′-5′ riboexonucleases that functions in the nuclei and the cytoplasm of all eukaryotic organisms. In the cytoplasm of yeast, the exosome interacts with the Ski complex, composed of Ski2p, Ski3p, and Ski8p, and functions in a minor 3′-5′ degradation pathway for normal and nonsense-sequence-containing mRNAs and in a pathway that degrades mRNAs lacking stop codons (14, 26, 35). In mammalian cells, the cytoplasmic exosome appears to function in the major mRNA degradation pathway, which contrasts with the predominant role of the 5′-3′ pathway of mRNA degradation in yeast (23a, 26a, 37). Experiments with mammalian cells indicate that RNA-binding proteins may play a critical role in the degradation of certain normal mRNAs by the cytoplasmic exosome. Specifically, mRNAs carrying destabilizing AU-rich elements (AREs) in their 3′ untranslated regions (3′ UTRs) require ARE-binding proteins to activate transcript degradation by the exosome (8).
The exosome also exists as a distinct nuclear complex distinguished by the presence of the nuclear 3′-5′ riboexonuclease Rrp6p (1, 5, 7). A significant fraction of the nuclear exosome exists in the nucleolus, where it plays a role in the processing of rRNA precursors (7). Experiments with yeast show that the exosome degrades the 5′ external transcribed spacer released by an early endonucleolytic cleavage of the 35S pre-rRNA (1). The exosome produces the mature 3′ end of 5.8S rRNA in a two-step reaction in which the core exosome and the putative helicase Mtr3p/Dob1p trim the 7S precursor to within approximately 30 nucleotides of the mature 3′ end, followed by exonucleolytic removal of the remaining tail by Rrp6p (1, 5). The exosome also plays a role in the processing of 3′-end-extended precursors of snoRNAs and snRNAs (1, 36). Similar to what occurs in the processing of 5.8S pre-rRNA, the core exosome trims the 3′ end of snRNAs and snoRNAs, followed by removal of the last several nucleotides by Rrp6p. Thus, the core exosome and Rrp6p appear to play temporally distinct roles in the 3′-end formation of these noncoding RNAs.
The nuclear exosome appears to function in an mRNA surveillance system that degrades pre-mRNAs that fail to complete the splicing, 3′-end processing, and transport steps of mRNA biogenesis. The first clue to the existence of this nuclear mRNA surveillance system came from the discovery of RRP6 as a suppressor of a conditional defect (pap1-1) in poly(A) polymerase (Pap1p) (5, 6). Mutations in RRP6 allow the growth of pap1-1 strains at the nonpermissive temperature, which presumably reflects an increase in the levels of poly(A)+ mRNA and the release of transcripts from nuclear foci in the suppressor strains (6, 20). Likewise, deletion of RRP6 (rrp6-Δ) increases the levels of cyc1-512 mRNAs, which fail to accumulate to normal levels due to a deletion of the CYC1 polyadenylation site (11). Defects in nuclear-exosome components also suppress the growth and mRNA accumulation defects caused by the mutation of Rna14p, a factor required for the cleavage and polyadenylation steps of mRNA 3′-end processing (24, 34). In these cells, the core exosome appears to trim the 3′ ends of extended transcripts to a position near the poly(A) site, but in this case, Rrp6p degrades the remainder of the transcript, or the mRNA undergoes polyadenylation at or near the normal site (34). Thus, 3′ trimming of aberrant, extended transcripts by the nuclear exosome may give rise to a normally polyadenylated mRNA if poly(A) polymerase is active.
These findings suggested that the nuclear exosome possesses the ability to play positive and negative roles when mRNA 3′-end formation fails. While experiments indicated a role for the exosome in the degradation of mRNAs suffering from cis- or trans-acting defects in 3′-end processing, there was no evidence that the exosome plays a role in controlling the level of normal mRNA. We report here that the nuclear exosome controls the level of NAB2 mRNA, which encodes a nuclear poly(A)+-mRNA-binding protein, Nab2p. This regulation requires an A26 sequence in the 3′ UTR of NAB2 that causes inefficient 3′-end formation and sensitizes the transcript to degradation. Moreover, Nab2p negatively regulates NAB2 mRNA levels in an A26-dependent manner, suggesting that Nab2p and the nuclear exosome cooperate to control the levels of NAB2 mRNA.
Yeast media and experimental reagents were prepared by standard protocols (18). Yeast strains are described in Table Table1.1. Oligonucleotides are listed in Table Table2.2. Chemicals were obtained from Sigma, U.S. Biological, or Fisher.
Oligonucleotides used to construct the plasmids in this study are described in Table Table2.2. Plasmids expressing various gene sequences under the control of the MET17 promoter were constructed by either direct cloning or Gap repair (28) of PCR-amplified DNA into pGFP-N-FUS (27), as described below. pGFPNAB2 was constructed by inserting a PCR product (primers OSB240 and OSB218) into the XbaI and ClaI sites of pGFP-N-FUS. pGFPHSP12 was constructed by inserting a PCR product synthesized with primers OSB187 and OSB188 into the XbaI and ClaI sites of pGFP-N-FUS. The NAB2 open reading frame (ORF) synthesized by PCR with primers OSB29 and OSB30 was cloned into the EcoRI and BamHI sites of pGAD424 to create pGAD424-NAB2 as described previously (5). The following plasmids were constructed by inserting PCR products into the indicated plasmid sites by Gap repair. pGFPNAB2-Δ3′UTR was constructed by inserting a PCR product synthesized with primers OSB271 and OSB270 into the XbaI site of pGFP-N-FUS. pGFPNAB2-C18 was constructed by inserting a PCR product synthesized with primers OSB297 and OSB296 into the SpeI site of pGFPNAB2-Δ3′UTR. pGFPNAB2-U26 was constructed by inserting a PCR product into the SpeI site of pGFPNAB2-Δ3′UTR; the inserted PCR product was generated using the primer pair OSB316-OSB296 to amplify hybridized templates produced by PCR with the primer pairs OSB316-OSB315 and OSB314-OSB296. pGFPNAB2-C26 was constructed by inserting a PCR product into the SpeI site of pGFPNAB2-Δ3′UTR; the inserted PCR product was generated using the primer pair OSB316-OSB296 to amplify hybridized templates produced by OSB316-OSB327 and OSB326-OSB296. pGFPNAB2-G18 was constructed by inserting a PCR product into the SpeI site of pGFPNAB2-Δ3′UTR; the inserted PCR product was generated using the primer pair OSB298-OSB299 to amplify a template produced by hybridization of OSB62-OSB63, followed by PCR amplification of the product with OSB295-OSB296. YEpNAB2 was constructed by amplifying NAB2 with OSB256 and OSB262 by PCR, hydrolyzing the ends of the product with BamHI and KpnI, and inserting the NAB2 fragment into the BamHI and KpnI sites of YEplac181 (16).
Total RNA was isolated from yeast strains grown to an A600 of 0.8 to 1.2 as described previously (29), and Northern blot analysis was carried out as described in reference 5. Oligonucleotides used for detection of RNA in Northern blot analyses are described in Table Table2.2. The procedure for generating [γ-32P]ATP-labeled DNA oligonucleotide probes is described in reference 5. 7S rRNA and ACT1 mRNA were detected with DNA oligonucleotides OSB157 and OSB184, respectively, whose 5′ ends were labeled with 32P. The 5′ ends of randomly primed hexamer probes were generated with a Random Primers DNA labeling system (Invitrogen) and labeled with [α-32P]CTP. GFP-NAB2 fusion transcripts were detected with 32P-labeled, randomly primed probes generated from a PCR template generated with primers OSB272 and OSB273 by using pGFP-N-FUS as a template. NAB2 mRNA was detected with 32P-labeled, randomly primed probes generated from a BamHI- and EcoRI-digested NAB2-containing fragment of pGAD424-NAB2.
Western blotting to determine Nab2p levels was carried out as described previously (6). Cells were grown in yeast extract-peptone-dextrose to an A600 of 1.0 at 30°C. Nab2p was detected with monoclonal antibody 3F2 (1:500 dilution) (3). Pgk1p was detected with a polyclonal antibody (A6457, 1:200 dilution; Clontech).
We analyzed two-color cDNA microarray data that compared global transcript levels in normal and rrp6-Δ cells (21). Comparison of transcript levels in the two strains with a compendium of reference data from the normal strain facilitated the identification of false-positive results arising as a result of normal transcript fluctuations in the control cells (21). Duplicate analyses with a correlation coefficient of 0.91 indicated increased or decreased expression of 45 transcripts in the rrp6-Δ strain relative to expression in the normal strain. No significant relationships emerged from an analysis of the common properties of these genes. We focused our initial attention on NAB2, the gene encoding a nuclear poly(A)+-mRNA-binding protein, Nab2p, whose mRNA is increased threefold in rrp6-Δ cells. Our interest in this gene arose from its role in nuclear polyadenylation and transport of mRNAs to the cytoplasm, both of which respond to defects in RRP6 (7, 17, 19, 20).
Northern blot analysis of NAB2 mRNA levels in RRP6 and rrp6-Δ strains revealed a three- to fivefold increase in transcript levels in the mutant strain compared to the level of ACT1 mRNA, which showed no change in the microarray analysis (21) (Fig. (Fig.1A).1A). The increase in NAB2 mRNA levels in the rrp6-Δ strain is reflected by an increase in the amount of Nab2 protein, as determined by Western blot analysis (Fig. (Fig.1B).1B). The NAB2 Northern blot reveals several longer products in addition to the major NAB2 transcript, suggesting inefficient 3′-end processing of the pre-mRNA (Fig. (Fig.1A1A).
Next, we analyzed NAB2 mRNA levels in strains with mutations affecting components of the nuclear and cytoplasmic exosomes to determine if NAB2 transcript levels respond to defects in these components (Fig. (Fig.2A).2A). Northern blot analysis of NAB2 mRNA levels included, as a control, quantitation of 7S rRNA, a precursor to 5.8S rRNA that increases in cells with defects in nuclear, but not cytoplasmic, exosome function (1, 5, 12). NAB2 mRNA levels increase in strains with mutations affecting the core exosome (rrp4-1) and the nuclear exosome (rrp6-Δ) but not the cytoplasmic exosome (ski3-Δ). Analysis of NAB2 mRNA levels in the mtr4-1 strain at the permissive temperature (25°C) (Fig. (Fig.2A)2A) and at the nonpermissive temperature (37°C) (data not shown) revealed only modest increases in NAB2 mRNA levels. Effects of the nuclear exosome on the amount of NAB2 mRNA imply control by a degradation mechanism operating in the nucleus.
We constructed a test system to determine whether the increases in NAB2 mRNA levels result from direct effects on mRNAs levels after transcription (Fig. (Fig.3A).3A). The test involves placing the coding sequence and the 3′ UTR of the gene in question between a heterologous promoter (MET17) and the strong CYC1 poly(A) site (MET17 and CYC1 mRNAs show no change in abundance in rrp6-Δ cells, as measured by microarray or Northern blot analysis [data not shown]). Fusion of the coding sequence to the ORF of green fluorescent protein (GFP) creates a chimeric gene. An inserted sequence containing a weak poly(A) site in its 3′ UTR has the potential to give rise to two transcripts, one ending in the inserted 3′ UTR and one ending at the CYC1 poly(A) site (Fig. (Fig.3A).3A). The relative levels of these two transcripts in RRP6 and rrp6-Δ backgrounds should change if the inserted 3′ UTR confers sensitivity to Rrp6p. For example, a mRNA ending at a weak poly(A) site might be sensitive to the level of Rrp6p, while a mRNA ending at the CYC1 site would not.
Insertion of the NAB2 sequence from codon 2 to a position 390 bp past its stop codon created a GFP-NAB2 fusion sequence under the control of the MET17 promoter. This chimeric gene produces a transcript ending at the CYC1 site in an RRP6 strain and low levels of a second transcript ending in the NAB2 3′ UTR (Fig. (Fig.3B,3B, lane 3). The absence of RRP6 enhances by three- to fivefold the amount of the transcript ending in the NAB2 3′ UTR without significantly affecting the level of the longer transcript ending at CYC1 (Fig. (Fig.3B,3B, lane 2). This result suggests that RRP6 specifically limits the levels of mRNAs ending at the 3′ end of NAB2 and that this effect occurs independently of the NAB2 promoter.
The role of Rrp6p and the nuclear exosome in controlling NAB2 mRNA levels implies degradation of the transcript in a 3′-to-5′ direction. Accordingly, we investigated the role of 3′ exonucleolytic degradation by inserting a stretch of 18 consecutive G residues (G18 block) 12 bp 3′ of the NAB2 stop codon, which lies 141 bp 5′ of the A26 polyadenylation site (see below; Fig. Fig.4A).4A). Previous studies showed that this sequence inhibits exonucleolytic degradation of RNAs in yeast but that a C18 block does not (13). The G18 block, but not the C18 block, enhances the accumulation of the transcript ending in the NAB2 3′ UTR in a normal strain, consistent with the view that the transcript is degraded from its 3′ end by the exosome (Fig. (Fig.4B,4B, lanes 1and 2). Combination of the G18 block and rrp6-Δ results in NAB2 mRNA levels higher than that produced by either mutation alone (Fig. (Fig.4B,4B, lane 6). The additional effect of the G18 block in the absence of Rrp6p most likely results from slowing NAB2 mRNA degradation by the remaining components of the nuclear exosome, such as Rrp4p.
In a complementary approach, we measured NAB2 mRNA levels in strains with mutations that differentially affect the 3′-end formation and degradation functions of the nuclear exosome (30). The rrp6-3 mutation, in the exonuclease domain of Rrp6p, abolishes the ability of the exosome to properly form the 3′ ends of 5.8 S rRNA and snoRNAs, as well as its ability to degrade the 5′ external transcribed rRNA spacer and poly(A)+ snoRNAs. In contrast, the rrp6-13 mutation, in the HRDC region of Rrp6p, causes defects in 3′-end formation but has no effect on degradation by the nuclear exosome. Analysis of NAB2 mRNA levels in cells with these mutations revealed elevated levels of the transcript and rrp6-3 cells, but not rrp6-13 cells (Fig. (Fig.4C).4C). These results suggest that the degradation activity of the nuclear exosome limits the concentration of NAB2 mRNA.
S. cerevisiae NAB2 contains a remarkable sequence of 26 consecutive adenosines (A26) 141 nucleotides 3′ of its stop codon (Fig. (Fig.5A).5A). Comparison of the NAB2 3′ UTRs from other species of Saccharomyces (9) reveals conservation of consecutive adenosines at very similar distances from the stop codons of the ORFs (Fig. (Fig.5A).5A). These similarities are consistent with a conservation of function for the adenosine-rich sequences in these organisms.
Based on sequence analysis of cDNAs, the mature 3′ end and poly(A) tail of NAB2 mRNA appears to occur within the A26 sequence (3). However, oligo(dT)-directed cDNA synthesis may lead to incorrect assignment of the polyadenylation site since priming of reverse transcription may occur at the template-encoded A26 sequence. We analyzed the site of polyadenylation using ligation-mediated poly(A)-tail analysis, a method that avoids artifacts arising from the priming of reverse transcription-PCR at poly(A) sequences internal to the NAB2 transcript and allows the determination of the poly(A) site and poly(A) tail length (32). We cloned the products of the entire NAB2 ligation-mediated poly(A) tail analysis reaction from the RRP6 strain into plasmids and analyzed the DNA sequences of individual clones. The results showed that 15 of 26 clones analyzed contained sequence from the NAB2 region and that 11 of 26 contained unrelated sequences. Nine of the 15 sequences from the NAB2 region end within the A26 sequence, while 6 contained more than the template-encoded 26 adenosines (data not shown). These findings confirm that polyadenylation of the NAB2 mRNA occurs within, or immediately adjacent to, the A26 sequence in the NAB2 3′ UTR.
We tested the requirement for the A26 sequence in the control of NAB2 mRNA levels by Rrp6p by precisely deleting the 3′ UTR of NAB2 from the test construct (Fig. (Fig.5)5) and analyzing the levels of transcript produced. The result shows that removal of the NAB2 3′ UTR results in a transcript ending only at the CYC1 polyadenylation site (Fig. (Fig.5B,5B, lanes 3 and 4). The deletion of the 3′ UTR also increases the amount of transcript fivefold and renders it insensitive to Rrp6p (Fig. (Fig.5B).5B). These results suggest that 3′-end formation of NAB2 and its negative regulation by Rrp6p require the 3′ UTR of NAB2.
Next, we tested the requirement for the A26 sequence in the control of NAB2 mRNA levels by precisely replacing it with the same number of U or C residues. Both replacements result in increased levels of mRNAs ending in the NAB2 3′ UTR, suggesting that the presence of the A26 sequence normally inhibits 3′-end formation in this region (Fig. (Fig.5C).5C). The replacements also abolish the sensitivity of NAB2 mRNA levels to Rrp6p, indicating that control of NAB2 mRNA levels by the exosome requires the presence of the A26 sequence in the 3′ UTR.
The presence of the NAB2 A26 sequence in its 3′ UTR and the ability of Nab2p to bind to poly(A) sequences prompted us to examine the effect of Nab2p activity on the level of its own mRNA. First, we expressed Nab2p from a high-copy-number plasmid (YEpNAB2) in an rrp6-Δ strain and measured its effect on the expression of GFP-NAB2 mRNA from our test plasmid. The results show that the presence of the Nab2p-expressing plasmid, but not the control plasmid (YEp), causes a decrease in the amount of the mRNA ending in the NAB2 3′ UTR but not in the amount of the mRNA ending at the CYC1 poly(A) site (Fig. (Fig.6A,6A, lanes 1 to 2). This effect requires the presence of the A26 sequence, as its replacement with U26 or C26 renders the transcript ending in the NAB2 3′ UTR insensitive to the expression of Nab2p from the high-copy-number plasmid (Fig. (Fig.6A,6A, lanes 3 to 6).
Next we examined the effect of the nab2-1 mutation on NAB2 mRNA levels. This mutation deletes amino acids 4 to 97 and results in a protein (Nab2ΔNp) defective in the nuclear export of mRNA (25). This deletion results in the synthesis of a shorter mRNA (nab2-1 mRNA) that accumulates at a fivefold-higher level than that of the normal NAB2 mRNA (Fig. (Fig.6B).6B). The nab2-1 defect results in a significantly larger increase in NAB2 mRNA levels than deletion of Rrp6p, suggesting that Nab2p may exert its negative effect by interacting with components of the exosome in addition to Rrp6p. Since both Nab2p and Rrp6p have A26-dependent negative effects on the levels of NAB2 mRNA, we analyzed the effect of a nab2-1 rrp6-Δ double mutation. The results show that deletion of RRP6 does not enhance the negative effect of the nab2-1 mutation on its own mRNA levels (Fig. (Fig.6,6, compare lanes 2 and 4), suggesting that the ability of Rrp6p to control the level of NAB2 mRNA requires Nab2p activity. On the other hand, the loss of Nab2p activity caused by nab2-1 increases NAB2 mRNA levels twofold in the absence of Rrp6p (Fig. (Fig.6B,6B, compare lanes 3 and 4). This effect suggests that Nab2p may recruit the exosome, albeit inefficiently, to degrade its own mRNA in the absence of Rrp6p.
Investigations of the function of the nuclear form of the exosome indicate that it participates in a surveillance pathway that degrades mRNAs that fail to complete mRNA-processing steps due to defects in processing factors or the mRNA sequences required by these factors (4, 6, 11, 34). Experiments presented here provide the first evidence that the nuclear exosome participates in posttranscriptional regulation of a normal mRNA encoding the nuclear poly(A)+-RNA-binding protein Nab2p. First, defects in activities required for the function of the nuclear exosome, but not those required solely for the activity of the cytoplasmic exosome, result in increased levels of NAB2 mRNA (Fig. (Fig.11 and and2).2). Second, this regulation occurs with a chimeric NAB2 transcript synthesized from a heterologous promoter, indicating that control most likely occurs at the level of the mRNA (Fig. (Fig.3).3). Third, a cis-acting sequence that inhibits riboexonucleases or trans-acting mutations that inhibit exonucleolytic degradation by the nuclear-exosome subunit Rrp6p result in elevated levels of NAB2 mRNAs, suggesting that the nuclear exosome controls NAB2 mRNA by degradation of the transcript (Fig. (Fig.4).4). Fourth, control of NAB2 mRNA levels by the nuclear exosome requires an unusual sequence of 26 consecutive adenosine residues in the 3′ UTR of the transcript (Fig. (Fig.5).5). Removal or replacement of this sequence abolishes control by the exosome and results in a significant increase in the amount of NAB2 mRNA, implicating the A26 sequence as a cis-acting negative-control element (Fig. (Fig.5).5). Presumably, replacement of the A26 sequence with U26 or C26 enhances mRNA 3′-end formation in the NAB2 3′ UTR by a mechanism that awaits experimental clarification. Finally, expression of Nab2p from a high-copy-number plasmid causes an A26-dependent decrease in the level of a chimeric GFP-NAB2 reporter, and a NAB2 mutation (nab2-1) results in elevated levels of its mRNA (Fig. (Fig.6).6). These findings suggest a novel autoregulatory mechanism in which Nab2p controls its own levels by interacting with the A26 sequence in its 3′ UTR and causing degradation of the transcript. The negative effect of Nab2p on its own mRNA levels appears to involve the nuclear exosome since Rrp6p also negatively affects NAB2 mRNA levels in an A26-dependent manner. Moreover, the double nab2-1 rrp6-Δ mutation does not cause an additive increase in NAB2 mRNA levels, suggesting that Nab2p and Rrp6p act together in this pathway. The ability of Nab2p to limit the level of its own mRNA may explain the lack of a significant increase in NAB2 mRNA levels in polyploid strains of S. cerevisiae (15).
Several observations suggest that the control of NAB2 mRNA levels by Rrp6p and Nab2p may act at the mRNA 3′-end-formation step. We confirmed that polyadenylation of NAB2 mRNA occurs at the A26 sequence in the 3′ UTR. However, placement of the NAB2 3′ UTR containing this sequence in cis prior to the CYC1 polyadenylation site results in the majority of the transcripts ending at the CYC1 polyadenylation site in an RRP6 strain (Fig. (Fig.3B).3B). In the absence of Rrp6p and Nab2p, the level of transcripts ending in the NAB2 3′ UTR increases without significantly changing the number of transcripts ending at the CYC1 polyadenylation site. This increase suggests that Rrp6p and Nab2p enhance the degradation of transcripts that fail to reach the CYC1 polyadenylation site or that cleavage and polyadenylation occurs at the NAB2 site and that the transcripts thus formed are unusually susceptible to degradation by the exosome. Degradation of NAB2 mRNA by Rrp6p requires the A26 sequence, but the sequence also inhibits the ability of the NAB2 3′ UTR to function in 3′-end formation in vivo. Removal of this element, or its replacement by C26 or U26, results in efficient 3′-end formation in the NAB2 3′ UTR and abolishes the production of transcripts ending at the downstream CYC1 polyadenylation site (Fig. (Fig.5).5). These findings suggest that the presence of the A26 sequence in the NAB2 3′ UTR results in inefficient 3′-end processing and sensitizes the transcripts to the negative effects of Nab2p and the nuclear exosome.
We propose that the control of Nab2p levels results from an autoregulatory mechanism in which inefficient 3′-end processing of NAB2 mRNA results in degradation of the transcript by the nuclear exosome. In this model, the binding of Nab2p to the A26 sequence in the 3′ UTR of its own mRNA would have either of two consequences. In one case, the binding of Nab2p would inhibit cleavage and polyadenylation of the transcript, thereby allowing degradation of the transcript by the exosome. In the second case, weak cleavage and polyadenylation of NAB2 mRNA would occur independently of Nab2p interaction, but Nab2p binding to the A26 sequence would accelerate NAB2 pre-mRNA degradation by the exosome. Our findings support the second mechanism, because NAB2 mRNA appears to have a weak polyadenylation site and because high-copy-number expression of Nab2p decreases the level of transcripts ending at NAB2, without significantly increasing the level of transcripts ending at the downstream CYC1 site in our test plasmid (Fig. (Fig.6).6). Moreover, Nab2p defects result in the hyperadenylation of mRNAs in vitro, and the addition of Nab2p to cell-free polyadenylation reaction mixtures inhibits hyperadenylation (19). Thus, Nab2p may function with a nuclease to limit the initial length of mRNA poly(A) tails synthesized in the nucleus. The ability of these proteins to initiate and then halt poly(A) trimming probably depends on the length of the poly(A) tail and the identity of the proteins bound to it and to the 3′ UTR of the transcript. We suggest that NAB2 pre-mRNAs that fail to undergo normal cleavage and polyadenylation may lack such factors. As a result, the binding of Nab2p to the A26 in the NAB2 3′ UTR may initiate uncontrolled 3′-5′ degradation of NAB2 mRNA by the exosome. The ability of Nab2p to activate the degradation of its own mRNA is reminiscent of the ability of hnRNP proteins to activate exosomal degradation of ARE mRNAs and may reflect a more general role for Nab2p and the exosome in mRNA 3′-end processing (26a). Experiments to test directly the effects of Nab2p and the exosome on NAB2 3′-end processing in vitro are under way.
We are grateful to Lynne Maquat and the members of the Butler and Sherman labs for helpful discussions and comments on the manuscript. We thank Anita Corbett, Roy Parker, and Maurice Swanson for providing strains and antibodies and Dan Shoemaker and Tim Hughes for providing data prior to publication.
This work was supported by a grant (GM59898) from the NIH.