In a screen for factors involved in mRNA turnover, four temperature-sensitive yeast strains (ts1189, ts942, ts817, and ts1100) exhibited defects in the decay of several mRNAs. Complementation of the growth and mRNA decay defects, and genetic experiments, revealed that ts1189 is mutated in the previously unknown MRT4 gene, ts942 is mutated in GRC5 (encoding the L9 ribosomal protein), ts817 contains a mutation in SLA2 (encoding a membrane protein), and ts1100 contains a mutation in THS1 (encoding the threonyl-tRNA synthetase). Three of the four mutants (mrt4, grc5, and sla2) were not defective in protein synthesis, suggesting that these strains contain mutations in factors that may play a specific role in mRNA decay. The mRNA stabilization observed in the ths1 strain, however, could be due to the significant drop in translation observed in this mutant at 37°. While the three interesting mutants appear to encode novel mRNA decay factors, at least one could be linked to a previously characterized mRNA decay pathway. The growth and mRNA decay defects of ts942 (grc5) cells were suppressed by overexpression of the NMD3 gene, encoding a protein shown to participate in a two-hybrid interaction with the nonsense-mediated decay protein Upf1p.
THE rate at which a specific mRNA is degraded can be a significant contributing factor to the overall expression levels of the protein encoded by that mRNA, with short-lived transcripts particularly susceptible to rapid, regulated changes in gene expression. The mechanisms of mRNA turnover have been the subject of intensive study, particularly in the yeast Saccharomyces cerevisiae (Caponigro and Parker 1996; Jacobson and Peltz 1996), and have led to the delineation of at least three mRNA decay pathways. In two of these, 5′ → 3′ exonucleolytic digestion of the mRNA follows a decapping event (Decker and Parker 1993, 1994), whereas in the third pathway, the transcript is degraded by 3′ 5′ exonucleolytic digestion (Muhlradet al. 1995). The initiation of decapping is known to be promoted by shortening the mRNA poly(A) tail or by premature translational termination (Muhlrad and Parker 1994; Muhlradet al. 1994). A number of trans-acting factors that function in these pathways have been identified and include the Xrn1p 5′ → 3′ exonuclease (Hsu and Stevens 1993); the Dcp1p decapping enzyme (Beelmanet al. 1996); the poly(A)-binding protein, Pab1p (Adamet al. 1986; Sachset al. 1986); two proteins that regulate decapping (Mrt1p and Mrt3p; Hatfieldet al. 1996); and several factors required for the decay of mRNAs with premature termination codons (Upf1p, Nmd2p/Upf2p, and Upf3p; Leeds et al. 1991, 1992; Peltzet al. 1994; Cuiet al. 1995; He and Jacobson 1995; Lee and Culbertson 1995; Heet al. 1997).
In addition to roles for specific factors, there is also abundant evidence for an important role for translation in mRNA decay. Inhibition of translational elongation has been shown to stabilize mRNAs (Peltzet al. 1992; Beelman and Parker 1994) while premature translation termination enhances mRNA decay rates (Peltzet al. 1994). Instability elements have been localized to mRNA coding regions (Parker and Jacobson 1990; Jacobson and Peltz 1996), and the activity of some of these depends on ribosome translocation up to or through the element (Parker and Jacobson 1990; Peltzet al. 1993a; Hennigan and Jacobson 1996). Other sequences implicated in mRNA decay, including the poly(A) tail and some 3′-UTR instability elements, have been found to influence protein synthesis (Kruyset al. 1989; Marinxet al. 1994; Jacobson 1996; Sachset al. 1997). Finally, factors involved in the decay process have been shown to be polysome associated (Peltzet al. 1993b; Caruccio and Ross 1994; Atkin et al. 1995, 1997; Zhanget al. 1997; Mangus and Jacobson 1999).
Despite the complexity of the mRNA decay process and the extensive link with protein synthesis, until recently only three mRNA decay factors (Pab1p, Mrt1p, and Mrt3p) appeared to be encoded by essential genes (Adamet al. 1986; Sachset al. 1986; Hatfieldet al. 1996). This suggested that the complement of factors involved in mRNA decay may include additional factors encoded by essential genes and led us to screen ∼4000 temperature-sensitive yeast strains for those harboring defects in mRNA decay (Zuk and Jacobson 1998). In this screen, we identified five temperature-sensitive (ts) yeast mutants that stabilized several different mRNAs at the nonpermissive temperature. We previously described the ts1159 strain, showing that its mutation in TIF51A implicates the putative translation factor elF-5A in the regulation of 5′ → 3′ exonucleolytic decay (Zuk and Jacobson 1998).
Here we describe the initial characterization of the remaining four strains, ts1189, ts942, ts817, and ts1100, which respectively contain mutations in MRT4 (encoding a previously uncharacterized protein), GRC5 (encoding a 60S ribosomal subunit protein; Dicket al. 1997; Eisingeret al. 1997; Nikaet al. 1997), SLA2 (encoding a membrane protein; Holtzmanet al. 1993; Rathset al. 1993; Naet al. 1995), and THS1 (encoding the threonyltRNA synthetase; Pape and Tzagoloff 1985).
MATERIALS AND METHODS
Yeast strains and media: The collection of temperature-sensitive mutants used in this study was generated from strain SS330 (MATa ade2-101 his3-200 tyr1 ura3-52) by EMS mutagenesis as described in Vijayraghavan et al. (1989). A related strain, SS328 (MATα ade2-101 his3-200 lys2 ura3-52) was used for crosses. All other strains, media, and growth conditions were described previously (Zuk and Jacobson 1998).
Temperature shifts and measurement of mRNA decay rates: mRNA decay rates in cells incubated at 37° for 1 hr were measured after inhibition of transcription with thiolutin (TL), as described previously (Zuk and Jacobson 1998). RNA was isolated by the hot phenol method (Herricket al. 1990) and subjected to Northern blot analysis as detailed previously (Zuk and Jacobson 1998).
Complementation by the YCp50 library, open reading frame subclones, and high-copy NMD3: The Ts– strains were electroporated (Becker and Guarente 1991) with 1 gμg (0.7–1 × 104 plasmids) of a YCp50 library (Roseet al. 1987). The electroporated cells were grown on SC-ura plates containing 1 m sorbitol at 37° for 5 days. The complementing plasmids were isolated from yeast cells and amplified in Escherichia coli using standard procedures (Sambrooket al. 1989). The resulting plasmids were transformed into the Ts– cells using the modified PEG-LiAc procedure (Soniet al. 1993) and shown to retain complementation of the growth defect. All the complementing plasmids were cut with EcoRI and SalI (which flank the insert in the YCp50 vector) to generate a preliminary map of the inserts. ts817, ts942, and ts1189 were each complemented by one plasmid (817-1, 942-1, and 1189-1, respectively), while ts1100 was complemented by two different plasmids (1100-9 and 100-11) that were subsequently found to contain the same DNA fragment in opposite orientations. Complementing DNAs were sequenced with the SequenaseII kit (BRL) using a primer (5′ CTTCGCTACTTGGAGCCACT 3′) complementary to a region 40 nt upstream of the BamHI site of YCp50 into which the library inserts had been cloned (all primers were made by Operon Technology Inc., Alameda, CA). The sequence obtained was compared to the Saccharomyces Genome Database at Stanford University (http://genome-www.stanford.edu/-Saccharomyces/) using the GCG BLAST program (Altschulet al. 1990).
The insert in 817-1 contained four open reading frames (ORFs), yNL245c, SUI1, SLA2, and yNL242w (using the ORF numbers assigned by the genome sequencing project for all unknown genes). Plasmid 942-1 contained yLR072w, yLR073c, yLR074c, GRC5, and yLR076c. The plasmids that complemented ts1100 contained yIL079c, THS1, yIL077c, and yIL076w. Plasmid 1189-1 contained yKL010c, yKL009w (MRT4), yKL008c, and CAP1. Using the sequence information, we excised fragments containing a single ORF plus enough sequence of flanking regions to overlap with neighboring ORFs from the YCp50 plasmids and subcloned the fragments into the URA3-containing vector pRS316 (Sikorski and Hieter 1989). Each subclone was transformed into the relevant Ts– strain as described above. Eight clones from each transformation experiment were patched onto SC-ura plates and grown at 24° for 3 days. These plates were replica-plated to two SC-ura plates, one of which was kept at 24° and the other of which was incubated at 37° for an additional 3 days to test for complementation of the growth defect. As a negative control, cells were also transformed with the empty pRS316 vector and analyzed in the same manner.
ts817 was complemented by a plasmid expressing the SLA2 ORF (pDZ57) containing a 4.2-kb BamHI/ClaI fragment of plasmid 817-1 cloned into the same sites of pRS316, including 769 nucleotides (nt) 5′ to the SLA2 ATG and 607 nt 3′ to the stop codon. ts942 was complemented by a plasmid expressing the GRC5 ORF (pDZ38) containing a 1.7-kb fragment of plasmid 942-1 (generated by PCR with primers 5′ CGGGATCCCT TGCCCAACAGCA 3′ and 5′ GGGGTACCCTCTTCACCGAA TATCTACCC 3′) cloned into the KpnI and BamHI sites of pRS316, including 493 nt 5′ to the GRC5 ATG and 543 nt 3′ to the stop codon. ts1100 was complemented by a plasmid expressing the THS1 ORF (pDZ28) containing a 3.5-kb EcoRI fragment of plasmid 1100-9 cloned into the same site of pRS316, including 1213 nt 5′ to the THS1 ATG and 129 nt 3′ to the stop codon. ts1189 was complemented by a plasmid expressing the MRT4 ORF (pDZ21) containing a 1.7-kb DraI fragment of plasmid 1189-1 cloned into the SmaI site of pRS316, including 377 nt 5′ to the MRT4 ATG and 637 nt 3′ to the stop codon.
Complementation of the RNA decay defect in cells harboring the different plasmids was assessed by comparing the half-lives of the CYH2 transcripts in wild-type and Ts– cells after a 1-hr shift to 37°. These experiments utilized Northern blot analyses performed as described above for the measurement of mRNA decay rates.
The YEp-HA-NMD3 construct contained the NMD3 promoter, a triple-HA tag (Kolodziej and Young 1991) inserted after the NMD3 ATG and fused to the rest of the ORF, and 3′ flanking sequences cloned into the YEp24 episomal vector (Botsteinet al. 1979). Complementation tests with this plasmid were performed as described above for the pRS316 constructs.
Strain construction and segregation analysis: To test the linkage of the ts817, ts942, and ts1100 mutants to the complementing genes, three marked strains were constructed, in which the wild-type allele of the relevant gene was linked to the URA3 marker. DNA fragments containing the complementing ORFs were cloned into the pJJ242 vector (containing yeast URA3 embedded into the multiple cloning site of pUC18; Jones and Prakash 1990), generating integrating plasmids in which the complementing gene was cloned upstream of the URA3 allele. The resulting plasmids [pDZ59 (SLA2), pDZ45 (GRC5), and pDZ58 (THS1)] were each linearized at a unique site in the cloned gene and transformed separately into the SS328 strain. Colonies capable of growth on SC-ura plates were isolated and subjected to Southern blot analysis (Sambrooket al. 1989) to verify intergration of the URA3 gene at the correct locus. Clones containing the correct alleles were crossed to the relevant Ts– mutants.
Diploids (selected on SC-ura, -lys, -tyr plates; see genotypes above) were isolated and grown in sporulation medium (1% KAc supplemented with 1% adenine) at 24° for 4 days. The resulting tetrads were digested with glusulase (1/10 dilution for 10 min at room temperature; New England Nuclear, Boston, MA) and dissected (Shermanet al. 1986). The spores were grown at 24° for 5 days and then tested for growth on rich medium at 24° and at 37°, for growth on SC-ura, SC-lys, SC-tyr media, and for the a and α mating types. True tetrads were defined as those that segregated 2:2 for growth on selective media and for mating type. Linkage was defined as Ts– and Ura–.
To generate the mrt4Δ strain, a DNA fragment containing the URA3 gene flanked on each end by 40 nt spanning the start and stop codons of the MRT4 ORF was amplified by PCR using the URA3-containing plasmid pJJ242 (Jones and Prakash 1990) and the primers MRT4URA3-5 (5′ GATGC CAAGATCAAAACGTTCCAAGCTAGTCACTTTAGCACGAT TCCGGTTTCTTTGAA 3′) and MRT4URA3-3 (5′ AGCTTAT TCCATGTTGATGTTAGTGCTTTCAACAGTGGAGCTTAC GACCGAGATTCCCG 3′). This DNA fragment was transformed into the SS330 × SS328 diploid strain, and clones heterozygous for the MRT4 locus (with one allele in which the MRT4 ORF was replaced with URA3) were identified by PCR. These clones were sporulated, dissected, and analyzed as described above.
PCR amplification of the MRT4 locus: The MRT4 locus was amplified from genomic DNA isolated from SS330 and ts1189 cells, using the primers 5′ GGGGTACCCCAAAATTTTCA CACCTGAG 3′ (335 nt 5′ to the ATG plus a KpnI site) and 5′ GCTCTAGACCAACTGACAGTACTCCAAC 3′ (378 nt 3′ to the stop codon plus an XbaI site). The resulting PCR products were cloned into the KpnI and XbaI sites of pRS316 and transformed into ts1189 cells. Complementation of the Ts– growth and mRNA decay phenotypes was tested as described above. The entire ORFs of both plasmids were sequenced with the SequenaseII kit and the sequences of the wild-type and Ts– alleles were compared directly.
Analysis of the Mrt4p sequence: The Mrt4p sequence was compared to the available databases (http://www.ncbi.nlm.nih.gov) using the BLAST algorithm (Altschulet al. 1990). The alignment of the Mrt4p sequence with those of Rpp0p and the product of F10E7.5 was generated by the Pileup program (Wisconsin Package version 9.1, Genetics Computer Group, Madison, WI). The output was produced by MacBoxshade.
Measurement of amino acid incorporation: Analysis of the incorporation of 35S-labeled amino acids was carried out after a shift to 37°, as described previously (Zuk and Jacobson 1998). Each experiment was repeated at least three times.
Polysome analysis: Extracts of cells grown at 24° or shifted to 37° for 1 hr were prepared and analyzed on sucrose gradients as detailed previously (Zuk and Jacobson 1998).
Primer extension: Total RNA samples (10 μg) were prepared from cells grown at 24° or shifted to 37° for 1 hr and analyzed essentially as described in He and Jacobson (1995). RNA was hybridized to an end-labeled oligonucleotide complementary to nts 351-327 of PGK1 mRNA by incubation at 80° for 4 min and 50° for 30 min in 10 μl hybridization buffer (300 mm NaCl, 10 mm Tris-HCl pH 7.6, and 2 mm EDTA pH 8.0). The RNA was reverse transcribed with AMV reverse transcriptase (RT; Boehringer Mannheim, Indianapolis) in 50 μl RT buffer (1 mm of each nucleotide, 10 mm DTT, 10 mm Tris pH 8.4, and 8 mm MgCl2) for 30 min at 42°. The reaction was stopped by the addition of 6 μl 1 m NaOH, incubation at 55° for 30 min, and subsequent addition of 6 μl 1 m HCl. Glycogen (30 μg; Boehringer Mannheim) was added and the samples were precipitated, resuspended in sequencing gel-loading buffer (98% formamide, 10 mm EDTA pH 8.0, 0.025% xylene cyanol FF, and 0.025% bromophenol blue), and run on a 6% polyacrylamide-urea gel alongside a sequence reaction of the PGK1 gene (from pRIPPGK1; Peltzet al. 1993a) done as described above, using the same primer. The gel was dried and subjected to autoradiography.
The four Ts– mutants have different mRNA decay defects: To characterize the mRNA decay defects observed in the four strains, we analyzed the turnover kinetics of several mRNAs after a shift of the strains to their nonpermissive temperature. The cells were grown at 24° to midlog phase and shifted to 37° for 1 hr. Transcription was then inhibited by the addition of thiolutin (TL; Jimenezet al. 1973). Decay rates of five transcripts were measured by probing Northern blots of RNA isolated at different times thereafter. The five RNAs represented different classes of transcripts found in the cell and included the CYH2 pre-mRNA and the CYH2, STE2, URA5, and PAB1 mRNAs. The CYH2 pre-mRNA is a substrate of the nonsense-mediated mRNA decay pathway (Heet al. 1993), and the STE2 and URA5 mRNAs have short half-lives (<7 min, see Table 1; Herricket al. 1990). In this strain, the CYH2 mRNA is a moderately stable transcript (t1/2 = 10 min) and the PAB1 mRNA has a relatively long half-life (20 min).
The results of these experiments are summarized in Table 1, and representative blots from which some of the half-lives (i.e., those of the CYH2 and URA5 transcripts) were derived are shown in Figure 1. The ts817 strain had the most extensive defect, with up to a fourfold increase in the half-lives of all the RNAs tested, suggesting a mutation in a factor involved in the decay of all mRNAs. All four mutants exhibited reductions in the decay rates of the unstable STE2 and URA5 mRNAs, but the stable PAB1 mRNA was not affected in ts942 and ts1100. This may be due to the fact that relatively small differences are easier to observe in mRNAs with short half-lives, or may indicate a more fundamental difference in the decay mechanism of the different transcripts. ts1100 also had an increased half-life of the CYH2 pre-mRNA, suggesting that this mutant is involved in the decay of both nonsense-containing mRNAs and other mRNAs.
Since all these strains were originally identified as having an elevated steady-state level of the CYH2 pre-mRNA it was odd that ts942 and ts1189 did not show substantially decreased decay rates of this RNA (Figure 1). This discrepancy was resolved by an experiment in which we grew these cells to different densities, shifted the cells to 37°, and analyzed the level of the CYH2 transcripts by Northern blotting. The results of this experiment indicated that there was an elevated level of the CYH2 pre-mRNA when the cells were maintained in early log phase, but not when the cultures were grown to a higher density (data not shown). Therefore, the increased steady-state level observed in these cells in the original blots may have been attributable to the density of the cultures and not a reflection of a defect in the decay of this transcript. However, ts942 cells did exhibit bona fide decreases in the decay rates of the STE2 and URA5 mRNAs, and the CYH2, STE2, and URA5 half-lives were increased by two- to sevenfold in ts1189; therefore these two mutants were also deemed worthy of further study.
Complementation of the Ts– mutants with cloned DNAs: The four Ts– strains were transformed with a yeast genomic DNA library, and plasmids that enabled temperature-independent growth in each strain were isolated. Cells from each Ts– strain, or from that strain harboring a complementing plasmid, were grown at 24° and 37°. Transformation with the relevant YCp50 clones was shown to restore growth at 37° to the mutant cells (data not shown). Growth at 37° was shown to be plasmid dependent by (i) retransforming the complementing plasmids back into the mutant cells and verifying the complemented phenotype and (ii) allowing the cells to lose the plasmids and demonstrating that all clones lacking the plasmid could no longer grow at 37° (data not shown).
The vector-insert junctions of the complementing YCp50 plasmids were sequenced and compared to the Saccharomyces Genome Database (http://genome-www.stanford.edu/-Saccharomyces/). Each plasmid contained several ORFs. These were then subcloned into the yeast centromere vector, pRS316 (Sikorski and Hieter 1989), along with enough flanking DNA to ensure inclusion of regulatory sequences. Each ORF construct was transformed into the relevant Ts– mutant and tested for complementation of the growth defect. In each case, only one ORF enabled the cells to grow at the nonpermissive temperature. The growth of the Ts– strains containing the complementing plasmid or the empty pRS316 vector is shown in Figure 2.
ts817 was complemented by the SLA2 gene (also called MOP2 and END4), previously identified in screens for factors involved in cytoskeletal membrane assembly, endocytosis of α-factor, and maintenance of the PMA1 ATPase (Holtzmanet al. 1993; Rathset al. 1993; Naet al. 1995). ts942 was complemented by expression of GRC5 (also called QSR1 andQM1), encoding the 60S ribosomal subunit protein L9, which plays a role in subunit joining (Dicket al. 1997; Eisingeret al. 1997; Nikaet al. 1997). THS1, the gene that complemented ts1100, encodes the cytoplasmic threonyl-tRNA synthetase (Pape and Tzagoloff 1985). ts1189 was complemented by a previously uncharacterized gene that we termed MRT4 (mRNA Turnover 4, yKL009w).
The Ts– growth phenotype and the mRNA decay defect in the mutants were complemented by the same genes: To demonstrate that the ORFs that complemented the Ts– growth defects in the mutant strains could also complement their mRNA decay defects, we measured the decay rates of the CYH2 transcripts in the four strains transformed with either the complementing ORF or with the empty pRS316 vector. The cells were grown to midlog phase at 24°, shifted to 37° for 1 hr, and then treated with thiolutin. RNA extracted from aliquots taken at various times thereafter was analyzed by Northern blotting using the CYH2 probe. The results in Figure 3 (with the calculated half-lives in Table 2) show that, in all cases, expression of the complementing ORF reduced the half-lives of both the CYH2 pre-mRNA and mRNA to wild-type levels or below (compare to the WT panel in Figure 1).
Even the ts942 cells, which did not originally show a severe defect in the decay of the CYH2 transcripts, exhibited increased decay of these RNAs in the presence of the wild-type GRC5 gene. To verify this result, however, we also tested the decay of the STE2 mRNA, since it was more significantly affected in the ts942 strain. This experiment showed that the half-life of the STE2 mRNA decreased twofold to WT levels (5 min) in cells containing the complementing plasmid (data not shown).
These results show that, in all four strains, the same ORF complemented both the Ts– and mRNA decay defects, suggesting that these two effects result from the same mutant alleles.
The defects in ts817, ts942, and ts1100 are respectively attributable to mutations in the SLA2, GRC5, and THS1 genes: To verify that the complementing genes were indeed those that were mutated in the ts817, ts942, and ts1100 strains, we constructed marked strains in which the URA3 gene was integrated adjacent to the respective wild-type locus (SLA2 for ts817, GRC5 for ts942, and THS1 for ts1100) in the temperature-independent SS328 parent strain. Each Ts– mutant was then crossed to the strain containing the relevant URA3-marked allele. The resulting diploids were sporulated, tetrads were dissected, and spores were analyzed for growth at 37° and for growth on media lacking uracil. Using this approach, if a Ts– strain is mutated in the marked gene then the URA3 marker will be linked to the wild-type allele of that gene so that essentially all Ts– spores will be Ura– and essentially all spores wild-type for growth will be Ura+.The results of these analyses, presented in Table 3, demonstrate the following: (i) all three crosses segregated 2:2 for temperature-sensitive growth, indicating that, in these mutants, this phenotype is the result of a defect in a single gene; (ii) 10 of 10 dissected tetrads from the ts817 cross exhibited cosegregation of the Ts– and Ura– phenotypes, indicating linkage of the defect with SLA2; (iii) of 21 tetrads dissected in the ts942 cross, 19 showed cosegregation of the Ts– and Ura– phenotypes, indicating linkage of the defect with GRC5; and (iv) 12 of 13 dissected tetrads from the ts1100 cross exhibited cosegregation of the Ts– and Ura– phenotypes, indicating linkage of the defect with THS1. These results, in combination with the complementation tests described above, strongly suggest that the phenotypes observed in these three Ts– strains are due solely to mutations in the respective genes.
Characterization of the mutated mrt4 allele in ts1189: Since ts1189 was the only strain complemented by a previously uncharacterized gene (MRT4), identified only in the Saccharomyces genome sequencing project, we decided to characterize the mutated allele in this strain in greater detail. We amplified the MRT4 locus from SS330 wild-type cells and from ts1189 cells, cloned the resulting DNA fragments into the pRS316 vector, and transformed both alleles into ts1189 cells. Expression of the MRT4 gene isolated from the wild-type cells (pDZ63) in ts1189 cells enabled their growth at the nonpermissive temperature, while the allele isolated from the ts1189 cells (mrt4-1189; pDZ64) could not complement the growth defect (Figure 4A).
The decay rates of a number of mRNAs were measured in ts1189 cells expressing the wild-type gene or the mutant mrt4-1189 allele to reiterate the link between the growth and mRNA decay defects in these cells. As shown in Figure 4B, the half-life of the URA5 mRNA was 6 min in cells expressing the wild-type gene (as compared to 4 min in the wild-type cells; see Table 1), while the URA5 mRNA half-life was 20 min in cells expressing the mrt4-1189 allele. Similarly, the STE2 and CYH2 transcripts were destabilized in ts1189 expressing the MRT4 gene but not in cells harboring the mutated allele (data not shown). Thus, the mrt4-1189 allele isolated from ts1189 cells could not complement either the Ts– growth or mRNA decay defects of these cells.
We then sequenced the two PCR products and identified two nucleotide sequence differences between the mrt4-1189 allele and the wild-type gene isolated from the SS330 parental cells. The first is a T to C change at position 552 of the MRT4 ORF, which is in the wobble position and does not change the sequence of the protein. The second mutation is an A to G change at position 647, which results in a lysine to arginine change in Mrt4p (Figure 4C). These results, together with the observation that a cross of ts1189 and wild-type cells segregated 2:2 for Ts+:Ts– growth (data not shown), suggest that the phenotypes exhibited by the ts1189 cells are attributable to the single amino acid substitution in Mrt4p.
The recovery of a Ts– allele of MRT4 suggested that this gene might be essential for viability. To test this possibility, the MRT4 ORF was replaced by the URA3 locus in a wild-type diploid strain and the heterozygous strain was sporulated and dissected. As can be seen in Figure 5, spores deleted for the MRT4 gene (all Ura+) exhibited markedly inhibited growth but were not completely dead. Therefore, it seems that Mrt4p is required for normal cell growth. Similar to the effect seen in the mrt4-1189 Ts– mutant (Table 1), the mrt4Δ strain exhibited a 2.5- to 3-fold stabilization of the URA5 mRNA (data not shown).
Protein synthesis in the mutant strains: GRC5 encodes a ribosomal protein (Dicket al. 1997; Nikaet al. 1997), THS1 encodes a tRNA synthetase (Pape and Tzagoloff 1985), and MRT4 encodes a protein that is similar to the P0 family of 60S ribosomal proteins (Figure 4C). The mRNA decay phenotypes of strains harboring mutations in these genes may thus reflect the indirect effects of translational impairment, similar to that observed when cells are treated with translation inhibitors such as cycloheximide (Herricket al. 1990; Peltzet al. 1992; Beelman and Parker 1994; Zuk and Jacobson 1998). Therefore, it was important to investigate whether the Ts– strains were defective in protein synthesis and, if so, to what extent. This was done by measuring the incorporation of 35S-labeled amino acids (methionine and cysteine) into the cells after a shift to the nonpermissive temperature. Duplicate aliquots of cells were taken at various times after the shift, incubated with the labeled amino acids for 4 min, and then precipitated and washed with trichloroacetic acid. The resulting precipitates were counted and compared to the amount of label incorporated at the time of the shift to 37° (t = 0; Figure 6).
We have shown previously (Zuk and Jacobson 1998) that protein synthesis in SS330 wild-type cells continues to increase throughout the duration of such experiments (after a slight drop, probably due to the heat shock of the shift to 37°). ts817 (sla2), ts942 (grc5), and ts1189 (mrt4) displayed profiles similar to that of wild-type cells (Figure 6), i.e., there was a slight drop in amino acid incorporation immediately after the shift to 37° and then a recovery by 60 min, with the increase in ts817 slightly less than in the other two strains. A substantial reduction in protein synthesis was observed only in the ts1100 (ths1) cells, with incorporation dropping to ∼40% of initial levels by 60 min after the temperature shift.
Since the ts1100 strain exhibited a translation defect in the 35S-amino acid incorporation experiments, we extended the evaluation of protein synthesis in these cells by fractionating cytoplasmic extracts on sucrose gradients and monitoring the distribution of ribosomes in these cells relative to that observed in the wild-type parental strain SS330. As can be seen in Figure 7, the wild-type cells displayed a polysome peak of eight to nine ribosomes at both temperatures (Figure 7, A and B) with a slight increase in the amount of 80S monosomes after the shift to the higher temperature. In ts1100 cells grown at 24° (Figure 7C), the general shape of the profile was relatively normal, with about five ribosomes per mRNA. After the incubation at the nonpermissive temperature (Figure 7D), however, there was a significant shift of ribosomes into the 80S fraction and a redistribution of polyribosomes to three to four ribosomes per mRNA. This is indicative of an overall decrease in the rate of translation initiation, which is consistent with the results of the 35S incorporation experiments described above.
These results suggest that the mRNA decay defects seen in ts817, ts942, and ts1189 are not the consequence of a decrease in general protein synthesis and the genes mutated in these strains may be involved in mRNA turnover directly. It is possible, however, that translation of a small subset of mRNAs could be inhibited in these cells. In contrast, the decay defect seen in the ts1100 cells may very well be due to an indirect effect of impaired translation.
The four Ts– mutant strains accumulate a full-length transcript at the nonpermissive temperature: Proteins involved in mRNA decay could function at any one of the previously identified steps of the principal pathway, i.e., deadenylation, decapping, or exonucleolytic decay (Caponigro and Parker 1996; Jacobson and Peltz 1996). Previous work has shown that cells deficient in the Xrn1p exonuclease accumulate a PGK1 transcript shortened by two nucleotides at its 5′ end, while cells deficient in earlier steps of the pathway (deadenylation or decapping) accumulate full-length PGK1 transcripts (Muhlradet al. 1995). To test whether the four Ts– mutants are deficient in an early or late step of the pathway, RNA was prepared from cells grown at 24°, or after a shift to 37°, and the 5′ end of the PGK1 transcript was measured by primer extension. RNA isolated from xrn1Δ cells was used to identify the full-length (ending in G at position –43 with the A of the translation start designated as 1) and shortened (ending in T at position –41) PGK1 transcripts (Figure 8). The results show that all four Ts– mutants accumulated only the full-length transcript at both temperatures (Figure 8). This indicates that these mutants are defective at an early step of mRNA decay.
Overexpression of NMD3 suppresses the Ts– growth and mRNA decay defects of ts942: ts942 cells did not exhibit a drop in protein synthesis, but had only a slight mRNA decay defect. Therefore, it was unclear whether this mutant would yield substantial insights into mRNA decay mechanisms. However, previous work had identified an nmd3 mutation that suppressed the Ts– growth phenotype of another grc5 mutant allele (Karlet al. 1999). NMD3 is an essential gene that was originally isolated in a two-hybrid screen with UPF1 (a factor involved in nonsense-mediated mRNA decay; He and Jacobson 1995) and that has recently been shown to encode a protein that co-localizes with the 60S ribosomal subunit (Belket al. 1999; Ho and Johnson 1999). To further examine the relationship between NMD3 and GRC5 we tested whether overexpressing NMD3 would suppress the growth and mRNA decay phenotypes of the ts942 mutant.
The NMD3 gene tagged with the triple-HA epitope (Kolodziej and Young 1991) was cloned into the YEp24 episomal vector (Botsteinet al. 1979) and transformed into ts942 cells. This epitope-tagged allele could complement an NMD3 null strain (data not shown). ts942 cells harboring the high-copy NMD3 plasmid were able to grow at 37° (Figure 9A; compare to the ts942 cells transformed with the empty vector in Figure 2), indicating that the Ts– growth phenotype of these cells had been suppressed.
To determine whether overexpression of NMD3 could also suppress the mRNA decay defect, we measured decay rates in the transformed ts942 cells and compared them to those in the cells containing an empty vector (Figure 3D). As can be seen in Figure 9B, the cells transformed with the NMD3 construct had wild-type half-lives for the CYH2 pre-mRNA (4 min) and for the CYH2 (10 min) and STE2 mRNAs (4 min), comparable to the values measured in ts942 cells complemented by the centromeric GRC5 construct and decreased from the values in the originial ts942 cells (see Tables 1 and 2).
These results show that overexpressing Nmd3P results in the same phenotype obtained by complementation with wild-type Grc5p. Since GRC5 encodes a 60S ribosomal protein (Dicket al. 1997; Nikaet al. 1997), further characterization of the genetic relationship between these two proteins may provide greater insight into the link between mRNA decay and protein synthesis.
To identify factors involved in mRNA turnover that are also essential for cell growth, we screened a panel of Northern blots derived from ∼4000 temperature-sensitive yeast strains for elevated levels of the CYH2 pre-mRNA. Five mutant strains (ts817, ts942, ts1100, ts1159, and ts1189) exhibited defects in the decay of several mRNAs. One of these, ts1159, harboring a mutation in TIF51A, the gene coding for elF-5A, has been described in detail elsewhere (Zuk and Jacobson 1998).
The initial characterization of the decay kinetics of wild-type mRNAs in the four mutant strains showed that ts817 and ts1189 had the most extensive decay defects, with up to sevenfold increases in the half-lives of the mRNAs tested (Table 1, Figure 1). ts942 and ts1100 exhibited two- to fourfold increases in the half-lives of the STE2 and URA5 mRNAs, with ts1100 also stabilizing the CYH2 pre-mRNA. The discrepancies among the four strains may be due to the fact that relatively small differences are easier to observe in mRNAs with short half-lives, or may indicate a more fundamental difference in the decay mechanisms of the different transcripts.
Although all the strains were initially chosen on the basis of an increase in the amount of the CYH2 pre-mRNA at the nonpermissive temperature, actual half-life measurements revealed little if any defect in the decay of this transcript in ts942 and ts1189 (Table 1). This seemed odd but additional experiments demonstrated that the increase in CYH2 levels could be attributed to the growth stage at which the cells used for the original blots were harvested (increased levels at low cell densities, but not at the midlog phase at which the half-life measurements were performed). These results suggested that these two mutant strains did not have a defect in the nonsense-mediated decay pathway (of which the CYH2 pre-mRNA is an endogenous substrate; Heet al. 1993). However, evaluation of the decay rates of additional mRNAs showed that these strains did have defects in the decay of inherently unstable transcripts.
To identify the genes mutated in the four strains, we complemented the temperature-sensitive growth phenotype of these strains with a yeast genomic DNA library. This strategy yielded unique DNA fragments (∼10 kb) that complemented each strain. Further dissection of these sequences resulted in the identification of single genes that were able to complement the growth defect of each strain at the nonpermissive temperature (Figure 2), as well as to alleviate the respective mRNA decay defects (Figure 3).
ts1189 was complemented by expression of a previously uncharacterized gene that we termed MRT4. Interestingly, comparison of the Mrt4p sequence to the available databases (Altschulet al. 1990) revealed that this protein has 44% similarity to Rpla0p, the S. cerevisiae member of the P0 family of 60S ribosomal proteins (Figure 4C), one of the components (together with P1 and P2) of the eukaryotic ribosome stalk (Remachaet al. 1995). This may indicate an additional link between the ribosome and components of the mRNA decay pathways. Additionally, Mrt4p is even more homologous to the product of an uncharacterized DNA (F10E7.5) sequenced in the C. elegans genome project (Wilsonet al. 1994).
To further characterize the mrt4-1189 allele, we amplified this locus from ts1189 cells and showed that the mutant allele could not complement the growth and mRNA decay defects of the original cells (Figure 4). Comparison of the mrt4 sequence isolated from the ts1189 cells to that found in the SS330 parental cells revealed two mutations, one of which is silent, and another that results in a lysine to arginine switch at position 216 near the C-terminal end of Mrt4p. Although these two amino acids are chemically very similar, they are subject to substrate-specific methylation. A number of studies have shown that simply substituting an arginine residue for a methylated lysine results in a defective protein, including impaired development of transgenic tobacco plants expressing a mutated calmodulin gene (Robertset al. 1992), and a decrease of >-97% in the ribonucleolytic activity of mutated human afngiogenin (Shapiroet al. 1989).
ts942 was complemented by GRC5 (also called QSR1 and QM1), which encodes the 60S ribosomal subunit protein L9 (Dicket al. 1997; Nikaet al 1997). Grc5p is required for the joining of the 40S and 60S subunits into the 80S monosome (Eisingeret al. 1997) and for the stability of the 60S ribosomal subunit (Dick and Trumpower 1998). In light of these results, it was surprising that the ts942 mutant showed no defect in translation in 35S-amino acid incorporation experiments (Figure 6). However, similar incorporation experiments done with another grc5 mutant (Eisingeret al. 1997) showed no significant effect of the nonpermissive temperature until about 12 hr after a temperature shift. Amino acid incorporation was measured in the ts942 cells after 1 hr at 37°, because the RNA decay defect was already apparent at that time. This raises the possibility that the mutant Grc5p has an effect on mRNA decay before it becomes deficient in translation. Additionally, it is interesting that the growth and mRNA decay defects in ts942 could be suppressed by overexpression of NMD3 (Figure 9), which encodes another 60S-associated protein, originally isolated in a two-hybrid interaction with Upf1p (He and Jacobson 1995). The latter is a protein required for nonsense-mediated mRNA decay (Leeds et al. 1991, 1992; Peltzet al. 1994). Therefore, the functional relationship between Nmd3p and Grc5p may reflect a specific link between the mechanisms of mRNA decay and translation.
ts817 was complemented by expression of SLA2 (also known as END4 and MOP2), which encodes a protein containing an actin-binding talin-like domain, previously identified as being required for accumulation of the PMA1 ATPase, for endocytosis of α-factor and its receptor, and for organization of the actin cytoskeleton (Holtzmanet al. 1993; Rathset al. 1993; Naet al. 1995). The connection between these processes and mRNA turnover is not self-evident. However, it is interesting to note that there is evidence for the influence of the cytoskeleton on the stability of specific mRNAs. It has been demonstrated (Henicset al. 1997) that disrupting the cytoskeleton with cytochalasin leads to the stabilization of lymphokine mRNAs in human peripheral blood lymphocytes, perhaps through the action of AU-rich sequence binding proteins. Similar drug treatments in rat alveolar cells affected the half-lives of surfactant protein mRNAs (albeit these were destabilized in the presence of cytochalasin; Shannonet al. 1998). Additionally, experiments done with a yeast strain harboring a Ts– mutation in ACT1 showed increased stability of the PGK1 and ACT1 mRNAs at the nonpermissive temperature (D. Herrick and A. Jacobson, unpublished results). Further investigation of Sla2p function in ts817 cells may identify a link between the cytoskeleton and mRNA stability.
ts1100 was complemented by expression of THS1, which encodes the yeast cytoplasmic threonyl-tRNA synthetase (Pape and Tzagoloff 1985). As expected from cells with a mutation in a factor required for protein synthesis, ts1100 exhibited a 60% drop in 35S-amino acid incorporation at 37° (Figure 6). This observation was confirmed by a comparison of the polysome profiles of ts1100 cells at the permissive and nonpermissive temperatures (Figure 7). These profiles showed a shift of material into the 80S peak and a smaller number of ribosomes per mRNA after 1 hr at 37° when compared to profiles obtained from wild-type cells. In addition, mapping the 5′ end of the PGK1 mRNA in this mutant (Figure 8) demonstrated that these cells accumulated full-length transcripts, consistent with an early block in mRNA decay. Previous work (Beelman and Parker 1994) has shown that the mRNA accumulated in cells treated with cycloheximide (as a general inhibitor of translation) was predominantly capped, again indicating a block in an early step of the pathway. It is possible, therefore, that the mutation in the THS1 gene leads to an overall increase in mRNA stability solely by virtue of the general drop in translation. However, this was the only tRNA synthetase to be identified in this screen, so its effect on the mRNA decay process may be more specific. Moreover, the effect of the temperature shift on the polysome profile of ts1100 cells was not consistent with a generalized reduction in translation elongation rates (Peltzet al. 1992).
Like the ths1-1100 mutant, strains with the mrt4, grc5, and sla2 Ts– mutations also accumulated full-length PGK1 mRNA at the nonpermissive temperature (Figure 8). In the deadenylation-dependent mRNA decay pathway, this effect is characteristic of the inhibition of mRNA decay before decapping (Muhlradet al. 1995) and indicates that all of the respective gene products are involved in the regulation of either decapping of poly(A) shortening. In light of the homologies and/or previously described roles of these genes it appears that, in addition to mRNA translation (Jacobson and Peltz 1996), other cellular events (such as the state of the cytoskeleton) may regulate mRNA decay rates. Further investigation of the roles played by the mutants identified in this study may shed some light on these regulatory connections.
We are grateful to John Abelson for providing blots, strains, information, and hospitality essential for the success of this project. We thank the members of the Jacobson lab for their comments on the manuscript. This work was supported by National Institutes of Health grant GM-27757 and American Cancer Society Research Project grant 93-026-05-GMC to A.J. and by a postdoctoral fellowship from the Human Frontiers Science Program Organization (LT-78/94) and an Institutional Postdoctoral Training Grant (T32-HD07312-14) to D.Z.
Communicating editor: A. G. Hinnebusch
- Received January 6, 1999.
- Accepted May 17, 1999.
- Copyright © 1999 by the Genetics Society of America