Proc. Natl. Acad. Sci. USA Vol. 87, pp. 2780-2784, April 1990 Biochemistry and a 42- segment within the coding region of the mRNA encoded by the MATal are involved in promoting rapid mRNA decay in yeast (cis-acting instability element/rare codons/) RoY PARKERt AND ALLAN JACOBSON Department of Molecular and Microbiology, University of Massachusetts Medical School, Worcester, MA 01655 Communicated by Gerald R. Fink, January 16, 1990 (received for review November 20, 1989)

ABSTRACT In yeast, the mRNA encoded by the MATal vation consistent with the rapid to a new set of gene is unstable (tl2 = 5 min) and the mRNAs encoded by the specific functions that occurs during mating-type switching ACTI gene (til2 = 30 min) and the PGKI gene (t.l2 = 45 min) are (18). To understand the structural basis for the instability of stable. To understand the RNA structural features that dictate the MATal mRNA, we have constructed several chimeras mRNA decay rates in yeast, we have constructed PGKI/ between the MATal gene and the PGKI and ACT] , MATal and ACTJ/MATal gene fusions and analyzed the both of which stable mRNAs. The decay rates of the decay rates of the resultant chimeric transcripts. Fusion of a hybrid mRNAs identify a 42-nucleotide (nt) coding-region MATal segment containing 73% of the coding region and the segment that is required for rapid decay. The of a 3' to either of the stable genes is sufficient nonsense codon immediately 5' to this 42-nt segment pre- to cause rapid decay ofthe chimeric mRNAs (t.12 = 6-7.5 min). vents rapid mRNA turnover, suggesting that the decay ofthe Sequences required for this rapid decay are not found in the MATal mRNA is promoted by the translocation of ribo- MATal 3' untranslated region but are located within a 42- somes through a specific region of the coding sequence. nucleotide segment of the coding region that has a high content (8 out of 14) ofrare codons. Introduction ofa translational upstream of this region stabilizes the hybrid mRNAs, MATERIALS AND METHODS indicating that the rapid decay promoted by these sequences is Strain and Media. The yeast strain was Y262 (MATa, his4- dependent on ribosomal translocation. 519, ura3-52, rpbl-l; provided by M. Nonet and R. Young; ref. 19). Synthetic media lacking (20) were used to select for Individual mRNAs are degraded at rates that vary by more and maintain plasmids introduced by transformation (21). than an order of magnitude. Although such differences can Plasmid Constructions. The DNA sequences of all junc- affect the intrinsic level ofgene expression and serve as a site tions within the translated regions of gene fusions were of posttranscriptional regulation, little is known about the determined by the dideoxy method (22), using oligonucleo- features of mRNAs that dictate their respective decay rates. tides complementary to flanking sequences as primers. Plas- Initially, it was suggested that the stability of mRNAs is related to basic mRNA properties such as size (1, 2), poly(A) mid constructions (23) were as follows. tail length (3), and translational efficiency (4, 5). However, Pal. Plasmid pRIP4 (a pUC19 derivative containing the several comparisons have failed to reveal any correlation PGKI gene) was cut with Bgl II. Overhanging ends were between these variables and the rate of mRNA turnover, filled in with Klenow enzyme and the DNA was recut with suggesting that general mRNA features are not the primary HindIII. The 1.6-kilobase (kb) EcoRV-HindIII fragment determinants of mRNA decay rate (6, 7). An alternative from pl.9, a subclone of MATal in pBR322 (ref. 25; obtained explanation for differences in decay rate is that unstable from K. Tatchell), was inserted into these sites to yield mRNAs contain specific sequences that promote their rec- pUC19.Pal. The resulting hybrid gene was then subcloned ognition by the cellular turnover machinery. This view is into a yeast centromere vector, pRIPlH [a derivative of supported by recent experiments in which specific sequences pUN50 (26) in which the HindIII site flanking the URA3 gene derived from unstable metazoan mRNAs promoted rapid had been filled in] to yield pRIPlH.Pal. decay when transferred to stable mRNAs. To date, these Aal. First, the Cla I-HindIlI fragment of the PGKI gene "instability elements" have been found in the 3' untranslated in pRIP4 was replaced with the Asu II-HindIII fragment of region (UTR) (8-11) and in coding regions (12-15). Although the MATal gene to yield pUC19.Pal*. The BamHI-HindIII the mechanisms by which such elements promote mRNA fragment of this hybrid gene was then subcloned into the turnover are unknown, there appears to be a close link BamHI and HindIII sites of pRIP1H to yield pRIPlH.Pal*. between the translation process and mRNA decay (16, 17). This plasmid was then digested with BamHI and partially In this paper we describe experiments to identify specific digested with EcoRV (to liberate the remaining PGKI frag- sequences that promote mRNA decay in the yeast Saccha- ment and a small segment of the MATal coding region) and romyces cerevisiae. We recently reported the decay rates of the BamHI-HindIII fragment of the actin gene from pYactl a number of yeast mRNAs, including both stable and unsta- (27) was inserted. The resulting plasmid was designated ble species (18). In agreement with previous work, no cor- pRIPlH.Aal. relation was found between mRNA decay rate and several P(ut)al. pRIP4 (see above, Pal) was cut with Cla I, filled general mRNA properties. Among the unstable mRNAs in with Klenow enzyme, digested with HindIII, and ligated to identified were those encoded by genes involved in mating the 1.1-kb Sca I-HindIII fragment of MATal isolated from (e.g., MATa), STE2, STE3, MFal, and MFA2), an obser- pl.9. This plasmid was designated pRIPl.P(ut)al.

The publication costs of this article were defrayed in part by page charge Abbreviations: nt, nucleotide(s); UTR, untranslated region. payment. This article must therefore be hereby marked "'advertisement" tCurrent address: University Department of Molecular and Cellular in accordance with 18 U.S.C. §1734 solely to indicate this fact. Biology, University of Arizona, Tuscon, AZ 85721. 2780 Downloaded by guest on September 27, 2021 Biochemistry: Parker and Jacobson Proc. NatL. Acad. Sci. USA 87 (1990) 2781 PalP. The 0.26-kb Cla I (filled-in)-HindIII fragment of Minutes at 36-C PGKI that contains the 3' UTR was inserted into pRIPlH.- Pal that had been digested partially with Sca I (site just 0 5 10 20 30 40 50 60 downstream ofthe MATal translational stop) and completely with HindIII, yielding pRIPlH.PalP. ;ii.- so Deletions of Pal. pUC19.Pal.UAG (see below) was cut PGK1 --..w 45 min with Bgl II and then treated with BAL-31 nuclease for various MA&-.-a 5 min times. Single-stranded overhangs were filled in with Klenow MATa 1 _ enzyme and then ligated to Bgl II linkers (no. 1001; New England Biolabs). The DNA was then cut with HindIII and the fragments containing the MATal region were purified and 100 0 inserted into Bgl II/HindIII-cut pRIPlH.Pal.UAG. Plas- mids these deletions are as U O containing designated pRIPlH.- CZ PalAx, where x is the number of deleted in the 3' E direction from the EcoRV site of the MATal gene. For a) the frame was 10 pRIPlH.PalA52, reading restored, following z sequencing, by filling in the Bgl II site at the new junction CC between the PGKI and MATal sequences. oPso Pal. UAG. pUC19.Pal* (see above) was cut with EcoRV and ligated to the linker 5'-CTAGATCTAG-3'. The DNA was 1 then cut with Bgl II (cutting the linker and the Bgl II site in 20 40 60 the PGKI sequences) and the plasmid was ligated to itself. Minutes at 36°C This plasmid, pUC19.Pal.UAG, was cut with BamHI and HindIII, and the fragment containing the hybrid gene was FIG. 1. Decay of stable and unstable mRNAs in yeast. Relative inserted into the BamHI and HindIII sites of levels of the PGKJ and MATal mRNAs were measured at different pRIPlHB (a times after a shift to 36°( in a temperature-sensitive RNA polymer- derivative ofpRIPlH in which the Bgl II site in the backbone ase II mutant. (Upper) Northern analysis ofmRNA levels at different had been filled in with Klenow enzyme), yielding the plasmid times after the temperature shift. The probe for the PGKJ mRNA pRIPlHB6.Pal.UAG. was a random-primed Dra 1-HindlIl restriction fragment containing Aal. UAG. The 1.4-kb BamHI-HindIII (filled-in) fragment the entire PGKI coding sequence. The probe for the MATal mRNA containing the 5' portion of the ACT) gene was inserted into was the same as for Fig. 2. (Lower) Quantitation ofthe Northern blot. the BamHI and Bgl II (overhang removed with mung bean mRNA levels are normalized to the level at time zero. *, MATal; o, nuclease) sites of pRIPlHB.Pal.UAG to yield pRIPlHB6.- PGKI. Half- (45 min and 5 min) calculated from the slopes are Aal.UAG. shown to the right of the autoradiogram. mRNA Decay Measurements, RNA Preparation, and RNA Analysis. mRNA decay rates were measured as described (18). mRNA (see schematic, Fig. 2). The only MATal sequences In brief, was inhibited by thermal inactivation of absent from this hybrid are the 21-nt 5' UTR and the initial RNA polymerase II and mRNA levels were quantitated by 140 nt of coding sequence. Northern blotting (28) with DNA probes labeled to high Decay of Pal mRNA was assayed by transferring the specific activity (29). Northern blots were quantitated by hybrid gene to a yeast centromere plasmid and introducing densitometry or by directly counting the ( decays occurring in the resulting plasmid into a strain (RY262) containing the each band with a Betascope (Betagen, Waltham, MA; ref. 18). rpbl-) temperature-sensitive RNA polymerase II . Data are expressed as the logl0 ofthe percentage ofeach RNA mRNA synthesis was inhibited by shifting these cells to the remaining vs. time at 360C. In multiple determinations half- restrictive temperature (360() and mRNA decay rates were (ti12) values varied by a maximum of ±23%. measured by Northern blot analysis of equal amounts of RNA isolated at various times after the temperature shift (Fig. 2A). Quantitation of this blot indicates that the decay RESULTS rate ofthe hybrid Pal mRNA is quite rapid (til2 = 6 ± 0.9 min) Chimeric Transcripts Containing MATal Sequences Decay and is comparable to the decay rate ofthe MATal mRNA (ti1/2 Rapidly. Recently, we measured the decay rates of 20 yeast = 5 min). This rapid decay differs significantly from the mRNAs by quantitating the levels ofindividual remain- turnover kinetics of both the endogenous PGKI mRNA (ti/2 ing after thermal inactivation of RNA polymerase II in a = 45 min) and a chimeric mRNA containing the same temperature-sensitive rpbl-l mutant (18). One of the least fragment ofPGK) mRNA fused to the 3' portion ofthe stable stable mRNAs examined in that study was the mRNA encoded ACT) mRNA (t1/2 = 45 min; unpublished work). by the MATal gene. Its rate of decay (tl2 = 5.0 min) is 6- to To rule out the possibility that the rapid decay of the Pal 9-fold faster than that ofstable mRNAs, such as those encoded mRNA was a fortuitous consequence of the specific hybrid by the ACT) (actin, til2 = 30 min) or PGKI (phosphoglycerate that was formed (e.g., that the junction sequence caused kinase, til = 45 min) genes. For comparison, Northern blots rapid mRNA decay), we created a second gene fusion, Aal. showing the decay ofthe PGKI and MATal mRNAs following In the transcripts derived from this construct, the same the inhibition of transcription in a temperature-sensitive RNA sequences of MATal that were used in the Pal hybrid are polymerase II mutant are shown in Fig. 1. fused downstream of a fragment of the stable ACT) mRNA. One explanation for the instability ofthe MATal mRNA is A Northern blot illustrating the decay of this mRNA in the presence within this mRNA of a specific sequence temperature-shifted Y262 cells is shown in Fig. 2B. This blot element that promotes rapid degradation. Our approach to shows that fusion of MATal sequences to a second stable delineating such sequences is to identify segments ofunstable mRNA also promotes rapid decay (til2 = 7.5 min), although mRNAs that will promote rapid mRNA decay when trans- a small fraction (10%6) of this mRNA is resistant to decay. ferred to mRNAs that are normally stable (24). As a first step These results indicate that sequences sufficient to promote in using this approach for the analysis ofthe MATal mRNA, rapid mRNA degradation are contained within the 580 nt of we created a gene fusion, designated Pal, that produces a the MATal mRNA present in the Pal and Aal hybrids. transcript containing 1205 (of 1380) nt from the stable PGKI The 3' UTR of MATal Is Not Sufficient or Required for mRNA fused, in frame, to 580 (of 736) nt from the MATal Rapid mRNA Decay. Studies with several mammalian Downloaded by guest on September 27, 2021 2782 Biochemistry: Parker and Jacobson Proc. Natl. Acad Sci. USA 87 (1990)

MlinuLtes a' 36 C below). This lag does not correlate with any particular stage of t A 0 5 10 2030 40 50 6E cellular growth, the degree of temperature shift, or the par- ticular construct, but it is suggestive of a possible aging PRl mRNA component in the decay of these mRNAs.] UAAN _ mla_ 6 . i, Sequences Within the MATal Coding Region Are Required for Rapid mRNA Decay. To test whether specific sequences from the MATal coding region would promote rapid decay of B the chimeric mRNAs, we constructed a series of AAlrmRNA constructs in which coding sequences were iAU ; .iss _ MATal progres- -; UAA __. -*Ate>,? sively removed from the Pal fusion gene. These deletions, shown schematically in Fig. 4, commence at the PGKI/ 250 nt MATal junction and remove 52-247 nt of MATal sequence. All fusions are in the proper reading frame to translate the FIG. 2. Northern blot analysis of the decay of the Pal (A) and remaining MATal coding sequence (confirmed by sequencing Aal (B) mRNAs. In the diagrams to the left (and in all other figures), the hybrid genes). Each deletion gene was introduced into PGKJ sequences are shown as black boxes, MATal sequences as RY262 cells and the consequences of these deletions on the open boxes, and ACT) sequences as gray boxes. The probe for both decay of the respective chimeric transcripts are shown in Fig. blots was an EcoRV-Sca I restriction fragment containing 400 nt of 4. Compared to the Pal mRNA (tl, = 6.0 min), deletion of 52 MATal coding sequence. nt has little or no effect on the miRIA decay rate (Pal.A52; tl,2 = 7.5 min, Fig. 4A). In contrast, deletion of 94 nt (Fig. 4B) or mRNAs have shown that structural determinants that con- 140 nt (Fig. 4C) results in a 4- to 5-fold increase in the half-life tribute to the rapid decay of those mRNAs can be localized of the corresponding mRNA (e.g., Pal.A94, ti,2 = 35 min). to 3' UTRs (8-11). In this light we sought to determine the mRNAs with larger deletions, of 188 or 247 nt (Fig. 4D and E), role of the 3' UTR in the rapid decay of the Pal and Aal decay with kinetics similar to that of the PGK) mRNA (tl, - mRNAs. To test whether the 3' UTR of MATal mRNA was >45 min). These results indicate that specific sequences within sufficient to promote rapid decay, we constructed a hybrid the 42 nt between the endpoints of A52 and A94 are required gene, P(ut)al, in which the majority of the MATal 3' UTR for the rapid decay of the hybrid mRNAs. The possibility that (170 of 200 nt) was fused 18 nt downstream of the PGKI other specific sequences also contribute to the small difference translational termination codon. The chimeric mRNA tran- (1.5-fold) in mRNA turnover rates observed upon deletion of scribed from this hybrid gene contains 85 nt more of the >140 nt cannot be ruled out at present. PGKI mRNA, including the termination codon, than the Pal We have compared the 42-nt region between the endpoints nmRNA and has a half-life of 45 min (Fig. 3A). Since this decay of A52 and A94 to other unstable yeast mRNAs and have yet rate is the same as that of the PGKI mRNA, we conclude that to find any striking homologies in either the RNA or the the majority of the 3' UTR of the MATal mRNA is not sequence. This may reflect the possibility that the sufficient to promote rapid mRNA decay. 42-nt region is part of a larger element whose 3' boundary has The failure of the MATal 3' UTR to destabilize the PGK) yet to be defined. However, we do note that the 42-nt region mRNA does not preclude the possibility that an "instability is rich in rare codons (8 out of 14), where a rare codon is element" of the MATal mRNA includes the 3' UTR. To test defined by its occurrence fewer than 13 times per 1000 yeast whether the MATal 3' UTR was required for rapid mRNA codons (31). Including one codon at the deletion boundary, decay, we created a hybrid gene in which we replaced the this region also contains a stretch of6 out of7 contiguous rare majority of the 3' UTR of the Pal hybrid (170 of 200 nt) with codons (Fig. 5). This concentration of rare codons suggests the 3' UTR from the PGKI gene (30). This hybrid, PalP, either that a occurs within this region or produces an mRNA that contains 1200 nt of the PGK) mRNA that there may be a specific requirement for a translational fused to 410 nt of the MATal mRNA (including the translation pause atthis site (see Discussion). stop codon and 30 nt of the 3' UTR) fused to the 75-nt 3' UTR Upstream Nonsense Codons Prevent Rapid Turnover Pro- of the PGKI mRNA. Following a brief lag, the PalP mRNA moted by MAT41 Coding Sequences. The results described decays with a tx, of 5.0 mini (Fig. 3B); i.e., the PalP mRNA above demonstrate that specific sequences within the MATal decays at essentially the same rate as the Pal mRNA. We coding region are required for rapid mRNA decay. The conclude that the majority of the MATal 3' UTR is not involvement of the coding region suggests that, as observed sufficient or required for rapid decay. [The brief lag seen in the for several other mRNAs (32), there may be a requirement for decay of the PalP mRNA is seen occasionally with other translation in the degradation of the MATal mRNA. To test rapidly decaying mRNAs (e.g., Pal, Aal, and Pal.A52; see whether translation is required for mRNA turnover, we constructed two hybrid genes in which ribosome transloca- Minutes at 36 C tion through the MATal sequences was prevented by up- stream stop codons. The first, Pal.UAG, is A 5 10 20 30 405060 identical to Pal except that an amber stop codon has been introduced at the Ptut)(0 mRNA junction between the PGK) and MATal sequences. Simi- _'I- - 45 min larly, Aal.UAG is identical to Aal except for a stop codon at the junction of the ACT) and MATal sequences. Fig. 6 shows that the decay of the mRNAs containing the stop B codons is significantly slower than the decay of the corre- PRi P ITiRNA sponding mRNAs in which translation continues through the _ _ - -s _5_5 Mir, MATal sequence. The Pal.UAG mRNA(ti/2 = 37 min) decays 5- to 6-fold slower than Pal mRNA (tl,2 = 6 min), and 250 nt the Aal.UAG mRNA(ti/2 = 17 min) decays 2- to3-fold slower than Aal mRNA (ty,2 = 7.5 min). Since an out- FIG. 3. Decay of P(ut)al (A) and PalP (B) mRNAs analyzed by of-frame UAG normally exists within the 42-nt MATal Northern blotting. The probe for the blot in A was a 1.1-kb Sca instability element, we consider it unlikely that the inserted I-HindI11 restriction fragment from the 3' end of the MATal gene. UAG simply provides a site for direct binding of some cellular The probe for the blot in B was the same as for Fig. 2. factor. Rather, since the reduction in decay rates is attribut- Downloaded by guest on September 27, 2021 Biochemistry: Parker and Jacobson Proc. Natl. Acad. Sci. USA 87 (1990) 2783

Minutes at 360C 0 5 10 20 30 40 50 60 tAl2 A A52 iUAA | -I g-~~~.PM40.w.o d By - 75 min

B AA94

HIIdZ -*O W, a_ - 35min

c A140 =mo * -"'AO* " AM 33 min FIG. 4. Deletion analysis of D MATal sequences contributing to A188 rapid decay. The effect ofdeleting MATal sequences from the Pal AU JA j-. *ga. Upa. 4-- - 48 min mRNA was quantitated by North- ern blotting. In the schematic, de- leted sequences are shown as sin- gle lines with the number of nu- E cleotides deleted shown above A247 each deletion. Half-lives shown at 1 -- * _ W 53 min the are the average of multi- E~Z 4 p4U a. right ple experiments (e.g., for A52, t1/2 = 7.5 ± 0.25 min, n = 3; for A94, 250 nt ti/2 = 35 + 8 min, n = 3).

able to the introduction of stop codons, we conclude that these stable reporter mRNAs results in accelerated mRNA ribosome translocation through MATal sequences is re- degradation. Formally, this accelerated decay could be due to quired for facilitating rapid mRNA decay. (Although we do the loss of specific sequences within the stable mRNAs that not fully understand the faster decay rate of the Aal.UAG protect those mRNAs from decay. However, since either hybrid mRNA, we note that the ACT) mRNA is normally less deletion (Fig. 4) or changes in the translation status (Fig. 6) stable than the PGK1 mRNA.) of specific MATal sequences convert the hybrid mRNA into a stable form, the rapid decay is clearly dependent on specific DISCUSSION MATal sequences. Based on an analysis ofchimeric mRNAs, virtually none of Nucleotides Within the MATa) Coding Region Can Promote the MATal 3' UTR is either sufficient or required for rapid Rapid mRNA Decay. In eukaryotic cells, the relationship mRNA decay. In contrast, a 42-nt segment within the MATal between mRNA structure and stability has yet to be deter- coding sequence is required for rapid decay. Further work mined. Experiments with several mammalian mRNAs have will be necessary to test the sufficiency ofthis "element" and identified specific structural determinants, primarily con- to define its 3' limit. Fig. 3 indicates that sequences required fined to the 3' UTR, that appear to dictate rapid decay (8-11, for the function ofthis element cannot extend beyond the first 33). We have begun an analysis ofmRNA decay in yeast with 30 nt of the 3' UTR. Along with other work (13, 15), these the goal of identifying both cis-acting sequences and trans- results add to the evidence that stability determinants will be acting components ofthe cellular mRNA turnover machinery (18, 24). In this paper we have described the decay rates of Minutes at 36 C

chimeric mRNAs composed of fragments of the unstable A 0 5 10 20 3040 5060 t mRNA encoded by the MATa) gene and the stable mRNAs Pacl.UAG mRNA encoded by the PGKI and ACT) genes. We find that fusion of the majority of the unstable MATal mRNA to either of _ UAG UAA _ 37 min A52 B Aal.UAG mRNA 66 + 74 MAUG UAA _ 6 _ 17 min .... AUU CAC AGG AUA GCG UCU GGA AGU CAA

75 82 250 nt MATa1 AAU ACU CAG UUU CGA CAG UUC AAU AAG .... FIG. 6. Translation is required for rapid mRNA decay. The decay ofhybrid mRNAs that contained an amber stop codon at thejunction A94 of the stable reporter mRNA and the MATal sequences was mea- sured by Northern blotting. The probe for these blots was the same FIG. 5. The segment required for rapid decay is rich in rare as for Fig. 1. (A) Pal.UAG. (B) Aal.UAG. The RNAs shorter than codons. Shown is a portion of the MATal sequence in which the 3' intact hybrid mRNAs are apparent only when nonsense codons are endpoints of A52 and A94 are indicated. The small numbers identify introduced upstream of the PGKI/MATal or ACTJ/MATal junc- codon positions within the MATa) mRNA. Codons with a frequency tions. These RNAs lack sequences complementary to probes specific of occurrence <13/1000 are underlined. for the 3' end of MATal and thus may be decay products. Downloaded by guest on September 27, 2021 2784 Biochemistry: Parker and Jacobson Proc. Natl. Acad. Sci. USA 87 (1990) found not only in the 3' UTR but also in the coding regions pausing (36, 37), suggests a model in which mRNA structural of unstable mRNAs. changes may be affected by the particular positioning of a Ribosome Translocation Stimulates Rapid Decay Promoted paused ribosome. However, since the stable PGKI mRNA by the MAT61 mRNA. An emerging theme in the study of can be altered to include up to 40% rare codons with, at most, mRNA turnover is a strong link between the translation and a 3-fold effect on steady-state mRNA level (38), and this degradation of mRNA. Perturbations that alter the transla- difference may actually be due to a change in transcription tion status of an mRNA often have dramatic effects on its rates (39), it seems unlikely that ribosome pausing per se is decay rate. For example, treatment ofcells with inhibitors of sufficient to promote rapid mRNA decay. Thus, a ribosome protein synthesis can drastically reduce the decay rate of paused at a specific site may expose downstream nuclease- many mRNAs (11, 17, 18). Similarly, premature nonsense recognition sites that can then be cleaved by either a soluble codons can lead to an increase in the rate (4, 5). One of the or a ribosome-bound nuclease. clearest examples of a relationship between translation and mRNA decay comes from experiments with mammalian We thank Dorothy Haffey for her role in the conception of this tubulin mRNAs, which show that the signal recognized by project and Stuart Peltz, Laura Steel, David Munroe, David Herrick, the cellular turnover machinery is found in the nascent and Janet Donahue for helpful discussions, miscellaneous reagents, polypeptide (17, 32, 34). To determine whether the function and short-term, low-interest loans. This work was supported by a ofthe MATal instability element is dependent on translation, grant to A.J. (GM27757) and a postdoctoral fellowship to R.P. we inserted stop codons upstream of these sequences in (GM11479) from the National Institutes of Health. hybrids formed with both PGKI and ACT). Ribosome trans- 1. Singer, R. H. & Penman, S. (1973) J. Mol. Biol. 78, 321-334. location up to or through the 42-nt MATal element defined 2. Meyuhas, 0. & Perry, R. P. (1979) Cell 16, 139-148. by deletion must be involved in the stimulation of mRNA 3. Nudel, U., Soreq, H. & Littauer, U. Z. (1976) Eur. J. Biochem. 64, decay since insertion of these stop codons increases the 115-121. half-lives of the Pal and Aal mRNAs from 6 min to 37 min 4. Losson, R. & Lacroute, F. (1979) Proc. Natl. Acad. Sci. USA 76, 5134-5137. and from 7.5 min to 17 min, respectively. Consistent with this 5. Maquat, L. E., Kinniburgh, A. J., Rachmilewitz, E. A. & Ross, J. (1981) conclusion, treatment of cells with cycloheximide stabilizes Cell 27, 543-553. the MATa) mRNA (18). The exact role of translation in the 6. Santiago, T. C., Bettany, A. J. E., Purvis, I. J. & Brown, A. J. P. (1987) turnover process is unknown. Translation may stimulate the Nucleic Acids Res. 15, 2417-2429. 7. Shapiro, R. A., Herrick, D., Manrow, R. E., Blinder, D. & Jacobson, A. activity ofan element that normally functions inefficiently, or (1988) Mol. Cell. Biol. 8, 1957-1969. translation may be absolutely required for the MATal se- 8. Shaw, G. & Kamen, R. (1986) Cell 46, 659-667. quences to function. In the latter case, the decay (albeit slow) 9. Pandey, N. B. & Marzluff, W. F. (1987) Mol. Cell. Biol. 7, 4557-4559. of the Pal.UAG and Aal.UAG hybrids could be due to 10. Casey, J. L., Hentze, M. W., Koeller, D. M., Caughman, S. W., Rouault, T. A., Klausner, R. D. & Harford, J. B. (1988) Science 240, alternative decay elements. 924-928. Decay rates ofthe chimeric mRNAs analyzed in this study 11. Mullner, E. W. & Kuhn, L. C. (1988) Cell 53, 815-825. have been obtained by direct measurement, using a temper- 12. Gay, D. A., Lau, T. J. Y. & Cleveland, D. W. (1987) Cell 50, 671-679. ature-sensitive RNA polymerase II mutant. A second, but 13. Harland, R. & Misher, L. (1988) Development 102, 837-852. 14. Kabnick, K. S. & Housman, D. E. (1988) Mol. Cell. Biol. 8, 3244-3250. indirect, approach to measuring mRNA decay rates com- 15. Shyu, A.-B., Greenberg, M. E. & Belasco, J. G. (1989) Genes Dev. 3, pares the steady-state levels of chimeric mRNAs to those of 60-72. endogenous mRNAs that have a known half-life and utilize 16. Graves, R. A., Pandey, N. B., Chodchoy, N. & Marzluff, W. F. (1987) the same transcriptional (24). Consistent with its Cell 48, 615-626. 17. Gay, D. A., Sisodia, S. S. & Cleveland, D. W. (1989) Proc. Nat!. Acad. rapid decay rate, the steady-state level of Pal mRNA is Sci. USA 86, 5763-5767. reduced relative to PGK) mRNA (data not shown). How- 18. Herrick, D., Parker, R. & Jacobson, A. (1990) Mol. Cell. Biol., in press. ever, the steady-state level ofthe Pal.UAG mRNA is similar 19. Nonet, M., Scafe, C., Sexton, J. & Young, R. (1987) Mol. Cell. Biol. 7, to that of Pal (data not shown). Although the indirect nature 1602-1611. 20. Sherman, F., Fink, G. R. & Hicks, J. B. (1986) Laboratory Course of this assay makes it very difficult to interpret this result, it Manualfor Methods in Yeast Genetics (Cold Spring Harbor Lab., Cold is tempting to speculate that the requirement for translation Spring Harbor, NY). is more stringent under our assay conditions (higher temper- 21. Ito, H., Fukuda, Y., Murata, K. & Kimura, A. (1983) J. Bacteriol. 153, ature; inhibited transcription). This reasoning is most con- 163-168. 22. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl. Acad. Sci. sistent with a model in which the role of translation is to USA 74, 5463-5467. stimulate the activity of an element that can function at a 23. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular Cloning: A reduced rate when not translated. Laboratory Manual(Cold Spring Harbor Lab., Cold Spring Harbor, NY). Possible Mechanisms for the Linkage of mRNA Decay to 24. Parker, R., Herrick, D., Peltz, S. W. & Jacobson, A. (1990) Methods Enzymol., in press. Both Translation and Specific mRNA Sequences. The require- 25. Astell, C. R., Ahlstrom-Jonasson, L., Smith, M., Tatchell, K., Nasmyth, ment of both a specific sequence and translation for mRNA K. A. & Hall, B. D. (1981) Cell 27, 15-23. decay suggests several nonexclusive mechanisms by which 26. Elledge, S. J. & Davis, R. W. (1988) Gene 70, 303-312. turnover could be affected: (i) the actual recognition event 27. Ng, R. & Abelson, J. (1980) Proc. Natl. Acad. Sci. USA 77, 3912-3916. 28. Thomas, P. S. (1980) Proc. Nat!. Acad. Sci. USA 77, 5201-5205. may occur, not at the RNA level, but as with mammalian 29. Feinberg, A. P. & Vogelstein, B. (1984) Anal. Biochem. 137, 266-267. tubulin mRNAs (17, 34), as the nascent polypeptide emerges 30. Hitzeman, R. A., Hagie, F. E., Hayflick, J. S., Chen, C. Y., Seeburg, from the ribosome (since the putative "target" region is 4100 P. H. & Dernck, R. (1982) Nucleic Acids Res. 10, 7791-7808. amino acids from the C terminus of the MATal polypeptide, 31. Aota, S., Gojobori, T., Ishibashi, F., Maruyama, T. & Ikemura, T. (1988) Nucleic Acids Res. 16, r315-r402. it would emerge from the ribosome while the mRNA was still 32. Cleveland, D. W. (1988) Trends Biochem. Sci. 13, 339-343. being translated); (ii) the ribosome itselfmay be critical to the 33. Luscher, B., Stauber, C., Schindler, R. & Schumperli, D. (1985) Proc. degradation process, either by delivering a nuclease (16, 34, Natl. Acad. Sci. USA 82, 4389-4393. 35) or by triggering a catalytic event itself; or (iii) the passage 34. Yen, T. J., Machlin, P. S. & Cleveland, D. W. (1988) Nature (London) 334, 580-585. of a ribosome through this region may alter the secondary 35. Ross, J. & Kobs, G. (1986) J. Mol. Biol. 188, 579-593. structure of the mRNA in such a way as to expose sequences 36. Varenne, S., Buc, J., Lloubes, R. & Lazdunski, C. (1984) J. Mol. Biol. containing nuclease-recognition sites that would normally 180, 549-576. not be available (16). The concentration ofrare codons in the 37. Wolin, S. L. & Walter, P. (1988) EMBO J. 7, 3559-3569. 38. Hoekema, A., Kastelein, R. A., Vasser, M. & deBoer, H. A. (1987) Mol. sequences required for rapid decay (Fig. 6), coupled with the Cell. Biol. 7, 2914-2924. prevalence of rare codons in unstable yeast mRNAs (18) and 39. Mellor, J., Dobson, M. J., Kingsman, A. J. & Kingsman, S. M. (1987) the known ability of rare codons to induce translational Nucleic Acids Res. 15, 6243-6259. Downloaded by guest on September 27, 2021