A turnover pathway for both stable and unstable mRNAs in yeast: evidence for a requirement for deadenylation

Turnover mechanisms of the stable yeast PGK1 mRNA

The first step in the decay of several yeast mRNAs is the shortening of the poly(A) tail, which for the MFA2 transcript triggers decapping and 5'-to-3' degradation. To understand the basis for differences in mRNA decay rates, it is important to determine if deadenylation-dependent decapping is specific to the unstable MFA2 transcript or is a general mechanism of mRNA degradation. To this end, we analyzed the turnover of the stable PGK1 mRNA by monitoring the decay of a pulse of newly synthesized transcripts while using two strategies to trap decay intermediates. First, we used strains deleted for the XRN1 gene, which encodes a major 5'-to-3' exonuclease in Saccharomyces cerevisiae. In xrn1 delta cells, PGK1 transcripts lacking the 5' cap structure and a few nucleotides at the 5' end were detected after deadenylation. Second, we inserted into the PGK1 5' untranslated region strong RNA secondary structures, which can slow exonucleolytic digestion and thereby trap decay intermediates. These secondary structures led to the accumulation of PGK1 mRNA fragments, following deadenylation, trimmed from the 5' end to the site of the secondary structure. The insertion of strong secondary structures into the 5' untranslated region also inhibited translation of the mRNA and greatly stimulated the decay of the PGK1 transcripts, suggesting that translation of the PGK1 mRNA is required for its normally slow rate of decay. These results suggest that one mechanism of degradation of the PGK1 transcript is deadenylation followed by decapping and subsequent 5'-to-3' exonucleolytic degradation. In addition, by blocking the 5'-to-3' degradation process, we observed PGK1 mRNA fragments that are consistent with a 3'-to-5' pathway of mRNA turnover that is slightly slower than the decapping/5'-to-3' decay pathway. These observations indicate that there are multiple mechanisms by which an individual transcript can be degraded following deadenylation.

Polyadenylylation helps regulate mRNA decay in Escherichia coli

As part of our genetic analysis of mRNA decay in Escherichia coli K-12, we examined the effect of the pcnB gene [encoding poly(A) polymerase I] on message stability. Eliminating poly(A) polymerase I (delta pcnB) dramatically stabilized the lpp, ompA, and trxA transcripts. The half-lives of individual mRNAs were increased in both a delta pcnB single mutant and a delta pcnB pnp-7 rnb-500 rne-1 multiple mutant. We also found mRNA decay intermediates in delta pcnB mutants that were not detected in control strains. By end-labeling total E. coli RNA with [32P]pCp and T4 RNA ligase and then digesting the RNA with RNase A and T1, we showed that many RNAs in a wild-type strain contained poly(A) tails ranging from 10 nt to > 50 nt long. When polynucleotide phosphorylase, RNase II, and RNase E were absent, the length (> 100 nt) and number (10- to 20-fold) of the poly(A) tails increased. After transcription initiation was stopped with rifampicin, polyadenylylation apparently continued. Deleting the structural gene for poly(A) polymerase I (pcnB) reduced the amount of 3'-terminal poly(A) sequences by > 90%. We propose a model for the role of polyadenylylation in mRNA decay.

Poly(A) tail length control is caused by termination of processive synthesis

Poly(A) polymerase synthesizes poly(A) tails rapidly and processively only when the substrate RNA is bound simultaneously by two stimulatory proteins, the cleavage and polyadenylation specificity factor (CPSF) and poly(A)-binding protein II (PAB II). A burst of synthesis terminates after the addition of about 250 nucleotides, a length corresponding to that of newly synthesized poly(A) tails in vivo. Further elongation is slow. Length control can be reproduced with premade poly(A) tails of different lengths and is insensitive to large changes in the elongation rate. Thus, the control mechanism truly measures the length of the poly(A) tail. The stimulatory action of PAB II is similar on long and short tails. Coating of poly(A) with one PAB II molecule for approximately 30 nucleotides is required, such that the number of PAB II molecules in the polyadenylation complex is a direct measure of poly(A) tail length. CPSF also stimulates poly(A) polymerase on long and short tails. Long tails differ from short ones only in that they do not permit the simultaneous stimulation of poly(A) polymerase by CPSF and PAB II. Consequently, elongation of long tails is distributive. Thus, length control is brought about by an interruption of the interactions responsible for rapid and processive elongation of short tails. The 3'-end of the poly(A) tail is not sequestered in the protein-RNA complex when the correct length has been reached. Neither ATP hydrolysis nor turnover of the polymerized AMP is involved in length control.

Characterization of cis-acting sequences and decay intermediates involved in nonsense-mediated mRNA turnover

Several lines of evidence indicate that the processes of mRNA turnover and translation are intimately linked and that understanding this relationship is critical to elucidating the mechanism of mRNA decay. One clear example of this relationship is the observation that nonsense mutations can accelerate the decay of mRNAs in a process that we term nonsense-mediated mRNA decay. The experiments described here demonstrate that in the yeast Saccharomyces cerevisiae premature translational termination within the initial two-thirds of the PGK1 coding region accelerates decay of that transcript regardless of which of the stop codons is used. Nonsense mutations within the last quarter of the coding region have no effect on PGK1 mRNA decay. The sequences required for nonsense-mediated mRNA decay include a termination codon and specific sequences 3' to the nonsense mutation. Translation of two-thirds of the PGK1 coding region inactivates the nonsense-mediated mRNA decay pathway. This observation explains why carboxyl-terminal nonsense mutations are resistant to accelerated decay. Characterization of the decay of nonsense-containing HIS4 transcripts yielded results mirroring those described above, suggesting that the sequence requirements described for the PGK1 transcript are likely to be a general characteristic of this decay pathway. In addition, an analysis of the decay intermediates of nonsense-containing mRNAs indicates that nonsense-mediated mRNA decay flows through a pathway similar to that described for a class of wild-type transcripts. The initial cleavage event occurs near the 5' terminus of the nonsense-containing transcript and is followed by 5'-->3' exonucleolytic digestion. A model for nonsense-mediated mRNA decay based on these results is discussed.

Control of RAS mRNA level by the mevalonate pathway

The ability of Ras proteins to initiate eukaryotic cell proliferation requires the post-translational attachment of a farnesyl group, an isoprenoid lipid moiety derived from mevalonate, to the carboxyl-terminus of the protein. This modification is essential for the subsequent processing and intracellular targeting of the Ras protein. Here we report that mevalonate is also required for the efficient synthesis of Ras proteins in Saccharomyces cerevisiae. Depletion of intracellular mevalonate resulted in decreased steady-state levels of Ras1p and Ras2p, an effect that was mediated at the level of mRNA accumulation. The sequences controlling the response of RAS2 mRNA level to mevalonate availability, mapped to the coding region of the RAS2 gene. Mevalonate starvation also had a significant effect on the expression of some, but not all, genes encoding prenylated proteins. The regulatory effect on RAS2 mRNA did not require a functional farnesyl transferase. These results uncover a novel regulatory role for mevalonate-derived products and expand the potential for inhibitors of mevalonate metabolism as anti-cancer agents.

A guanosine quadruplex and two stable hairpins flank a major cleavage site in insulin-like growth factor II mRNA

Insulin-like growth factor II (IGF-II) mRNAs are cleaved by an endonucleolytic event in a conserved part of their 3' untranslated region that is predicted to exhibit a complex higher-order RNA structure. In the present study, we have examined the putative secondary structures of in vitro transcripts from the conserved part of human and rat mRNAs by enzymatic and chemical probing. The results show that the cleavage site is situated between two highly structured domains. The upstream domain consists of two large hairpins, whereas the downstream domain is guanosine-rich. The guanosine-rich domain adopts a compact unimolecular conformation in Na+ or K+ but not in Li+, and it completely arrests reverse transcription in K+ but only partially in Na+, indicating the presence of an intramolecular guanosine quadruplex. The flanking higher-order structures may ensure that the cleavage site is not sequestered in stable RNA structures, thus allowing interactions with RNA or proteins at posttranscriptional stages of IGF-II expression.

AUUUA is not sufficient to promote poly(A) shortening and degradation of an mRNA: the functional sequence within AU-rich elements may be UUAUUUA(U/A)(U/A)

AU-rich elements (AREs) in the 3' untranslated regions of several cytokine and oncogene mRNAs have been shown to function as signals for rapid mRNA degradation, and it is assumed that the many other cytokine and oncogene mRNAs that contain AU-rich sequences in the 3' untranslated region are similarly targeted for rapid turnover. We have used a chimeric gene composed mostly of growth hormone sequences with expression driven by the c-fos promoter to investigate the minimal sequence required to act as a functional destabilizing element and to monitor the effect of these sequences on early steps in the degradation pathway. We find that neither AUUUA, UAUUUA, nor AUUUAU can function as a destabilizing element. However, the sequence UAUUUAU, when present in three copies, is sufficient to destabilize a chimeric mRNA. We propose that this sequence functions by virtue of being a sufficient portion of the larger sequence, UUAUUUA(U/A)(U/A), that we propose forms the optimal binding site for a destabilizing factor. The destabilizing effect depends on the number of copies of this proposed binding site and their degree of mismatch in the first two and last two positions, with mismatches in the AUUUA sequence not being tolerated. We found a strict correlation between the effect of an ARE on degradation rate and the effect on the rate of poly(A) shortening, consistent with deadenylation being the first and rate-limiting step in degradation, and the step stimulated by destabilizing AREs. Deadenylation was observed to occur in at least two phases, with an oligo(A) intermediate transiently accumulating, consistent with the suggestion that the degradation processes may be similar in yeast and mammalian cells. AREs that are especially U rich and contain no UUAUUUA(U/A)(U/A) motifs failed to influence the degradation rate or the deadenylation rate, either when downstream of suboptimal destabilizing AREs or when alone.

Selective degradation of early-response-gene mRNAs: functional analyses of sequence features of the AU-rich elements

The metabolic lifetime of mRNA can be specified by specific cis-acting elements within mRNA. One type of element is an adenylate- and uridylate-rich element (ARE) found in the 3' untranslated region of many highly unstable mRNAs for mammalian early-response genes (ERGs). Among the better-characterized members of the ERG family are certain genes encoding nuclear transcription factors. Of particular significance was the finding that their mRNAs decay rapidly with kinetics similar to those of c-fos mRNA. Our previous studies of the c-fos ARE-directed mRNA decay have revealed the existence in this ARE of two structurally distinct and functionally interdependent domains, termed domain I and domain II. We proposed that the c-fos ARE-directed decay is a two-step mechanism in which rapid shortening of the poly(A) tail leads to the decay of the mRNA body and further hypothesized that this is a general mechanism by which the ERG AREs mediate rapid mRNA degradation. To test this hypothesis and to further address the generality of the critical structural characteristics within the c-fos ARE, the RNA-destabilizing functions of more than 10 different AU-rich sequences from various nuclear transcription factor mRNAs have been tested. Consistent with the above-mentioned hypothesis is the observation that mRNAs carrying the functional AREs display a biphasic decay, which is characteristic of the proposed two-step mechanism. Our results indicated that the presence of AUUUA pentanucleotides in an AU-rich region does not always guarantee an RNA-destabilizing function for this region. Our results also led to the identification of a novel class of AU-rich destabilizing elements which contains no AUUUA pentanucleotide. The results of sequence comparison and functional tests revealed that a continuous U-rich sequence is a unique feature among the functional AREs. Finally, our experiments further showed that the c-fos ARE domain II has an RNA decay-enhancing ability upon its fusion to heterologous AU-rich regions and defined for the first time an RNA decay-enhancing element, which we termed the RDE element.

AUUUA is not sufficient to promote poly(A) shortening and degradation of an mRNA: the functional sequence within AU-rich elements may be UUAUUUA(U/A)(U/A)

AU-rich elements (AREs) in the 3' untranslated regions of several cytokine and oncogene mRNAs have been shown to function as signals for rapid mRNA degradation, and it is assumed that the many other cytokine and oncogene mRNAs that contain AU-rich sequences in the 3' untranslated region are similarly targeted for rapid turnover. We have used a chimeric gene composed mostly of growth hormone sequences with expression driven by the c-fos promoter to investigate the minimal sequence required to act as a functional destabilizing element and to monitor the effect of these sequences on early steps in the degradation pathway. We find that neither AUUUA, UAUUUA, nor AUUUAU can function as a destabilizing element. However, the sequence UAUUUAU, when present in three copies, is sufficient to destabilize a chimeric mRNA. We propose that this sequence functions by virtue of being a sufficient portion of the larger sequence, UUAUUUA(U/A)(U/A), that we propose forms the optimal binding site for a destabilizing factor. The destabilizing effect depends on the number of copies of this proposed binding site and their degree of mismatch in the first two and last two positions, with mismatches in the AUUUA sequence not being tolerated. We found a strict correlation between the effect of an ARE on degradation rate and the effect on the rate of poly(A) shortening, consistent with deadenylation being the first and rate-limiting step in degradation, and the step stimulated by destabilizing AREs. Deadenylation was observed to occur in at least two phases, with an oligo(A) intermediate transiently accumulating, consistent with the suggestion that the degradation processes may be similar in yeast and mammalian cells. AREs that are especially U rich and contain no UUAUUUA(U/A)(U/A) motifs failed to influence the degradation rate or the deadenylation rate, either when downstream of suboptimal destabilizing AREs or when alone.

Selective degradation of early-response-gene mRNAs: functional analyses of sequence features of the AU-rich elements

The metabolic lifetime of mRNA can be specified by specific cis-acting elements within mRNA. One type of element is an adenylate- and uridylate-rich element (ARE) found in the 3' untranslated region of many highly unstable mRNAs for mammalian early-response genes (ERGs). Among the better-characterized members of the ERG family are certain genes encoding nuclear transcription factors. Of particular significance was the finding that their mRNAs decay rapidly with kinetics similar to those of c-fos mRNA. Our previous studies of the c-fos ARE-directed mRNA decay have revealed the existence in this ARE of two structurally distinct and functionally interdependent domains, termed domain I and domain II. We proposed that the c-fos ARE-directed decay is a two-step mechanism in which rapid shortening of the poly(A) tail leads to the decay of the mRNA body and further hypothesized that this is a general mechanism by which the ERG AREs mediate rapid mRNA degradation. To test this hypothesis and to further address the generality of the critical structural characteristics within the c-fos ARE, the RNA-destabilizing functions of more than 10 different AU-rich sequences from various nuclear transcription factor mRNAs have been tested. Consistent with the above-mentioned hypothesis is the observation that mRNAs carrying the functional AREs display a biphasic decay, which is characteristic of the proposed two-step mechanism. Our results indicated that the presence of AUUUA pentanucleotides in an AU-rich region does not always guarantee an RNA-destabilizing function for this region. Our results also led to the identification of a novel class of AU-rich destabilizing elements which contains no AUUUA pentanucleotide. The results of sequence comparison and functional tests revealed that a continuous U-rich sequence is a unique feature among the functional AREs. Finally, our experiments further showed that the c-fos ARE domain II has an RNA decay-enhancing ability upon its fusion to heterologous AU-rich regions and defined for the first time an RNA decay-enhancing element, which we termed the RDE element.

Efficient translation of poly(A)-deficient mRNAs in Saccharomyces cerevisiae

The polyadenylate tail of eukaryotic mRNAs is thought to influence various metabolic phenomena including mRNA stability, translation initiation, and nucleo-cytoplasmic transport. We have analyzed the fate of mRNAs following inactivation of poly(A) polymerase in Saccharomyces cerevisiae containing a temperature-sensitive, lethal mutation (pap1-1) in the gene for poly(A) polymerase (PAP1). Inactivation of poly(A) polymerase (Pap1) by shifting cells to the nonpermissive temperature resulted in the loss of at least 80% of measurable poly(A) within 60 min. Northern blot analysis revealed the disappearance of some mRNAs (CYH2 and HIS4) consistent with a role for poly(A) tails in mRNA stability. However, other mRNAs (TCM1, PAB1, ACT1, and HTB2) accumulate as poly(A)-deficient (A < approximately 25) transcripts as defined by an inability to bind oligo(dT)-cellulose. Sucrose density gradient analysis of polyribosomes revealed a twofold reduction in the amount of each size class of polyribosomes in shifted cells and a commensurate increase in free ribosomes. However, poly(A)-deficient mRNAs in shifted cells remain associated with the same size polyribosomes as poly(A)+ mRNAs in unshifted cells, indicating normal initiation of translation. RNase mapping of transcripts from pap1-1 cells revealed PAB1 mRNA to be poly(A)- whereas TCM1 exists as equal amounts of poly(A)- and poly(A)+ mRNA 60 min after shift. Interestingly, both of these classes of TCM1 mRNA appear in similar amounts in each polyribosome fraction indicating that ribosomes may not distinguish between them. These findings suggest that under conditions of excess translational capacity, poly(A)- and poly(A)+ mRNAs may initiate translation with comparable efficiencies.

Inactivation of SSM4, a new Saccharomyces cerevisiae gene, suppresses mRNA instability due to rna14 mutations

Decay rates of mRNAs depend on many elements and among these, the role of the poly(A) tail is now well established. In the yeast Saccharomyces cerevisiae, thermosensitive mutations in two genes, RNA14 and RNA15, result in mRNAs having shorter poly(A) tails and reduced half-life. To identify other components interacting in the same process, we have used a genetic approach to isolate mutations that suppress the thermosensitivity of an rna14 mutant strain. Mutations in a single locus, named SSM4, not only suppress the cell growth phenotype but also the mRNA instability and extend the short mRNA poly(A) tails. The frequency of appearance and the recessive nature of these mutations suggested that the suppressor effect was probably due to a loss of function. We failed to clone the SSM4 gene directly by complementation, owing to its absence from gene banks; it later emerged that the gene is toxic to Escherichia coli, but we have nevertheless been able to clone the SSM4 sequence by Ty element transposition tagging. Disruption of the SSM4 gene does not affect cell viability and suppresses the rna14 mutant phenotypes. The protein encoded by the SSM4 gene has a calculated molecular mass of 151 kDa and does not contain any known motif or show homology with known proteins. The toxicity of the SSM4 gene in E. coli suggests that a direct biochemical activity is associated with the corresponding protein.

Metabolic stability of mRNA in yeast--a potential target for modulating productivity?

Attempts to improve the production of (heterologous) proteins in yeast cells have, to date, focused almost exclusively on increasing the transcriptional yield of the heterologous gene by raising the number of gene copies per cell, and/or putting the gene under the control of a strong homologous promoter. However, the cellular level of translatable mRNA is a function of the rate at which it is produced and the rate at which it is removed--or at least inactivated--by nucleolytic degradation. Recently, considerable progress has been made in unravelling the mechanism of mRNA decay in yeast cells and in identifying both the cis-acting stability determinants and the trans-acting factors involved in this process. This knowledge can be used as the basis for rational engineering of a given transcript to modulate its metabolic stability, and thus its cellular level.

Premature translational termination triggers mRNA decapping

The degradation of messenger RNA in eukaryotic cells is initiated by endonucleolytic cleavage or by shortening of the poly(A) tail, which for some mRNAs activates a deadenylation-dependent decapping reaction. One type of rapid mRNA degradation in eukaryotes is caused by premature termination of translation. This turnover process prevents the translation of aberrant mRNAs, may affect the abundance and splicing pattern of nuclear transcripts, and may be involved in the aetiology of human genetic disease. Here we show that premature translational termination in yeast triggers decapping, independent of deadenylation, thereby exposing the transcript to 5'-to-3' degradation. Inactivation of the 5'-to-3' exonuclease reveals an additional 3'-to-5' pathway of mRNA turnover. These observations provide in vivo evidence for two new mechanisms of mRNA decay.

Mechanisms of mRNA degradation in eukaryotes

Recent experiments have identified distinct mechanisms of eukaryotic RNA turnover. In one mechanism, deadenylation triggers decapping, exposing the messenger RNA to 5' to 3' degradation. This pathway may act at different rates on the majority of messenger RNAs. There are also degradation mechanisms, such as endonucleolytic cleavage, limited to messenger RNAs containing specific sequence elements.

An amino-terminal tetrapeptide specifies cotranslational degradation of beta-tubulin but not alpha-tubulin mRNAs

The steady-state level of alpha- and beta-tubulin synthesis is autoregulated by a posttranscriptional mechanism that selectively alters alpha- and beta-tubulin mRNA levels in response to changes in the unassembled tubulin subunit concentration. For beta-tubulin mRNAs, previous efforts have shown that this is the result of a selective mRNA degradation mechanism which involves cotranslational recognition of the nascent amino-terminal beta-tubulin tetrapeptide as it emerges from the ribosome. Site-directed mutagenesis is now used to determine that the minimal sequence requirement for conferring the full range of beta-tubulin autoregulation is the amino-terminal tetrapeptide MR(E/D)I. Although tubulin-dependent changes in alpha-tubulin mRNA levels are shown to result from changes in cytoplasmic mRNA stability, transfection of wild-type and mutated alpha-tubulin genes reveals that alpha- and beta-tubulin mRNA degradation is not mediated through a common pathway. Not only does the amino-terminal alpha-tubulin tetrapeptide MREC fail to confer regulated mRNA degradation, neither wild-type alpha-tubulin transgenes nor an alpha-tubulin gene mutated to encode an amino-terminal MREI yields mRNAs that are autoregulated. Further, although slowing ribosome transit accelerates the autoregulated degradation of endogenous alpha- and beta-tubulin mRNAs, degradation of alpha-tubulin transgene mRNAs is not enhanced, and in one case, the mRNA is actually stabilized. We conclude that, despite similarities, alpha- and beta-tubulin mRNA destabilization pathways utilize divergent determinants to link RNA instability to tubulin subunit concentrations.

An amino-terminal tetrapeptide specifies cotranslational degradation of beta-tubulin but not alpha-tubulin mRNAs

The steady-state level of alpha- and beta-tubulin synthesis is autoregulated by a posttranscriptional mechanism that selectively alters alpha- and beta-tubulin mRNA levels in response to changes in the unassembled tubulin subunit concentration. For beta-tubulin mRNAs, previous efforts have shown that this is the result of a selective mRNA degradation mechanism which involves cotranslational recognition of the nascent amino-terminal beta-tubulin tetrapeptide as it emerges from the ribosome. Site-directed mutagenesis is now used to determine that the minimal sequence requirement for conferring the full range of beta-tubulin autoregulation is the amino-terminal tetrapeptide MR(E/D)I. Although tubulin-dependent changes in alpha-tubulin mRNA levels are shown to result from changes in cytoplasmic mRNA stability, transfection of wild-type and mutated alpha-tubulin genes reveals that alpha- and beta-tubulin mRNA degradation is not mediated through a common pathway. Not only does the amino-terminal alpha-tubulin tetrapeptide MREC fail to confer regulated mRNA degradation, neither wild-type alpha-tubulin transgenes nor an alpha-tubulin gene mutated to encode an amino-terminal MREI yields mRNAs that are autoregulated. Further, although slowing ribosome transit accelerates the autoregulated degradation of endogenous alpha- and beta-tubulin mRNAs, degradation of alpha-tubulin transgene mRNAs is not enhanced, and in one case, the mRNA is actually stabilized. We conclude that, despite similarities, alpha- and beta-tubulin mRNA destabilization pathways utilize divergent determinants to link RNA instability to tubulin subunit concentrations.

Deadenylation of the unstable mRNA encoded by the yeast MFA2 gene leads to decapping followed by 5'-->3' digestion of the transcript

The first step in the decay of some eukaryotic mRNAs is the shortening of the poly(A) tail. To examine how the transcript body was degraded after deadenylation, we followed the decay of a pulse of newly synthesized MFA2 transcripts while utilizing two strategies to trap intermediates in the degradation pathway. First, we inserted strong RNA secondary structures, which can slow exonucleolytic digestion and thereby trap decay intermediates, into the MFA2 5' UTR. Following deadenylation, fragments of the MFA2 mRNA trimmed from the 5' end to the site of secondary structure accumulated as full-length mRNA levels decreased. In addition, in cells deleted for the XRN1 gene, which encodes a major 5' to 3' exonuclease in yeast, the MFA2 transcript is deadenylated normally but persists as a full-length mRNA lacking the 5' cap structure. These results define a mRNA decay pathway in which deadenylation leads to decapping of the mRNA followed by 5'-->3' exonucleolytic degradation of the transcript body. Because the poly(A) tail and the cap structure are found on essentially all mRNAs, this pathway could be a general mechanism for the decay of many eukaryotic transcripts.

Interplay of two functionally and structurally distinct domains of the c-fos AU-rich element specifies its mRNA-destabilizing function

AU-rich elements (ARE) in the 3' untranslated region of many highly labile mRNAs for proto-oncogenes, lymphokines, and cytokines can act as an RNA-destabilizing element. The absence of a clear understanding of the key sequence and structural features of the ARE that are required for its destabilizing function has precluded the further elucidation of its mode of action and the basis of its specificity. Combining extensive mutagenesis of the c-fos ARE with in vivo analysis of mRNA stability, we were able to identify mutations that exhibited kinetic phenotypes consistent with the biphasic decay characteristic of a two-step mechanism: accelerated poly(A) shortening and subsequent decay of the transcribed portion of the mRNA. These mutations, which affected either an individual step or both steps, all changed the mRNA stability. Our experiments further revealed the existence of two structurally distinct and functionally interdependent domains that constitute the c-fos ARE. Domain I, which is located within the 5' 49-nucleotide segment of the ARE and contains the three AUUUA motifs, can function as an RNA destabilizer by itself. It forms the essential core unit necessary for the ARE-destabilizing function. Domain II is a 20-nucleotide U-rich sequence which is located within the 3' part of the c-fos ARE. Although it alone can not act as an RNA destabilizer, this domain serves two critical roles: (i) its presence enhances the destabilizing ability of domain I by accelerating the deadenylation step, and (ii) it has a novel capacity of buffering decay-impeding effects exerted by mutations introduced within domain I. A model is proposed to explain how these critical structural features may be involved in the c-fos ARE-directed mRNA decay pathway. These findings have important implications for furthering our understanding of the molecular basis of differential mRNA decay mediated by different AREs.

Interplay of two functionally and structurally distinct domains of the c-fos AU-rich element specifies its mRNA-destabilizing function

AU-rich elements (ARE) in the 3' untranslated region of many highly labile mRNAs for proto-oncogenes, lymphokines, and cytokines can act as an RNA-destabilizing element. The absence of a clear understanding of the key sequence and structural features of the ARE that are required for its destabilizing function has precluded the further elucidation of its mode of action and the basis of its specificity. Combining extensive mutagenesis of the c-fos ARE with in vivo analysis of mRNA stability, we were able to identify mutations that exhibited kinetic phenotypes consistent with the biphasic decay characteristic of a two-step mechanism: accelerated poly(A) shortening and subsequent decay of the transcribed portion of the mRNA. These mutations, which affected either an individual step or both steps, all changed the mRNA stability. Our experiments further revealed the existence of two structurally distinct and functionally interdependent domains that constitute the c-fos ARE. Domain I, which is located within the 5' 49-nucleotide segment of the ARE and contains the three AUUUA motifs, can function as an RNA destabilizer by itself. It forms the essential core unit necessary for the ARE-destabilizing function. Domain II is a 20-nucleotide U-rich sequence which is located within the 3' part of the c-fos ARE. Although it alone can not act as an RNA destabilizer, this domain serves two critical roles: (i) its presence enhances the destabilizing ability of domain I by accelerating the deadenylation step, and (ii) it has a novel capacity of buffering decay-impeding effects exerted by mutations introduced within domain I. A model is proposed to explain how these critical structural features may be involved in the c-fos ARE-directed mRNA decay pathway. These findings have important implications for furthering our understanding of the molecular basis of differential mRNA decay mediated by different AREs.