Mutations in VPS16 and MRT1 Stabilize mRNAs by Activating an Inhibitor of the Decapping Enzyme

Abstract

Decapping is a rate-limiting step in the decay of many yeast mRNAs; the activity of the decapping enzyme therefore plays a significant role in determining RNA stability. Using an in vitro decapping assay, we have identified a factor, Vps16p, that regulates the activity of the yeast decapping enzyme, Dcp1p. Mutations in the VPS16 gene result in a reduction of decapping activity in vitro and in the stabilization of both wild-type and nonsense-codon-containing mRNAs in vivo. The mrt1-3 allele, previously shown to affect the turnover of wild-type mRNAs, results in a similar in vitro phenotype. Extracts from both vps16 and mrt1 mutant strains inhibit the activity of purified Flag-Dcp1p. We have identified a 70-kDa protein which copurifies with Flag-Dcp1p as the abundant Hsp70 family member Ssa1p/2p. Intriguingly, the interaction with Ssa1p/2p is enhanced in strains with mutations in vps16 or mrt1. We propose that Hsp70s may be involved in the regulation of mRNA decapping.

Rates of decay of individual mRNAs within a cell vary extensively. Some mRNAs turn over with half-lives that are on the order of minutes, while other transcripts are stable for days or weeks (reviewed in references 3, 6, 23, 34, 36, and 46). In addition, the rates of decay of certain mRNAs are modulated in response to environmental signals. Taken together, these results demonstrate that mRNA stability is an important mode of regulating gene expression. The goal of a number of laboratories has been the elucidation of the cellular mechanisms that control mRNA decay rates. These studies include the characterization of both the cis-acting elements and the trans-acting factors that are involved in the decay process.

Two mRNA turnover pathways that have been intensively investigated include the poly(A)-shortening-dependent and the nonsense-mediated decay (NMD) [poly(A)-shortening-independent] pathways (reviewed in references 3, 6, 7, 23, and 36). The poly(A)-shortening-dependent pathway has been demonstrated to degrade a large number of wild-type transcripts in both yeast and mammalian cells (12, 30, 41, 42). Removal of the poly(A) tail appears to be the first step in this pathway and precedes the degradation of the body of the transcripts. Mutations that cripple or inactivate sequence elements that promote poly(A) shortening result in the stabilization of these mRNAs. On the other hand, the NMD pathway in both yeast and mammalian cells degrades nonsense transcripts prior to shortening of the poly(A) tail (31, 42).

The rate-limiting events in the poly(A)-shortening-dependent and NMD pathways have been investigated most extensively with yeast cells and, most recently, mammalian cells (reviewed in references 3, 6, 8, 23, and 37). For the poly(A)-shortening-dependent pathway, the 5′ cap structure is removed by a decapping enzyme following deadenylation. Removal of the 5′ cap is also a rate-limiting event in the NMD pathway. Following decapping, the uncapped transcripts are degraded by the 5′→3′ Xrn1 exoribonuclease (4, 12, 17, 21, 29, 31).

The fact that decapping is an important event in the decay of both wild-type and nonsense transcripts indicates that the enzymes and factors that carry out and regulate this activity are important in controlling mRNA turnover rates. A highly purified yeast decapping activity has recently been isolated, and its gene has been identified (4, 26). Microsequencing of this protein led to the isolation of the DCP1 gene. Cells harboring a dcp1 disruption were viable, with a twofold increase in doubling time, and promoted the stabilization of mRNAs in both the poly(A)-shortening-dependent and the NMD pathways (4, 18). Further analysis demonstrated that the poly(A) tails of these transcripts were deadenylated normally but that their decapping rates were greatly reduced. Dcp1p has recently been shown to be a phosphoprotein (26). Phosphorylation may therefore regulate the activity of Dcp1p.

Mutations in two potential factors that regulate decapping activity have recently been described (18). Strains harboring the mrt1 and mrt3 alleles demonstrated reduced decapping activity. At present, however, the MRT1 and MRT3 genes have not been isolated, and the basis for the reduction in cellular decapping activity is not known.

In this report, we identify a novel regulator of the yeast decapping enzyme by using a biochemical screen. Mutations in the VPS16 gene result in a reduction of decapping activity in vitro and in the stabilization of wild-type and nonsense-codon-containing mRNAs in vivo. We show that extracts from vps16 and mrt1 (a gene previously shown to affect the turnover of mRNAs) mutant strains are able to inhibit the activity of purified Flag-Dcp1p. Purification of Flag-Dcp1p from wild-type and mutant strains revealed a 70-kDa interacting protein whose association was enhanced when Dcp1p was isolated from either vps16 or mrt1-3 strains. This protein was subsequently identified as the abundant Hsp70 family member Ssa1p/2p. We propose that the activity of the decapping enzyme may be regulated by Ssa1p/2p, and the implications are discussed.

MATERIALS AND METHODS

Strains, media, and general methods.

The yeast strains used in this study are listed in Table 1. Escherichia coli DH5α was used to amplify plasmid DNA. Yeast medium was prepared as described previously (40). Yeast transformations were performed by the lithium acetate method (22) as modified by Schiestl and Gietz (39).

TABLE 1.

Yeast strains used in this study

Strain(s)	Genotypea	Reference or source
Y137	MATa ura3-1 ade2-1(oc) his3-11,15 leu2-3,112 can1-100	32
Y138	MATα ura3-1 ade2-1(oc) trp1-1 leu2-3,112 can1-100	32
27	MATa ura3-1 ade2-1(oc) his3-11,15 leu2-3,112 can1-100 vps16-100	This study
466	MATa ura3-1 ade2-1(oc) his3-11,15 leu2-3,112 can1-100 dcp3-2	This study
Y368	MATa ura3-1 ade2-1(oc) his3-11,15 leu2-3,112 can1-100 vps16::hisG	This study
Y361	MATa ura3-1 ade2-1(oc) his3-11,15 leu2-3,112 can1-100 vps16	This study
Y132	MATa ura3-1 ade2-1(oc) his3-11,15 leu2-3,112 can1-100 ts+	32
	MATα ura3-1 ade2-1(oc) trp1-1 leu2-3,112 can1-100 ts+
YRP1070	MATa ura3-52 trp1-1 leu2-3,112 his4-539 dcp1::URA3 cup1::LEU2/MFA29/PGK19	18
vps and pep collection	wt; vps8 vps16 vps1	E. W. Jonesb
	wt; vps5 vps2 vps20
	wt; pep5-8 pep3-12 pep14-5 pep7-15 pep12-2
Y187	MATa mrt1-3 ura3-52 leu2-3,112 his4-539 trp1 cup1::LEU2/MFAZp9/PGK1p9	18

^awt, wild type. 

^bCarnegie Mellon University, Pittsburgh, Pa.  

Materials.

Restriction enzymes were obtained from Boehringer Mannheim, New England Biolabs, and Bethesda Research Laboratories. Radioactive nucleotides were obtained from Amersham ([α-³²P]dCTP and [α-³²P]GTP). Oligonucleotides used in these studies were purchased from Integrated DNA Technologies. Mass spectrometry was performed at the W. M. Keck Foundation in New Haven, Conn. Flag peptide, anti-Flag M2 antibody, and anti-Flag M2 affinity resin were obtained from Sigma.

Screening of a decapping-deficient temperature-sensitive (ts) yeast library.

Extracts were prepared by harvesting 10 ml of cells grown to an optical density at 600 nm of 0.7 to 1.0 and washing the cells with 1 ml of cold buffer XO (10 mM Tris [pH 7.6], 100 mM potassium acetate, 2 mM magnesium acetate, 2 mM dithiothreitol). The cells were resuspended in an equal volume of cold buffer XOR (buffer XO containing 2 μg each of leupeptin, pepstatin A, and aprotinin per ml), and an equivalent volume of acid-washed glass beads was added to the cells. The mixture was vortexed for 15 s and cooled on ice for 45 s. The vortexing protocol was repeated a total of six times. The extract was centrifuged at 13,000 rpm for 10 min at 4°C in a microcentrifuge, the supernatant was transferred to a fresh tube, and the protein concentration of the extract was determined. The decapping assay was performed as described previously (47).

mRNA decay rates.

mRNA decay rates were determined for wild-type strain W303, strains Y137 and Y138, and strains 27 and 466 (Table 2). The half-life of nonsense allele PGK1 expressed from the centromere-based plasmid pRIP1PGK(−AU)1UAAIN1 (17) was determined. Cultures (100 ml) of yeast cells were grown to the mid-log phase (A₆₀₀, 0.5 to 0.7) at 24°C on yeast extract-peptone-dextrose or medium lacking uracil, centrifuged, and resuspended in 36 ml of the same medium preheated at 36°C. After 0 or 20 min of incubation at 37°C, transcription was inhibited by adding 15 μg of Thiolutin per ml. At various times following transcription inhibition, 4-ml aliquots were removed and cells were collected by rapid centrifugation. The cell pellets were frozen in a dry ice-ethanol bath. Equal amounts (approximately 10 to 20 μg) of total RNA from each time point were analyzed by Northern blotting. The results of this analysis were quantified with either a Bio-Rad phosphorimager or densitometer. Half-lives were determined by expressing the data as the log of the percentage of each RNA remaining at time zero versus the time at which transcription was initially inhibited. mRNA decay rate measurements varied ±15%.

TABLE 2.

mRNA half-lives^a

Transcript	Half-life (min) of the following strain:
	Wild type	27	vps16 Δ
MFA2	5	20	18
URA4	4	19	20
Nonsense allele PGK1	<3	14	16

^aCell cultures grown to an optical density at 600 nm of 0.7 to 1.0 were collected, resuspended, and incubated at 37°C for 0 or 20 min before the addition of Thiolutin to a final concentration of 15 μg/ml. After inhibition of transcription, cells were harvested at various times. RNA was prepared, and Northern blotting was performed. 

DNA probes were labeled to a high specific activity with [α-³²P]dCTP (15). The DNAs used included the following: (i) the BglIII-HindIII fragment of the MFA2 gene, (ii) the BamHI-SspI fragment of the URA4 gene, (iii) the BamHI-HindIII fragment from the PGK1 gene, and (iv) the U3 gene.

Isolation and characterization of the VPS16 gene.

The VPS16 gene was cloned from a pYCp50 yeast genomic library (purchased from the American Type Culture Collection) prepared from a partial Sau3A digest of genomic DNA. Strains 27 and 466 were transformed with this library. Cells were grown at 24°C for 12 h and subsequently incubated at 37°C for 4 days. Totals of 40,000 Ura⁺ transformants for strain 27 and 5,500 Ura⁺ transformants for strain 466 were screened for plasmids that complemented the ts defect. Colonies that grew at 37°C on minimal medium lacking uracil were retested. Colonies that grew at 37°C were characterized further. To ensure that growth at 37°C depended on the gene products synthesized from genes on the plasmid, the cells were plated on 5-fluoro-orotic acid (5-FOA). 5-FOA selection identified cells that had lost the plasmid. Plasmids were isolated from cells that grew on 5-FOA and consequently demonstrated a ts growth phenotype. Five plasmids that, when transformed into strain 27 or 466, relieved the ts growth phenotype were isolated from the above-described screen. These plasmids were subsequently transformed and propagated in E. coli.

Primers that flanked the 5′ and 3′ ends of the yeast genomic DNA were used to obtain sequence information. The results demonstrated that there was an overlap among the four insertions, and a map of the yeast genomic DNA fragment revealed five open reading frames (ORFs). Plasmids containing subclones of the yeast genomic DNA fragment were constructed to identify the ORF that was responsible for complementing the ts growth phenotype of strains 27 and 466 (see Fig. 2A). The plasmids were transformed into strains 27 and 466, and the ts growth phenotype and decapping activity were subsequently analyzed as described in Results.

FIG. 2.

(A) The VPS16 gene complements both the ts and the decapping defects of strain 27. Two 9-kb genomic fragments that were able to complement the ts and decapping phenotypes were initially isolated when they were transformed into strain 27. A set of subclones that covers the different regions of the originally isolated genomic fragments was constructed. Their ability to complement the ts and decapping phenotypes of strain 27 were assayed, and the results are shown. +, ability to grow at 37°C or demonstrate decapping activity; −, converse of +. (B) Crossing with vps16Δ demonstrates that mutations in the VPS16Δ gene cause the ts phenotype of strains 27 and 466. The crosses of 466 and vps16Δ and of 27 and vps16Δ were duplicated, accounting for the larger segments shown on the right. wt, wild type.

Preparation of the VPS16 knockout plasmid.

pKOVPS16 was constructed in order to delete the VPS16 gene from the yeast chromosome. An SstI-SphI fragment containing the VPS16 gene was subcloned into pUC19. The StuI-SpeI fragment containing the entire VPS16 ORF was replaced with a DNA fragment containing the hisG-URA3-hisG cassette (1). Strains harboring a VPS16 disruption can be constructed by transforming yeast cells with a DNA fragment containing the hisG cassette and selecting for growth on medium lacking uracil. A disruption of the VPS16 gene was confirmed by Southern blotting analysis.

Purification of epitope-tagged Dcp1p, Vps16p, and Upf1p.

Flag-tagged alleles of DCP1 and VPS16 were constructed by PCR in which the Flag epitope was fused to the 5′ end of the ORF and inserted 3′ of the glyceraldehyde-3-phosphate dehydrogenase promoter in plasmid pG-1 (38). The Flag-tagged DCP1 allele was shown to complement a dcp1 strain in terms of both decapping activity and increasing growth rates (data not shown). The plasmids were transformed into wild-type (VPS16⁺), vps16, or mrt1-3 strains. The Flag-tagged proteins were purified on an anti-Flag antibody affinity column as described previously (11, 45) (see Fig. 5A). The column was subsequently washed three times with 5 ml of low-salt buffer, and the Flag-conjugated proteins were eluted with buffer containing Flag peptide. The collections were dialyzed against storage buffer (50 mM Tris-HCl [pH 7.5], 1 mM MgCl₂, 100 mM NaCl, 50% glycerol, 1 mM phenymethylsulfonyl fluoride, 1 mM dithiothreitol, 0.1% Triton X-100) overnight. ATP or ADP washing of the column was performed prior to Flag elution with ATP elution buffer (50 mM Tris-HCl [pH 7.5], 10 mM MgCl₂, 200 mM KCl, 2 mM ATP [or ADP], 10% glycerol). The proteins in each fraction were resolved by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE). Protein staining and Western blotting analysis were performed as described previously (11).

FIG. 5.

(A) Ssa1p/2p copurifies with Dcp1p. An anti-Flag antibody column loaded with cell extracts was washed and eluted with Flag peptide, and proteins were analyzed on a Coomassie blue-stained SDS-polyacrylamide gel. The sizes of markers are shown on the left. The identity of each band on the gel is shown on the right. Lane 1, fraction bound to the anti-Flag antibody column when the cells contained an empty vector; lanes 2 to 4, Flag-Dcp1p purified from wild-type (wt), vps16Δ, and mrt1-3 strains with the Flag peptide, respectively; lane 5, Flag-Vps16p purified from the wild-type strain; lanes 6 and 7, Flag-Upf1p purified from wild-type and vps16Δ strains, respectively. (B) Ssa1p/2p can be separated from Flag-Dcp1p by washing the column with 2 mM ATP. The Coomassie blue-stained gel shows proteins purified from a vps16Δ strain expressing Flag-Dcp1p. Lane 1, proteins are eluted when the column is washed with Flag peptide; lane 2, nothing is eluted after the column is washed with 2 mM ADP; lane 3, Ssa1p/2p can be eluted when the column is washed with 2 mM ATP; lane 4, Flag-Dcp1p is eluted with Flag peptide after the column is washed with 2 mM ATP. (C) When separated from Ssa1p/2p, Flag-Dcp1p is equally active whether purified from wild-type or mutant strains. Flag-Dcp1p (200 ng) purified from wild-type, mrt1-3, and vps16Δ strains and separated from Ssa1p/2p by washing with 2 mM ATP was used in a decapping assay. The m⁷GDP product of each reaction is depicted.

RESULTS

Isolation of strains with putative decapping defects.

We have developed an in vitro decapping assay which has been distributed and successfully used by other investigators prior to this publication (4, 18, 47). The assay uses thin-layer chromatography to monitor decapping activities in cell extracts (47; see also Materials and Methods). As expected, m⁷GDP was obtained as a reaction product (data not shown). To isolate putative decapping-defective strains, a collection of ts yeast strains derived from segregants of strain W303 were screened for mutations that altered decapping activity in yeast extracts (see Materials and Methods). Strains were grown at the permissive temperature, and extracts were prepared. Reaction mixtures containing equal amounts of protein derived from the cell extracts but lacking the cap-labelled substrate were incubated at 37°C for 10 min and subsequently placed on ice for 5 min. The capped RNA substrate was added to the reaction mixtures and incubated for an additional 40 min at 37°C. The reaction products of the decapping reaction were analyzed by thin-layer chromatography, and the decapping activities from these extracts were compared with the decapping activity from an extract prepared from wild-type cells.

Two strains exhibiting reduced decapping activity in this screen were extensively characterized. Extracts prepared from strains 27 and 466 demonstrated approximately a six- to eightfold reduction in decapping activities compared with an extract prepared from either a wild-type strain or another ts strain (Fig. 1A, compare wild-type strain and strain 26 with strain 27; data not shown for strain 466). Complementation analysis showed that strains 27 and 466 were allelic (data not shown).

FIG. 1.

(A) Identification of ts strain 27 with decreased decapping activity in vitro. A ts yeast strain collection was screened for decreased decapping activity by the in vitro decapping assay. Cell extracts were prepared from wild-type cells, ts isolate 26, and ts isolate 27. Decapping reactions were performed as outlined previously (47). The positions at which the products of the reaction migrated are indicated. Cap-labelled RNA was incubated in a reaction mixture lacking cell extract as a negative control (substrate lane). (B) (Top) Increased abundance of the PGK1 nonsense-codon-containing transcript in decapping-defective strain 27. The wild-type strain and strain 27 harboring the PGK1 nonsense allele on a centromere-based plasmid (see Materials and Methods) were grown to the mid-log phase, and cells were harvested at various times (minutes) after the culture was shifted to 37°C. Total RNA was prepared and subjected to RNA blotting analysis with a radioactive DNA fragment specific for the PGK1 transcript as a probe. wt, wild type. (Bottom) Plot of the abundance of the PGK1 nonsense transcript, normalized to the abundance of the wild-type PGK1 mRNA, at various times (minutes) after the cells were shifted to a nonpermissive temperature. (C) PGK1 nonsense mRNA and MFA2 and URA4 transcripts were stabilized in decapping-deficient strain 27. The wild-type strain and strain 27 were grown to the mid-log phase, and the cells were shifted to 37°C for either 0 or 20 min (only the results for 37°C and 0 min are shown). After the temperature shift, 15 μg of Thiolutin per ml was added to inhibit transcription. Cell cultures were harvested at various times (minutes) after transcription inhibition. U3 small nucleolar RNA (snoRNA) was also probed as an internal control. The half-life was determined by Northern blotting analysis as described in Materials and Methods.

Characterization of mRNA stability in strains that are putatively deficient in decapping activity.

We next determined whether strain 27 manifested an in vivo mRNA metabolism defect. As an initial test, the abundance of a PGK1 nonsense transcript in mutant and wild-type cells was monitored. The PGK1 nonsense transcript is normally rapidly degraded in a decapping-dependent manner by the NMD pathway (17, 31). The abundance of this transcript was determined before and after a temperature shift in wild-type and mutant strains. The results demonstrated that the abundance of the PGK1 nonsense transcript, normalized to the abundance of the wild-type PGK1 mRNA, increased approximately six- to sevenfold in strain 27 cells compared with wild-type cells (Fig. 1B). An increased accumulation of the PGK1 nonsense-codon-containing mRNA was observed in cells prior to the temperature shift (Fig. 1B, time zero), suggesting that strain 27 demonstrated deficient decapping activity at the permissive temperature. Consistent with this view, extracts from strain 27 that were not shifted to 37°C demonstrated reduced decapping activity (data not shown).

The results described above indicated that the mutation in strain 27 affected the abundance of an unstable mRNA that uses decapping in its turnover pathway. We next determined the rates of decay of mRNAs in wild-type and putative decapping-deficient strains. Both wild-type cells and mutant strain 27 cells were grown to the mid-log phase at 24°C and subsequently shifted to 37°C for either 0 or 20 min. Rates of decay of the MFA2, URA4, and PGK1 nonsense transcripts were determined by RNA blotting analysis of RNAs isolated at different times after inhibition of transcription by Thiolutin addition (see Materials and Methods). Previous studies have shown that inhibition of transcription by Thiolutin accurately reflects mRNA decay rates in yeast cells (19). Thiolutin was shown to inhibit the transcription of both wild-type and mutant strains by monitoring the incorporation of [³H]uridine (data not shown). The MFA2, URA4, and PGK1 nonsense transcripts were found to be rapidly degraded in wild-type cells but were stabilized five- to sevenfold in strain 27 cells, even when the cells were not preincubated at 37°C (Fig. 1C and Table 2). Similar results were obtained when the cells were preincubated at 37°C for 20 min (data not shown). These results indicated that both wild-type and nonsense mRNAs were stabilized in strain 27 cells, consistent with the notion that the reduced decapping activity in these cells leads to the stabilization of both wild-type and nonsense-codon-containing mRNAs.

Genetic characterization of putative decapping-deficient ts strains.

We examined whether strains 27 and 466 demonstrated a recessive or a dominant phenotype. These strains were crossed with an isogenic wild-type strain, and the growth phenotypes of the diploids were monitored. Both the ts growth phenotype and reduced decapping activity in the diploid extracts were recessive (data not shown). We also examined whether the ts growth phenotype of strain 27 cosegregated with the reduced decapping activity. To do so, strain 27 was mated with a wild-type strain, and the diploid cells were induced to sporulate. The four spores from each tetrad were grown at 24°C, replica plated, and grown at 37°C. Both the ts growth phenotype and the reduced decapping activity cosegregated in a 2:2 segregation pattern, indicating that the lesion causing reduced decapping activity was a result of a single ts allele (data not shown).

Cloning of the wild-type form of the gene from strains 27 and 466.

The wild-type gene was isolated by transforming strains 27 with a centromere-based yeast library, selecting for cells able to grow at 37°C, and determining whether extracts from these cells demonstrated wild-type decapping activity. Five plasmids with these characteristics were isolated, and transformation of these plasmids into strains 27 and 466 relieved the ts growth phenotype and increased the decapping activity observed in cell extracts prepared from these strains (data not shown). Sequencing of the 5′ and 3′ ends of the yeast genomic fragments identified overlapping fragments on chromosome 16 containing six potential ORFs, including NOP4, VPS16/VAM9/CLS17, CAM1, SRB10, and two small ORFs that have not been previously described (Fig. 2A). Additional subcloning and assaying for relief of the ts growth phenotype and reduced decapping activity in strain 27 indicated that the VPS16 gene complemented these defects (Fig. 2A). Consistent with strains 27 and 466 harboring mutations in the VPS16 gene, crossing strains 27 and 466 with a previously identified strain harboring a vps16 mutation (2) did not relieve the ts growth defect (Fig. 2B). Taken together, these results indicate that mutations in the VPS16 gene can affect decapping activity. Mutations in the VPS16/VAM9/CLS17 gene were previously identified in a variety of screens for mutations that altered either vacuolar functioning or morphology as well as cell wall biogenesis (see Discussion).

A vps16Δstrain has the same phenotype as strains harboring the vps16 alleles identified in the biochemical screen.

The VPS16 gene was deleted from wild-type strain Y137, and the growth and mRNA turnover phenotypes of the resulting strain were monitored. As in the mutant strains identified above, a vps16Δ mutation resulted in a ts growth phenotype at 37°C (Fig. 2B). Decapping activity in extracts prepared from the vps16 strain was reduced sevenfold compared to that in wild-type extracts (data not shown). Furthermore, both wild-type MFA2 and URA4 transcripts were stabilized sevenfold in the vps16 strain compared to in the wild-type strain (Table 2). Taken together, these results indicate that the mutations identified in the screen are located in the VPS16 gene and that a vps16Δ strain has the same mRNA decay phenotype as the ts strains that we have previously isolated.

Mutations in other genes that affect vacuolar biogenesis do not affect decapping activity.

The vacuolar morphology of strain 27 appeared normal, while strain 466 showed an altered morphology, suggesting that the decreased decapping activity of strain 27 may not be a consequence of a defect in vacuolar biogenesis (data not shown). If this notion is true, we predict that other vacuolar mutations will not result in a defect in decapping activity. To test this possibility, decapping activities were monitored in extracts prepared from strains harboring mutations in the VPS1, VPS2, VPS5, VPS8, VPS16, VPS20, PEP3, PEP5, PEP7, PEP12, and PEP14 genes. These alleles cover the entire spectrum of vacuolar mutations identified (35, 44). In particular, and like mutations in the VPS16 gene, the pep3, pep5, and pep14 mutations are also thought to be mutations in regulatory genes involved in vacuolar biogenesis. Strains harboring these alleles have abnormal vacuolar structures, and these alleles have also been suggested to be involved in vacuolar biogenesis, segregation, or inheritance (2). Strains harboring these mutations were grown to the mid-log phase, and decapping activity in the cell extracts was monitored as described above. The results of these experiments demonstrated that, unlike the vps16 strain, strains harboring the other mutations that affected vacuolar function demonstrated wild-type levels of decapping activity (Fig. 3). Only mutations in the VPS16 gene resulted in decreased decapping activity in the extracts.

FIG. 3.

Decapping activity of vps and pep mutants. Cell extracts were prepared from different classes of vps and pep mutants, and decapping activity was measured as described in Materials and Methods. vps8 is a class A mutant. vps5 is a class B mutant. Class C mutants include vps16, pep5-8, pep3-12, and pep14-5. pep12-2 is a class D mutant. The vps2 and vps20 mutants belong to class E. vps1 is a class F mutant. The bars were grouped to indicate that these mutants and the wild type (wt) share the same genetic background. The decapping activity of wild-type cells was defined as 100%.

The expression of Dcp1p is not affected in a vps16Δ strain.

We next asked whether the defect in decapping in a vps16 strain could be due to altered expression of the DCP1 gene. To test this possibility, the Dcp1p levels in extracts prepared from wild-type and vps16Δ strains were determined by Western blotting analysis. The results demonstrated that Dcp1p was expressed equivalently in extracts prepared from both wild-type and vps16Δ strains (data not shown).

We also determined whether the expression of Dcp1p from a strong heterologous promoter would suppress the vps16Δ ts growth defect. A DCP1 allele with a Flag epitope tag was expressed from the glyceraldehyde-3-phosphate dehydrogenase promoter (see Materials and Methods). Wild-type cells harboring the Flag-tagged DCP1 allele demonstrated greater decapping activity than cells lacking it (data not shown). Immunoblotting demonstrated that the expected band with an apparent molecular mass of approximately 34 kDa reacted with the anti-Flag antibody (data not shown). The overexpression of Dcp1p in a vps16Δ strain, however, did not alleviate either its ts growth or its reduced decapping activity phenotype (data not shown).

Extracts from vps16 and mrt1-3 strains contain a factor that inhibits the decapping activity of Dcp1p.

We investigated how deletion of the VPS16 gene results in reduced decapping activity. We hypothesized that an inhibitor of Dcp1p is activated in a vps16Δ strain. This notion was tested by monitoring the decapping activity of purified Dcp1p in the presence of various concentrations of extracts prepared from either wild-type or vps16Δ strains. Flag-Dcp1p used in this experiment was purified from extracts of wild-type yeast cells by affinity chromatography with an anti-Flag antibody column (see Materials and Methods). Reaction mixtures containing a cap-labelled RNA substrate, purified Dcp1p, and increasing concentrations of extracts from wild-type or mutant strains were prepared, and decapping activity was monitored as described above. The results of these experiments demonstrated that reaction mixtures containing extracts from vps16Δ cells inhibited the activity of purified Dcp1p (Fig. 4). Conversely, extracts from wild-type cells slightly increased decapping activity in these reactions (Fig. 4).

FIG. 4.

Cell extracts prepared from vps16Δ and mrt1-3 strains inhibit the decapping activity of purified Dcp1p. Purified Dcp1p (see Materials and Methods) was mixed with different amounts of cell extracts prepared from wild-type, vps16Δ, and mrt1-3 strains, and decapping activity was subsequently monitored. Decapping activity was plotted as a percentage of purified Dcp1p observed in the absence of extracts. The results did not vary by more than 10%.

The inhibitor appears to be heat labile, as boiling of the vps16Δ extracts prior to using them in the decapping assay abolished the inhibitory activity. The inhibitor also is not dialyzable, as dialyzed vps16Δ extracts did not regain decapping activity (data not shown).

We hypothesized that cells harboring the mrt1-3 allele, which stabilizes deadenylated mRNAs in vivo (18), may also activate an inhibitor of Dcp1p. We tested whether an extract prepared from an mrt1-3 strain would inhibit the decapping activity of Dcp1p as described above. As predicted, an extract from an mrt1-3 strain inhibited the activity of purified Dcp1p (Fig. 4). Although both mutations contain an inhibitor of decapping activity, complementation analysis demonstrated that VPS16 and MRT1 are not allelic (data not shown). Taken together, these results indicate that Dcp1p is inhibited by a component in extracts prepared from either a vps16Δ or an mrt1-3 strain.

Identification and characterization of heat shock protein Ssa1p/2p as a Dcp1p-interacting protein.

In principle, the inhibition of Dcp1p could be achieved by interaction with an inhibitory factor, by direct modification of Dcp1p itself, or by a combination of the two. We initially asked whether factors that interacted with Dcp1p from extracts prepared from vps16Δ or mrt1-3 but not wild-type cells could be isolated. Flag-Dcp1p was purified from wild-type, vps16Δ, and mrt1-3 extracts by affinity purification with an anti-Flag monoclonal antibody column (see Materials and Methods). The purified fractions were analyzed by SDS-PAGE and visualized by Coomassie blue staining. Intriguingly, an approximately 70-kDa protein copurified with Flag-Dcp1p from vps16 and mrt1-3 extracts but was much less abundant when Flag-Dcp1p was isolated from wild-type cells (Fig. 5A, compare lanes 3 and 4 to lane 2). The 70-kDa protein did not copurify with Flag-Upf1p isolated from wild-type or vps16 extracts (Fig. 5A, lanes 6 and 7), indicating that the interaction was not a consequence of the Flag epitope.

The identity of the protein that copurified with Dcp1p from vps16 or mrt1-3 extracts was determined by mass spectrometry of tryptic peptides. The results demonstrated that this protein was constitutively expressed Hsp70 encoded by either the SSA1 or the SSA2 gene. Because of the high homology between the proteins encoded by SSA1 and SSA2, mass spectrometry analysis was unable to differentiate between them (reviewed in reference 9). Consistent with this protein being Ssa1p/2p, the 70-kDa protein that copurified with Dcp1p in vps16 and mrt1-3 strains reacted with a monoclonal antibody raised against human Hsc70 in Western blotting analysis (data not shown). Interestingly, Ssa1/2p also copurified with Flag-Vps16p expressed in a wild-type strain (Fig. 5A, lane 5).

Taken together, these results indicate that there is an enhanced interaction of Ssa1p/2p with Dcp1p in extracts prepared from either vps16 or mrt1-3 strains. This enhanced interaction is not due to an increased expression of SSA1 or SSA2 in mrt1-3 and vps16 strains overexpressing Flag-Dcp1p, as the levels of Ssa1p/2p do not vary between strains in Western blots (data not shown).

We were able to separate the Ssa1p/2p protein from Flag-Dcp1p by washing the column with 2 mM ATP (Fig. 5B). ATP has previously been shown to modulate the interaction of Hsp70 with peptides (16, 49). When separated from Ssa1p/2p, Flag-Dcp1p had the same level of activity regardless of whether it was purified from a wild-type or a mutant strain (Fig. 5C). This result indicates that the inhibition of decapping activity in vps16 and mrt1-3 mutant strains is not due to a modification of the decapping enzyme but more likely is due to an interaction of Dcp1p with an inhibitory factor.

DISCUSSION

Decapping is a rate-limiting step in the decay of mRNAs by both the poly(A)-shortening-dependent and the NMD pathways. Thus, the regulation of decapping is vital to the mRNA turnover process. Using a combination of genetic, molecular, and biochemical approaches, we have identified factors that may be involved in modulating the decapping activity of Dcp1p in yeast cells. We showed that mutating or deleting the VPS16 gene led to a reduction of decapping activity in yeast extracts. Deletion of the VPS16 gene, as well as the presence of the mrt1-3 allele that was previously demonstrated to stabilize deadenylated mRNAs in yeast cells (18), led to a reduction of decapping activity in yeast extracts due to the activation of a heat-labile, nondialyzable inhibitor. Furthermore, these mutations promoted an increased interaction of the heat shock protein Ssa1p/2p with Flag-Dcp1p.

Deletion of the VPS16 gene results in inhibition of the decapping activity of Dcp1p.

Our results demonstrated that mutations in or deletion of the VPS16 gene resulted in reduced decapping activity and stabilization of cellular mRNAs. Mutations in the VPS16/VAM9/CLS17 gene were identified in a variety of screens for mutations that altered either vacuole functioning or morphology and the cell wall (28, 33, 35, 44). Because of the pleiotropic nature of mutations in the VPS16 gene, it was suggested that Vps16p is a regulator of vacuolar biogenesis. Unlike most vacuolar proteins, Vps16p is a hydrophilic, cytoplasmic protein which lacks secretion signals and posttranslational modification and does not enter the secretory pathway. Interestingly, Vps16p is a soluble protein that is part of a large cytoplasmic complex (20).

Mutations in mRNA turnover pathways can have pleiotropic effects on many cellular pathways (23). For example, mutations in the 5′→3′ exoribonuclease XRN1 gene have been isolated in a number of genetic and biochemical screens for factors involved in nuclear fusion, plasmid stability, and DNA strand exchange during genetic recombination (13, 14, 24, 25, 43). The simplest interpretation of these results is that a mutation in the XRN1 gene affects a number of decay pathways and that the different phenotypes are a consequence of altered mRNA turnover pathways. At present, the role of Vps16p in regulating mRNA turnover is not well understood. It is conceivable that mutations in Vps16p affect a number of cellular pathways in addition to affecting vacuolar biogenesis.

Heat shock protein Ssa1p/2p demonstrates an enhanced interaction with Dcp1p in mutant strains with decapping activity.

A clue as to how Vps16p may be involved in regulating decapping activity was uncovered when Flag-Dcp1p was purified from vps16Δ extracts. We demonstrated that constitutively and highly expressed Ssa1p/2p copurified with Dcp1p in vps16Δ and mrt1-3 extracts but to a much lesser extent in the wild-type extract.

Hsp70s are protein chaperones which have both an ATPase domain and a peptide-binding domain. Their activity can be modulated by partner proteins, some of which are homologues of the E. coli DnaJ protein (9, 10).

The interaction of Dcp1p with Ssa1p/2p invites speculation as to the involvement of other heat shock proteins in the decapping process (10, 48). Intriguingly, Sis1p, a DnaJ homologue, is required for translation initiation and hence could provide the long-suspected link between translation initiation and mRNA decapping (48). Sis1p has been genetically linked to the poly(A) binding protein (PABP) (48). PABP has been previously shown to be an inhibitor of decapping (5). Thus, there appears to be a “triangle” involving PABP, decapping, and translation initiation.

Intriguingly, Hsp70 has recently been implicated in the decay of AU-rich element-containing mRNAs in a mammalian system (27). Laroia et al. (27) demonstrated that AU-rich element-mediated decay is blocked by heat shock, perhaps due to nuclear sequestration of these mRNAs. It will be interesting to determine whether mutations in Ssa1p/2p affect mRNA turnover in yeast.

The stability of mRNAs can be regulated by modulating the activity of the decapping and exonuclease enzymes. The results presented here indicate that the decapping activity of Dcp1p can be regulated. Modulators include Vps16p, Mrt1p, and perhaps the heat shock protein Ssa1p/2p. Furthermore, the results suggest that Vps16p and Mrt1p may modulate decapping activity by altering the interaction of Ssa1p/2p with Dcp1p. We have demonstrated that Flag-Vps16p binds Ssa1p/2p (Fig. 5A, lane 5), and one hypothesis is that it normally prevents or alters the interaction between Ssa1p/2p and Dcp1p.

The results described here suggest that cellular decapping activity is regulated by multiple modulators that appear to regulate an inhibitor of this activity. Future experiments will aim to identify the inhibitor of the decapping enzyme as well as to understand how the protein chaperone Ssa1p/2p regulates decapping. In addition, further investigations are required to establish whether there is a link between the PABP, the heat shock proteins, translation initiation, and decapping. Such experiments will test the intriguing hypothesis that the decision to no longer initiate translation may be directly coupled to the decision to promote the rapid degradation of mRNA by stimulating decapping.