Report

Requirement of the DEAD-Box Protein Ded1p for Messenger RNA Translation

See allHide authors and affiliations

Science  07 Mar 1997:
Vol. 275, Issue 5305, pp. 1468-1471
DOI: 10.1126/science.275.5305.1468

Abstract

The DED1 gene, which encodes a putative RNA helicase, has been implicated in nuclear pre-messenger RNA splicing in the yeast Saccharomyces cerevisiae. It is shown here by genetic and biochemical analysis that translation, rather than splicing, is severely impaired in two newly isolated ded1 conditional mutants. Preliminary evidence suggests that the protein Ded1p may be required for the initiation step of translation, as is the distinct DEAD-box protein, eukaryotic initiation factor 4A (eIF4A). The DED1 gene could be functionally replaced by a mouse homolog, PL10, which suggests that the function of Ded1p in translation is evolutionarily conserved.

Eukaryotic translation initiation requires many factors to promote the binding of the 80S ribosome to the initiation codon in the mRNA (1, 2). Binding of the 43S preinitiation complex, which is derived from the 40S subunit of the 80S ribosome, to mRNA is a rate-limiting step that requires three eukaryotic initiation factors: 4A (eIF4A), 4B, and 4F (consisting of three subunits, eIF4A, eIF4E, and eIF4G). It has been proposed (3) that eIF4B and eIF4F together form an RNA-helicase complex that binds to the 5′ cap of the mRNA through eIF4E. This complex then unwinds duplex structures in the 5′ untranslated region, thereby permitting the 43S complex to scan the mRNA until the first AUG codon is selected. eIF4A is thought to play a major role in this unwinding process, because it can unwind short RNA duplexes in vitro in conjunction with eIF4B (4). eIF4A belongs to the evolutionarily conserved DEAD-box protein family (5), whose members share nine highly conserved amino acid regions, including the distinct Asp-Glu-Ala-Asp (DEAD) sequence. Here, we report that translation in yeast requires another DEAD-box protein, Ded1p, whose function appears to be conserved in evolution.

The DED1 gene was originally identified as an essential open reading frame adjacent to HIS3 (6). Ded1p was hypothesized to function in nuclear pre-mRNA splicing because spp81-1, a mutant allele of DED1, suppresses the growth and splicing defects caused by the prp8-1 mutation (7). More recently, it was reported that overexpression of DED1 can suppress the growth defect of an RNA polymerase III (Pol III) mutant, which suggests that Ded1p can also influence Pol III transcription, although it may not normally participate in this process (8). To investigate the function of Ded1p, we isolated two ded1 cold-sensitive mutants (9). The ded1-120 and ded1-199 alleles yield mutant forms of Ded1p with predicted amino acid substitutions of Gly108 → Asp and Gly494 → Asp (ded1-120) and Gly368 → Asp (ded1-199). At 25°C both mutants grew substantially slower than did the wild-type strain, and at 15°C they did not grow.

We first examined the ded1-120 and ded1-199 mutants for splicing defects by Northern (RNA) blotting. No splicing defects were detected in either ded1 mutant after shifting cultures to 15°C for 2 hours. This observation was in sharp contrast to the aberrant rise in ACT1 pre-mRNA levels and decline in mature CRY1 mRNA levels in control strains prp2 and prp11, which harbor splicing mutations (10).

Instead, we found that, at 15°C, the incorporation of [35S]methionine into acid-precipitable peptides in each ded1 mutant strain was ∼10% of that incorporated in the isogenic wild-type strain. This inhibition of protein synthesis seemed to be general, because the production of most, if not all, of the polypeptides appeared to be equally reduced in the ded1 mutants (Fig. 1A). When the ded1-199 mutant strain was shifted to 15°C for 2 hours, there was a marked increase in the level of 80S ribosomes and a decline in the level of polyribosomes (Fig. 1, B and C). The changes in the polyribosome profile were readily detected 5 min after temperature shift (Fig. 1D), which suggested that Ded1p likely plays a direct role in translation initiation. Fractionation of the ded1 mutant extracts in high-salt sucrose gradients resulted in a complete dissociation of the 80S ribosomes into 40S and 60S subunits, indicating that the accumulated 80S ribosomes were either pre-initiation complexes or 80S couples (10, 11). These phenotypes were not the result of mRNA export defects, because in situ mRNA localization assays (12) revealed no nuclear accumulation of polyadenylated [poly(A)+] RNAs in ded1 mutants after temperature shift for up to 22 hours. Nor were these phenotypes caused by rapid degradation of the poly(A)+ RNAs, because the steady-state levels of transcripts from ACT1, CYH2, CUP1, and CRY1 genes in ded1 mutants were similar to those of the wild-type cells 4 hours after temperature shift (10).

Fig. 1.

Defective protein synthesis in ded1-120 and ded1-199 mutants. (A) Protein synthesis as measured by [35S]methionine incorporation. Lanes 1 and 4, wild-type strain; lanes 2 and 5, ded1-120 strain; lanes 3 and 6, ded1-199 strain. Three OD600 units of cells were resuspended in 1 ml of YPD (1% yeast extract, 2% peptone, and 2% dextrose) and incubated at 30° or 15°C. After 5 min, 20 μCi of L-[35S]methionine (>1000 Ci/mmol, DuPont-NEN) was added and the cells were again incubated at 30° or 15°C for 1 hour. The collected cells were broken by vortexing with glass beads. Proteins in the supernatant were analyzed by electrophoresis on an 8% SDS polyacrylamide gel and autoradiography. (B and C) Polyribosome profiles of the wild-type (B) or ded1-199 (C) strains after temperature shift to 15°C for 2 hours. (D) Similar to (C), except that the cultures were shifted to 15°C for 0, 5, 10, and 15 min. Cells were grown initially to 0.5 to 0.8 OD600 unit at 30°C and shifted to 15°C for the time indicated. Polyribosomes were analyzed on sucrose density gradients (7 to 47%) as in (26).

To test whether Ded1p is present in the cytoplasm, we constructed a yeast strain in which the wild-type DED1 gene was replaced with a DED1-Protein A gene fusion. The fusion protein was functional in vivo, because the growth rate and the polyribosome profile of its host strain (DED1-PA) were nearly indistinguishable from those of the wild-type strain. A Ded1p-Protein A fusion protein (Ded1p-PA) of the predicted size (92 kD) was detected by normal rabbit serum in extracts prepared from the DED1-PA strains, but not from the wild-type strain (Fig. 2A). Indirect immunofluorescence microscopy revealed that Ded1p-PA staining was restricted to the cytoplasm (Fig. 2, C and D). Control experiments with strains expressing only the wild-type Ded1p yielded no fluorescence signal, and an identical cytoplasmic staining pattern was observed in a strain expressing a hemagglutinin-tagged Ded1p (10). Although we cannot rule out the possibility that a minor fraction of Ded1p is present in the nucleus or that Ded1p shuttles between the cytoplasm and the nucleus, our results strongly suggest that Ded1p is predominantly cytoplasmic, consistent with its proposed role in translation.

Fig. 2.

Cytoplasmic localization of Ded1p-PA. (A) Immunoblot analysis of crude extracts from the DED1-PA strain (lanes 1 and 2, two independent isolates) or from the wild-type strain (lane 3). Immunoblots were developed with normal rabbit serum at 1:2500 dilution and Protein G-horseradish peroxidase conjugate (Bio-Rad) at 1:6000 dilution, followed by chemiluminescence detection (ECL system, Amersham) of Ded1p-PA. The DED1-PA recombinant clone was constructed by fusing in frame a 651-bp polymerase chain reaction product containing the IgG-binding domain of Protein A to the Aat II site immediately upstream of the stop codon of the wild-type DED1 gene carried on pRS315. The Aat II site was engineered by site-specific mutagenesis using an oligonucleotide primer (5′-TGTCTGAAATCAGACGTCCCACCAAGAAGA-3′). This DED1-PA plasmid was introduced into a diploid strain [MATaded1::TRP1/ ded1::TRP1 ura3-52/ura3-52 lys2-801/lys2-801 ade2-101/ade2-101 trp1-Δ1/trp1-Δ1 his3-Δ200/his3-Δ200 leu2-Δ1/leu2-Δ1 pDED1008 (DED1/CEN/URA3)]. After 5-FOA counterselection (9) of pDED1008, the resulting diploid strain was used for immunoblot analysis and indirect immunofluorescence microscopy (27). (B to D) Indirect immunofluorescence microscopy of DED1-PA cells. Cells viewed by Normarski optics (B), cells stained first with purified normal rabbit IgG and then with Texas Red-conjugated goat antibody to rabbit IgG (C), and cells stained by DAPI (D) are shown. Scale bar, 10 μm.

Independent mutations in the same biochemical pathway may act synergistically to yield a lethal phenotype (13). We tested the possibility that the ded1-120 and ded1-199 mutations are synthetically lethal to a temperature-sensitive mutation in TIF1 (14), which encodes yeast eIF4A (15). Strains harboring the double mutations ded1 tif1, ded1 prp28, ded1 prt1, and ded1 sis1 were constructed and examined under growth conditions permissive to all the single mutants. Prp28p is a member of the DEAD-box protein family involved in pre-mRNA splicing (16), Prt1p is an eIF3 subunit (17), and Sis1p is thought to mediate the dissociation of protein complexes in the translation machinery (11). Lethality resulted only when the ded1-120 (10) or ded1-199 mutations were combined with the tif1 mutation (Fig. 3). Thus, the synthetic lethality was gene-specific and was not simply dependent on the presence of two mutant DEAD-box protein genes. Because DED1 and TIF1 are both essential genes and the observed synthetic lethality is not allele-specific, the simplest interpretation for these data is that Tif1p and Ded1p play two independent but nonetheless related roles in translation. The inability to suppress the growth defects of ded1 mutants by overexpression of TIF1 (10) is consistent with this proposal. Our genetic analysis is further supported by the finding that a ded1 mutation is also synthetically lethal to cdc33-1, which encodes a mutant eIF4E (18).

Fig. 3.

Synthetic lethality of the ded1-199 and tif1 mutations. Single and double mutants were streaked out on a 5-FOA plate and incubated at 30°C for 4 days. Double mutants were constructed by first crossing a tif1 temperature-sensitive strain (strain SS13-3A; MATa tif1::HIS3 tif2::ADE2 his3 ade2 leu2 trp1 ura3 YCpLac33- tif1ts) to a DED1 wild-type strain [MATα ded1::TRP1 ura3-52 lys2-801 ade2-101 trp1-Δ1 his3-Δ200 leu2-Δ1 pDED1009 (DED1/CEN/LEU2)] to obtain a combination of appropriate genetic markers. The tights allele on YCpLac33 was transferred to pRS317 to allow 5-FOA counterselection of cells harboring a URA3 plasmid containing the DED1 gene (pDED1008). The tester strains at the end are MATa (or α) tif1::HIS3 tif2::ADE2 ded1::TRP1 his3 ade2 leu2 trp1 ura3 lys2 pRS317-tif1ts pDED1008. To rule out synthetic lethality caused by genetic background variations, we used at least three independent tester strain isolates. To test for synthetic lethality, we individually transformed DED1, ded1-120, and ded1-199 alleles carried on pRS315 into tester strains. The resulting transformants were then streaked out on 5-FOA plates and incubated at 27° or 30°C for 4 days. The ded1 prt1, ded1 sis1, and ded1 prp28 double mutants were constructed similarly, using strain F294 (MATa prt1-1 ade1 leu2-3,112 ura3-52), strain CY732 (11) [MATa sis1::HIS3 ade2-1 trp1-1 leu2-3,112 his3-11,15 ssd1-d2 can1-100 (sis1-85 on a CEN/LEU2 plasmid)], and strain YTC95 (28) [MATα prp28::HIS3 ura3-52 lys2-801 ade2-101 trp1-Δ1 his3-Δ200 leu2-Δ1 (prp28-117 on a CEN/LEU2 plasmid)]. The prp28-117 mutant does not grow at 15°C.

A yeast in vitro translation system (19) was used to test whether Ded1p is directly involved in translation. Immunodepletion of Ded1p-PA by immunoglobulin G (IgG)-Sepharose beads nearly completely abolished translation activity as measured by a luciferase assay (Fig. 4A). This effect was apparently caused by the depletion of >90% of Ded1p-PA in the extract as judged by immunoblot analysis (10). Identical treatments of the wild-type extracts reduced the translation activity by only 40%, presumably as a result of nonspecific interactions of IgG-Sepharose beads with cellular components. Because the chemical half-lives of the luciferase transcripts in both cases were nearly identical (10), the loss of the translation activity was unlikely to have been a result of the rapid degradation of transcript upon Ded1p-PA depletion. Addition of normal rabbit IgG to the Ded1p-PA extracts yielded a dose-dependent reduction of the translation activity but had no inhibitory effect on the wild-type extracts (Fig. 4B). To rule out nonspecific depletion, we used a glutathione-S-transferase (GST)-Ded1p fusion protein purified from Escherichia coli to reconstitute the translation activity. GST-Ded1p enhanced the translation activity in a dose-dependent manner to more than eight times that of the depleted extracts, whereas purified GST had no stimulatory effect on the translation activity (Fig. 4C).

Fig. 4.

Loss and restoration of Ded1p-dependent translation activity. (A) Immunodepletion of Ded1p-PA results in loss of translation activity. Data are expressed as the percent of luciferase produced in extracts treated with IgG-Sepharose (solid bars) and without treatment (open bars; 100% activity). Yeast extract (25 μl) was mixed with IgG-Sepharose [10 μl (bed volume), Pharmacia] and incubated at 4°C for 30 min. After centrifugation, 10 μl of supernatant was used to assemble a 25-μl translation reaction in the presence of 100 ng of capped-luciferase transcript with poly(A) tail (19). Reactions were incubated at 20°C for 1 hour. Luciferase was quantitated by luminescence emission using a luciferase assay (Promega). Experiments were repeated at least three times. (B) Neutralization of Ded1p-PA results in loss of translation activity. Data are expressed as the percent of luciferase produced in the absence of normal rabbit IgG in the wild-type (□) and DED1-PA (▵) extracts. After 10 μl of yeast extract was preincubated with increasing amounts of normal rabbit IgG at 4°C for 30 min, translation was assayed as in (A). (C) Restoration of Ded1p-dependent translation activity by addition of GST-Ded1p. Data are expressed as activity increase relative to the translation activity of depleted extract upon addition of GST-Ded1p (open bars) or GST (solid bars). The DED1 coding region was fused in frame to the GST coding region on pGEX-2T (Pharmacia) and overexpressed in E. coli strain XL1-Blue (Stratagene). Affinity purification of GST-Ded1p and GST by glutathione agarose (Sigma) was done according to manufacturer's instructions (Pharmacia). After 10 μl of the depleted extract was preincubated with increasing amounts of purified GST-Ded1p or GST for 15 min at 20°C, translation was assayed as in (A).

Ded1p shares 53% amino acid sequence identity with the predicted gene product of a mouse transcript, PL10, which is thought to be involved in spermatogenesis (20). We thus tested whether PL10 and DED1 are functional homologs. When driven by the yeast glyceraldehyde-3-phosphate dehydrogenase promoter (PGPD) and carried on a yeast centromere plasmid, the PL10 cDNA rescued the lethality of cells with a chromosomal ded1 deletion (10). This functional complementation was specific, because a PGPD-PRP28 construct capable of complementing a lethal deletion of the PRP28 gene failed to do so in the ded1-deletion background (10). This observation suggests that PL10 may play a role in translation in the mouse.

It is unlikely that Ded1p plays a direct role in pre-mRNA splicing (Fig. 2C) (10) and Pol III transcription (8). We suggest that Ded1p may influence the production of factors involved in these two nuclear events, thereby mediating the genetic suppression indirectly. Although several novel factors, including a helicase-like protein, have been implicated in translation (21), subsequent biochemical analysis suggested otherwise (22). Our data thus provide evidence that at least two distinct DEAD-box proteins are indispensable for translation; such a situation is reminiscent of pre- mRNA splicing (23) and ribosomal biogenesis (24), in which multiple DEAD-box proteins are required. Although it is clear that eIF4A and Ded1p are functionally distinct, it remains possible that their functions partially overlap. Our results also raise the possibility that the involvement of PL10 during the meiotic and haploid stages of mouse spermatogenesis is mediated by its function in translation. Because multiple PL10- and DED1-like genes are expressed in different mouse tissues (25), these genes are functionally important and have potential roles in regulating gene expression by means of differential translation.

REFERENCES AND NOTES

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
View Abstract

Stay Connected to Science

Navigate This Article