The DEAD-Box RNA Helicase Dbp5 Functions in Translation Termination

See allHide authors and affiliations

Science  02 Feb 2007:
Vol. 315, Issue 5812, pp. 646-649
DOI: 10.1126/science.1134641


In eukaryotes, termination of messenger RNA (mRNA) translation is mediated by the release factors eRF1 and eRF3. Using Saccharomyces cerevisiae as a model organism, we have identified a member of the DEAD-box protein (DBP) family, the DEAD-box RNA helicase and mRNA export factor Dbp5, as a player in translation termination. Dbp5 interacts genetically with both release factors and the polyadenlyate-binding protein Pab1. A physical interaction was specifically detected with eRF1. Moreover, we show that the helicase activity of Dbp5 is required for efficient stop-codon recognition, and intact Dbp5 is essential for recruitment of eRF3 into termination complexes. Therefore, Dbp5 controls the eRF3-eRF1 interaction and thus eRF3-mediated downstream events.

DEAD-box RNA helicases are found in almost all organisms and function in many fundamental steps in the life of RNA molecules, ranging from transcription to decay. They use the energy from adenosine triphosphate hydrolysis to rearrange RNA structures or to dissociate RNA/protein complexes (1). The DEAD-box helicase Dbp5 shuttles between the nucleus and the cytoplasm and is involved in translocation of the mature mRNA/protein complex into the cytoplasm (2). Consistently, Dbp5 is localized to the nuclear rim, where it interacts with components of the nuclear pore complex (NPC) (2, 3). Dbp5 is also distributed within the cytoplasm of yeast, higher eukaryotes, and human cells, but a cytoplasmic function has not yet been defined (46).

To get insights into its cytoplasmic function, we used Saccharomyces cerevisiae to investigate whether Dbp5 associates with mRNAs during translation. Western blot analysis of sucrose density fractionation experiments revealed that, like the polyadenylate-binding protein Pab1, significant amounts of the extracted Dbp5 (∼60%, which equals ∼40% of the total Dbp5) but not of Gfd1 (a nuclear pore–associated factor) were detectable in polysome-containing fractions (Fig. 1, A and B). The addition of puromycin, which specifically disrupts polysomes, confirmed that Dbp5 co-sediments with polyribosome-containing mRNAs (fig. S1). A potential role of Dbp5 in translation is supported by the finding that dbp5/rat8 mutants are hypersensitive to translational inhibitors (Fig. 1C). This effect is not due to defects in mRNA export, because rat7-1, which in contrast to dbp5 displays strong mRNA export defects even at 25°C (Fig. 1D), is not hypersensitive to those inhibitors.

Fig. 1.

Dbp5 associates with polysomes during translation, and mutants are hypersensitive to translational inhibitors. (A) Supernatants of cell extracts of log-phase wild-type cells, expressing functionally green fluorescent protein (GFP)–tagged versions of GFD1, DBP5, or PAB1, were fractionated through 15 to 50% sucrose gradients and subjected to Western blot analysis. Antibodies to Rps3 detected endogenous ribosomal protein. Absorbance at 254 nm (OD, optical density) shows the distribution of ribosomes. The ratio of the extracted proteins is shown as follows: amount of the unbound proteins (left) or mono- and polysome-bound proteins (right). (B) Quantification of at least three independent experiments shown in (A). (C) Growth of wild-type (WT), rat7-1, rat8-2, and rat8-3 strains on yeast extract, peptone, and dextrose (YPD) medium with and without translational inhibitors in 10-fold serial dilutions of similar cell numbers. (D) mRNA localization in log-phase wild-type, rat7-1, and rat8 mutants at 25°C.

Different genetic approaches were carried out to investigate in which phase of translation Dbp5 might be acting. Dbp5 overexpression studies revealed that mutations in genes encoding initiation or elongation factors were not suppressed by high-copy DBP5, whereas mutations in both of the eukaryotic release factors eRF1 (SUP45) and eRF3 (SUP35), as well as a mutation in Pab1, were specifically suppressed (fig. S2). Moreover, similarly to the synthetic lethal effect seen when both release-factor mutants (sup35-21 and sup45-2) are combined, dbp5/rat8 mutants are synthetically lethal with mutant eRF1, and the growth of dbp5/rat8 mutants is severely inhibited when combined with mutant eRF3 or Pab1 (pab1-53) (Fig. 2A), indicating a potential function of Dbp5 in translation termination.

Fig. 2.

Genetic and physical interaction of Dbp5/Rat8 with translation termination factors. (A) Growth of all indicated double mutants carrying a 2μ URA3 vector encoding either DBP5 or SUP35 SUP45 and all single mutants carrying an empty vector was analyzed on –URA- and 5′-fluoroorotic acid (FOA)–containing plates that select for the loss of plasmid. (B) Immunoprecipitation of functionally tagged versions [or the endogenous release factors (see fig. S3)] of eRF1 (Sup45-GFP), eRF3 (Sup35-GFP), and Pab1-GFP with Myc-tagged Dbp5. Total lysates were split and incubated with or without RNase, and Dbp5-bound proteins were detected on Western blots with an antibody to GFP. Endogenous Por1 served as a negative control.

Translation termination is mediated by recognition of a stop codon via the transfer RNA (tRNA)–mimicking protein eRF1 and subsequent hydrolysis of the ester bond connecting the polypeptide chain and the tRNA, stimulated by the guanosine triphosphatase activity of eRF3. Although eRF3 is unable to promote in vitro termination on its own, it enhances eRF1 activity (7). Direct interactions between Pab1 and eRF3 have been described in yeast, frog, and mammalian cells (7, 8), and overexpression of PAB1 in yeast suppresses effects associated with mutant eRF3 in vivo, suggesting a functionally important interaction of these proteins for termination (9), possibly as bridging factors to channel the terminating ribosomes back to the 5′ end of an mRNA (8). To get insights into how Dbp5 might contact the termination machinery, we performed co-immunoprecipitation experiments that revealed a stable interaction between Dbp5 and eRF1 (Fig. 2B and fig. S3B). In contrast, no interaction between Dbp5 and eRF3 was found, whereas an interaction between Pab1 and Dbp5 was ribonuclease (RNase)–sensitive, indicating concurrent binding of Dbp5 and Pab1 to the same mRNA but no simultaneous presence of Dbp5 and eRF3.

To demonstrate an active role of Dbp5 in translation termination, we assayed the stop codon recognition under different conditions. First, we showed that like eRF1 and eRF3 mutants, dbp5/rat8 mutant cells show increased termination readthrough activity in luciferase dual-reporter assays (Fig. 3). The assay is based on compared expression of β-galactosidase and luciferase open reading frames, separated by a stem loop or a stop codon (fig. S4A), which allows us to compare the frequency of translational readthrough in different strain backgrounds (10). In agreement with previous results, we found a basal readthrough activity of ∼15 to 17% in wild-type cells (10). Although mutant dbp5 did not influence readthrough activities in the presence of the stem loop, Dbp5 was clearly required for efficient recognition of the termination codon, in contrast to Dhh1 (Fig. 3A). Dhh1 is another DEAD-box RNA helicase family member, which has been implicated in connecting translation to mRNA degradation, because it can act as a translational repressor and an activator of mRNA decapping (11). However, in contrast to dbp5 (Fig. 2A), no genetic interaction between dhh1Δ and any of the termination-factor mutants was detected (fig. S4C). The rate of read-through in dbp5 mutant cells was very similar to that seen with a mutation in eRF3 (Fig. 3B). The termination readthrough defects in eRF1 and eRF3 mutants were fully suppressed in the presence of high-copy Dbp5 but not Dhh1 (Fig. 3, B and C). This was confirmed in a different assay, in which the decreased translational fidelity of either defective release factor, reflected by decreased growth rates in the presence of paromomycin, was efficiently suppressed by high-copy DBP5 (fig. S4B). Moreover, catalytic RNA helicase activity of Dbp5 is required for this suppression, because a mutation in its DEAD-box (Glu240→Gln240) prevented suppression (Fig. 3C).

Fig. 3.

Dbp5 is required for efficient stop codon recognition. (A) Readthrough activities in wild-type, rat8-2, and dhh1Δ are shown. All strains carrying either of the reporter constructs were grown to log phase and shifted to 37°C for 15 min before cell lysis. β-galactosidase and luciferase activities were measured and their ratios were used to calculate the relative molar luciferase expression. (B) Suppression of the termination readthrough defects of mutations in eRF1 and eRF3 by increased (PGAL1DBP5) Dbp5 level. All strains carrying the reporter construct and a DBP5 plasmid were treated as described in (A). The readthrough activity of all strains is shown. (C) Catalysis activity of Dbp5 is required for sup45-2 suppression. The experiment was done as described in (B), with active Dbp5 (PGAL1DBP5) versus an inactive DEAD-box mutant (PGAL1dbp5 E240Q) and Dhh1 (PGAL1DHH1). The results of at least 10 independent experiments are shown.

Besides functioning as an eRF1-stimulating protein, eRF3 is a key mediator that transmits the termination signal to mRNA decay (7). The RNA helicase Upf1 (human RENT1 or HUPF1) is involved in the nonsense-mediated decay (NMD) of aberrant mRNAs, and deletion of UPF1 leads to an increased termination readthrough (12). Although Upf1 and its interacting proteins Upf2 and Upf3 are capable of binding to eRF3, it is hypothesized that Pab1 competes for this interaction and thereby precludes the binding and destabilizing activities of these NMD factors at normal terminators, supporting a role for Upf1 in coupling premature termination events to NMD rather than a function in regular translation (12). To investigate a potential involvement of Dbp5 in NMD, we tested whether mutations in DBP5 stabilize NMD substrates and found that this is not the case, in contrast to upf1Δ (fig. S5A). Also, neither genetic interactions between dbp5 mutants and upf1, upf2, or upf3 deletion strains, nor genetic interactions between upf1Δ and either eRF1 or eRF3 mutants, were detectable (fig. S5, B and C), in contrast to the synthetic growth defects detected between dbp5/rat8 and either release-factor mutant (Fig. 2A). Further, the readthrough defects of dbp5/rat8 and either upf mutant are not additive (fig. S5D). Together, these findings suggest a general function for Dbp5 in translation termination and support an exclusive function for Upf1 on NMD substrates.

To gain mechanistic insights into the function of Dbp5 during translation termination, we compared polysomes of rat8-2 to those of the wild-type and found that rat8-2 is defective in the recruitment of eRF3 into termination complexes (Fig. 4A). In a wild-type strain, roughly 60% of extracted Dbp5, eRF1, and eRF3 is ribosome-associated. In contrast, eRF3 is almost absent from the ribosomal fractions (12%) in rat8-2 cells at the nonpermissive temperature, whereas approximately 55% of both Dbp5 and eRF1 remained polysome-associated, indicating that intact Dbp5 is required for eRF3 entry into the termination complexes. The corresponding rRNA profiles revealed an elevated monosome peak in rat8-2 cells (resulting in three monosome fractions in the Western blot) as compared to the wild-type (two fractions) or rat7-1, reflecting the influence of Dbp5 on translation (fig. S6). The defect of the dbp5/rat8-mutant in eRF3 recruitment was further confirmed by analysis of the interaction between eRF1 and eRF3 (13), which was completely lost in dbp5/rat8 mutant cells, in contrast to wild-type or upf1Δ strains (Fig. 4B).

Fig. 4.

Active Dbp5 is required for the incorporation of eRF3 into translation termination complexes. (A) Sucrose fractionation experiments (with 15 to 50% sucrose gradients) of log-phase wild-type and rat8-2 strains shifted to 37°C for 20 min were performed, and Western blot detection of TAP-tagged SUP35, SUP45, and GFP-DBP5 or GFP-rat8-2, respectively, and Rps3 is shown in wild-type (upper panel) and rat8-2 (lower panel) cells. The ratio of the extracted proteins is shown as follows: amount of the unbound proteins (left) or mono- and polysome-bound proteins (right). (B) Top: Sup35-TAP purification from log-phase wild-type and rat8-2 cells expressing SUP45-GFP shifted to 37°C for 20 min prior to lysis is shown as total lysates and purified proteins (bound) after incubation with protein A for 3 hours with or without RNase. Sup45-GFP and Por1 (negative control) were detected with antibodies to GFP and Por1, respectively. Bottom: Immunoprecipitations of endogenous eRF1 from wild-type, upf1Δ, rat8-2, and rat8-3 cells treated as described above are shown. The presence of endogenous eRF3 and Fech1 (negative control) is shown.

Together our data reveal a requirement for functional Dbp5 for the entry of eRF3 into termination complexes and support the following model. Once the ribosome has reached a termination codon, eRF1 is recruited to the mRNA, possibly by Dbp5. The RNA helicase activity of Dbp5 might remodel the mRNA/protein complex to allow proper eRF1 positioning on the stop codon and thus an efficient termination reaction. Subsequent dissociation of Dbp5 from eRF1 is followed by the entry of eRF3 into the complex, which promotes the release of the peptide chain and allows eRF3-mediated downstream events. Thus, Dbp5 is involved in translation termination through its interaction with eRF1, and it controls the subsequent eRF1-eRF3 interaction through its dissociation from eRF1.

Supporting Online Material

Materials and Methods

Figs. S1 to S6

Tables S1 and S2


References and Notes

View Abstract

Navigate This Article