Cap-binding protein 1-mediated and eukaryotic translation initiation factor 4E-mediated pioneer rounds of translation in yeast

Abstract

Nonsense-mediated mRNA decay (NMD) in mammalian cells is restricted to newly synthesized mRNA that is bound at the 5′ cap by the major nuclear cap-binding complex and at splicing-generated exon–exon junctions by exon junction complexes. This messenger ribonucleoprotein has been called the pioneer translation initiation complex and, accordingly, NMD occurs as a consequence of nonsense codon recognition during a pioneer round of translation. Here, we characterize the nature of messenger ribonucleoprotein that is targeted for NMD in Saccharomyces cerevisiae. Data indicate that NMD targets both cap-binding complex (Cbc)1p- and eukaryotic translation initiation factor (eIF)4E-bound mRNAs, unlike in mammalian cells, where NMD does not detectably target eIF4E-bound mRNA. First, intron-containing pre-mRNAs in yeast are detectably bound by either Cbc1p, or, unlike in mammalian cells, eIF4E, indicating that mRNAs can be derived from either Cbc1p- or eIF4E-bound pre-mRNAs. Second, the ratio of nonsense-containing Cbc1p-bound mRNA to nonsense-free Cbc1p-bound mRNA, which was < 0.4 for those mRNAs tested here, is essentially identical to the ratio of the corresponding nonsense-containing eIF4E-bound mRNA to nonsense-free eIF4E-bound mRNA, and both ratios increase in cells treated with the translational inhibitor cycloheximide (CHX). These data, together with data presented here and elsewhere showing that Cbc1p-bound transcripts are precursors to eIF4E-bound transcripts, demonstrate that Cbc1p-bound mRNA is targeted for NMD. In support of the idea that eIF4E-bound mRNA is also targeted for NMD, eIF4E-bound mRNA is targeted for NMD in strains that lack Cbc1p. These results suggest that both Cbc1p- and eIF4E-mediated pioneer rounds of translation occur in yeast.

Studies of mammalian cells indicate that nonsense-mediated mRNA decay (NMD) targets newly synthesized mRNA as a consequence of nonsense codon recognition during what has been called a pioneer round of translation (1). For spliced mRNAs, the pioneer translation initiation complex is characterized by (i) the mostly nuclear but shuttling cap-binding heterodimer that consists of cap-binding protein (CBP)80 and CBP20; (ii) the exon junction complex of proteins that is deposited as a consequence of pre-mRNA splicing ≈ 20–24 nt upstream of exon–exon junctions, and includes NMD factors up-frameshift (Upf)2 and Upf3 or Upf3X (also called Upf3a and Upf3b, respectively); and (iii) poly(A)-binding protein (PABP) N1 [previously designated PABP2 (1, 2)] and PABPC [previously designated PABP (ref. 2 and F. Lejeune and L.E.M., unpublished data)]. By the time eukaryotic initiation factor (eIF) 4E, which is the major cytoplasmic cap-binding protein, replaces the CBP80/20 heterodimer at the mRNA cap, the exon–junction complex, and associated Upf NMD factors have been removed so that the resulting steady-state mRNA is immune to NMD (1–4). Thus, whereas the pioneer translation initiation complex has components in common to the steady-state translation initiation complex, it is structurally and functionally distinct (1, 2, 4, 5).

To date, comparable studies of Saccharomyces cerevisiae that examine the nature of messenger ribonucleoprotein (mRNP) that is targeted for NMD have not been performed. In this study, we demonstrate that NMD in S. cerevisiae targets mRNAs that are bound by the primarily nuclear cap-binding complex (Cbc)1p, which is orthologous to mammalian CBP80, and mRNAs that are bound by the primarily cytoplasmic eIF4E. Because the ratio of intron-containing pre-mRNA to the corresponding spliced mRNA product is similar before and after immunopurification (IP) with anti-eIF4E, but at least 4-fold larger after IP with anti-Cbc1p, we conclude that eIF4E replaces Cbc1p at the caps of most transcripts but, alternatively, may bind to the caps of some transcripts without prior Cbc1p binding. This replacement of Cbc1p by eIF4E apparently occurs despite the fact that Cbc1p facilitates pre-mRNA splicing (6) and, possibly, mRNA export (7, 8). Consistent with the ability of eIF4E to productively bind to newly synthesized caps without prior Cbc1p binding, and in keeping with previous data indicating that cells lacking Cbc1p are viable and able to support NMD (8–11), we find that NMD targets eIF4E-bound mRNA in a cbc1-Δ strain. These data and other data indicate that mRNAs in yeast, like mRNAs in mammalian cells, undergo a Cbc1p-mediated pioneer round of translation that is inhibited by CHX, and that nonsense codon recognition during this round of translation leads to NMD, provided the Upf NMD factors are present. Additionally, these data indicate that mRNAs in yeast also undergo NMD when bound by eIF4E, unlike mRNAs in mammalian cells.

Materials and Methods

Yeast Strains, Media, and Yeast Genetics. The genotypes of S. cerevisiae strains used in this study are listed in Table 1. Standard yeast extract/peptone/dextrose and omission media were used for yeast propagation and testing (12). The UPF3, UPF1, and UPF2 genes were sequentially disrupted in the strain B-10529 by using PCR-generated DNA fragments that harbored the appropriate gene fragment and the selective blaster hisG-URA3-hisG (13), thus generating B-15334 (Table 1).

Table 1. Yeast strains.

Strain no.	Genotype	Source
B-10529	MATα cyc1-512 lys2-187 leu2-1 ura3-52 his4-166 met8-1 (Normal A)	—*
B-15315	MATα cyc1-512 lys2-187 leu2-1 ura3-52 his4-166 met8-1 upf3::hisG URA3 hisG	B-10529
B-15316	MATα cyc1-512 lys2-187 leu2-1 ura3-52 his4-166 me 8-1 upf3::hisG	B-15315
B-15317	MATα cyc1-512 lys2-187 leu2-1 ura3-52 his4-166 met8-1 upf3::hisG upf1::hisG URA3 hisG	B-15316
B-15318	MATα cyc1-512 lys2-187 leu2-1 ura3-52 his4-166 met8-1 upf3::hisG upf1::hisG	B-15317
B-15333	MATα cyc1-512 lys2-187 leu2-1 ura3-52 his4-166 met8-1 upf3::hisG upf1::hisG upf2::hisG URA3 hisG	B-15318
B-15334	MATα cyc1-512 lys2-187 leu2-1 ura3-52 his4-166 met8-1 upf3::hisG upf1::hisG upf2::hisG	B-15333
B-11621	MATa cyc1-512 lys2-187 his4-166	—†
B-11622	MATa cyc1-512 lys2-187 his4-166 leu2-1 met8-1	—†
B-9037	MATa cyc1-512 trp2-1 ura3-52 (Normal B)	Ref. 11
B-11558	MATa cyc1-512 trp2-1 ura3-52 cbc1::hisG	Ref. 11

^*B-10529 is a segregant from the diploid strain D-1135.

^†B-11621 and B-11622 are segregants from the diploid strain B-11768.

Plasmids. The plasmids used in this study are listed in Table 2. The PGK1 gene lacking a premature termination codon (PTC) is denoted as PGK1 WT, even though it is not a WT gene.

Table 2. Plasmids used in this study.

Plasmid no.	Yeast genes	Ref.
pAB2992 [pRIP1PGK(—AU)]	URA3 PGK1 (WT)	14
pAB2996 [pRIPPGKH2(3)Δ4]	URA3 PGK1 (PTC)	14

IPs. Yeast cells were grown in 600 ml of yeast extract/peptone/dextrose or SC-Ura medium to an OD₆₀₀ of 1.5 (≈3 × 10⁷ cells per ml). Cells were pelleted at 5,000 × g for 5 min, washed with water, and resuspended in 2 ml of lysis buffer (50 mM Tris·HCl, pH 7.4/150 mM NaCl/0.05% Tween 20/5 mM MgCl₂/1 mM PMSF/1 mM benzamidine/10 mM NaF/10 mM β-glycerophosphate). Cells were broken by two passages through a French press at 750 psi, and then centrifuged at 10,000 × g for 10 min. For each IP, 1 ml of supernatant (that corresponded to ≈7.5 × 10⁹ cells) was precleared with 50 μl of protein A-conjugated agarose (Roche Diagnostics) by rotation for 30 min at 4°C. IP was then performed as described (1). However, 2 mg of bacterial tRNA (Sigma) was used to saturate the beads before IP, and the beads were washed 10 times before elution.

Western Blotting. Western blotting was performed (1) by using the following antibodies: anti-Cbc1p (15) (1:100,000 dilution) or anti-eIF4E (16) (1:1,000 dilution).

RT-PCR. Immunopurified transcripts or transcripts before IP were analyzed by RT-PCR (17). ACT1 pre-mRNA was amplified by using the primers 5′-GTCTCATGTACTAACATCGATTGC-3′ (sense) and 5′-CCCAGTTGGTGACAATACCGTG-3′ (antisense), and ACT1 mRNA was amplified, using primer 5′-TTTACTGAATTAACAATGGATTCTGAGG-3′ (sense) and the same antisense primer as used for pre-mRNA. TUB3 pre-mRNA was amplified by using the primers 5′-TTTGTGTCTTCT TCT TCGG-3′ (sense) and 5′-ATCGATA ACAT TGGGCTCT-3′ (antisense), and TUB3 mRNA was amplified, using primer 5′-GTCATTAGTATTAATGTTGGT-3′ (sense) and the same antisense primer as used for pre-mRNA. KIN28 pre-mRNA was amplified by using the primers 5′-GTACACAAAGGTAGTGGGGG-3′ (sense) and 5′-CTTAACTTCACGGATAGCTGAC-3′ (antisense), and KIN28 mRNA was amplified, using primer 5′-GAATATGGAGTACACAAAGGAAAAG-3′ (sense) and the same antisense primer as used for pre-mRNA. CYH2 pre-mRNA was amplified by using the primers 5′-GTATCAAATGGTTGTAGAGAGCGC-3′ (sense) and 5′-TGTGGAAGTATCTCATACCAACC-3′ (antisense), and CYH2 mRNA was amplified, using primer 5′-CAGAGGTCACGTCTCAGCC-3′ (sense) and the same antisense primer as used for pre-mRNA. URA3 mRNA was amplified by using the primers 5′-GTGCTTCATTGGATGTTCGTACC-3′ (sense) and 5′-CCACCACACCGTGTGCATTC-3′ (antisense). PGK1 mRNA that was either WT or PTC-containing was amplified by using the primers 5′-GTCTCTACTGGTGGTGGTG-3′ (sense) and 5′-GGGGAAAGAGAAAAGAAAAAAATTGATCTATCGATAGT-3′ (antisense). MET8 mRNA that was either WT or PTC-containing was amplified by using primers 5′-GAGGAATTAGGCTGCTGGCAC-3′ (sense) and 5′-CTGCAAGGAACAGTTCTGTTC-3′ (antisense).

Results

The Ratio of pre-mRNA to Its Product mRNA Is Higher in IPs That Used Anti-Cbc1p Instead of Anti-eIF4E. To gain insight into the precursor-product relationship between Cbc1p- and eIF4E-bound transcripts in S. cerevisiae, we compared the relative abundance of intron-containing pre-mRNA and its product, spliced mRNA, in IPs by using antibody to each cap-binding protein. Strain B-10529 (Normal A) was grown, and extracts were prepared with and without IP by using (i) anti-Cbc1p, (ii) anti-eIF4E, (iii) purified rabbit IgG, which controlled for nonspecific IP using anti-Cbc1p, or (iv) normal rabbit serum (NRS), which controlled for nonspecific IP, using anti-eIF4E. Western blotting demonstrated that anti-Cbc1p specifically immunopurified Cbc1p, but not eIF4E, whereas anti-eIF4E specifically immunopurified eIF4E, but not Cbc1p (Fig. 1A). RT-PCR demonstrated that the ratio of pre-mRNA to mRNA in the anti-Cbc1p IP was 15- to 21-fold, 4- to 7-fold, 5- to 6-fold, and 4- to 5-fold higher than in the anti-eIF4E IP, respectively, for ACT1, TUB3, KIN28, and CYH2 transcripts (Fig. 1B; data not shown for an independently performed experiment). Because the amplification efficiency of each pre-mRNA is likely to differ from that of the corresponding mRNA, it was not possible to determine the relative amount of each pre-mRNA and mRNA that was associated with each CBP. However, we can conclude that some pre-mRNA detectably associates with eIF4E in yeast, which is in contrast to the situation in mammalian cells (4). Whether eIF4E binds pre-mRNA instead of Cbc1p, after replacing Cbc1p, or both, remains to be determined (see below). Furthermore, we can conclude that mRNA preferentially associates with eIF4E because the ratio of pre-mRNA to mRNA is comparable before or after IP by using anti-eIF4E. Notably, the amount of Cbc1p-bound pre-mRNA was too small to detectably change the ratio of pre-mRNA to mRNA before and after IP with anti-eIF4E.

Fig. 1.

The ratio of pre-mRNA to its product mRNA is higher in IPs that used anti-Cbc1p instead of anti-eIF4E. (A) Anti-Cbc1p and anti-eIF4E specifically immunopurify Cbc1p and eIF4E, respectively. B-10529 (Normal A) cells were cultured in yeast extract/peptone/dextrose medium. Lysates were then generated and immunopurified by using anti-Cbc1p or anti-eIF4E. The specificity of each IP was controlled for by using rabbit IgG or NRS, respectively. Samples were analyzed by Western blotting. IP efficiencies were 20% for Cbc1p and 10% for eIF4E. The five leftmost lanes represent twofold dilutions of protein before IP (–) and demonstrate that the conditions used for Western blotting were semiquantitative. (B) Both intron-containing pre-mRNAs and spliced mRNAs immunopurify with Cbc1p and eIF4E, although to differing extents. RT-PCR was performed to detect the specified transcripts. PCR was initiated by using three amplification cycles and primers for the specified pre-mRNA and continued, using 19 cycles after adding primers for the corresponding mRNA, except for both KIN28 pre-mRNA and KIN28 mRNA, which were amplified using 22 cycles. The numbers below each lane designate the ratio of pre-mRNA to mRNA after subtracting the amount present in the appropriate control IP. The five leftmost lanes represent twofold dilutions of RNA before IP (–) and demonstrate that the conditions used for RT-PCR were semiquantitative.

NMD Targets Cbc1p-Bound PGK1 mRNA. Because data indicate that pre-mRNA is bound by either Cbc1p or eIF4E, it may be that Cbc1p is replaced by eIF4E before the first round of translation. If so, then a Cbc1p-initiated pioneer round of translation would not occur in yeast. To address this possibility, we determined whether Cbc1p is replaced by eIF4E before NMD. Strains B-10529 (Normal A) and B-15334 (upf1-Δ upf2-Δ upf3-Δ) were transformed with a centromeric plasmid that carried (i) a URA3-selective marker and (ii) a PGK1 gene that was either WT (pAB2992) or harbored a TAG PTC (pAB2996). Extracts were prepared with and without IP as described above. RT-PCR of samples before IP demonstrated that the level of PGK1 PTC mRNA was 28% the level of PGK1 WT mRNA in the Normal A strain and increased to 95% of PGK1 WT mRNA in the upf1-Δ upf2-Δ upf3-Δ strain (Fig. 2A), which is consistent with previous reports (14). In these analyses and subsequent experiments, the level of URA3 mRNA was used to control for variations in plasmid copy number and RNA recovery. Western blotting revealed that the IPs were specific (Fig. 2B Upper). RT-PCR of samples before and after IP revealed that the level of PGK1 PTC mRNA was 20–23% of PGK1 WT mRNA before or after IP with either anti-Cbc1p or anti-eIF4E (Fig. 2B Lower). Because at least some Cbc1p binds to transcripts before eIF4E, and because some mRNA is bound by Cbc1p, these data indicate that NMD targets Cbc1p-bound mRNA and, therefore, Cbc1p-bound mRNA must be translated. Consistent with this interpretation, the level of PGK1 PTC mRNA in the upf1-Δ upf2-Δ upf3-Δ strain was 96–102% of PGK1 WT mRNA before or after IP with either anti-Cbc1p or anti-eIF4E (Fig. 6, which is published as supporting information on the PNAS web site). Furthermore, the ratio of Cbc1p-bound PGK1 mRNA that does and does not harbor a PTC, and the ratio of eIF4E-bound mRNA that does and does not harbor a PTC, were increased to essentially the same extent as the ratio of total-cell PGK1 mRNA that does and does not harbor a PTC in cells treated with the translational inhibitor CHX (Fig. 7, which is published as supporting information on the PNAS web site). We conclude that a Cbc1p-mediated pioneer round of translation does occur in S. cerevisiae. However, NMD may also target eIF4E-bound mRNA (see below).

Fig. 2.

PGK1 mRNA that harbors a PTC is targeted for NMD when bound by Cbc1p. (A) PTC-containing PGK1 mRNA is subject to NMD in a way that requires Upf proteins. Yeast strains B-10529 (Normal A) and B-15334 (upf1-Δ upf2-Δ upf3-Δ) were transformed with a yeast-centromeric plasmid carrying either a WT or PTC-containing PGK1 gene and the URA3-selective marker. RNA was purified, and RT-PCR was performed to quantitate the levels of PGK1 mRNA after normalization to the level of URA3 mRNA. The normalized level of PGK1 WT mRNA in each strain was defined as 100%. The five leftmost lanes represent twofold dilutions of RNA and demonstrate that the conditions used for RT-PCR were semiquantitative. Each mRNA was amplified by using 19 cycles. (B) The levels of Cbc1p- and eIF4E-bound PGK1 PTC mRNAs were reduced to the same percentage of Cbc1p- and eIF4E-bound PGK1 WT mRNAs, respectively. IPs were performed as described in Fig. 1 A by using the same B-10529 (Normal A) lysates as described in A.(Upper) Western blotting demonstrated that IP efficiencies were 20% for Cbc1p and 10% for eIF4E. The six leftmost lanes represent twofold dilutions of protein before IP (–) and demonstrate that the conditions used for Western blotting were semiquantitative. (Lower) RT-PCR demonstrated that the normalized level of Cbc1p-bound PGK1 PTC mRNA, when presented as a percentage of the normalized level of Cbc1p-bound PGK1 WT mRNA, was comparable to the normalized level of eIF4E-bound PGK1 PTC mRNA, when presented as a percentage of the normalized level of eIF4E-bound PGK1 WT mRNA. The four leftmost lanes represent twofold dilutions of RNA before IP (–) and demonstrate that the conditions used for RT-PCR were semiquantitative. PGK1 mRNAs were amplified by using 22 cycles, whereas URA3 mRNA was amplified using 19 cycles.

NMD Targets Cbc1p-Bound MET8 mRNA. To determine whether our finding that NMD targets Cbc1p-bound PGK1 mRNA applies to another Cbc1p-bound mRNA, we assessed met8–1 mRNA, which contains an UAG PTC. Extracts were prepared from strain B-11621 (MET8), which harbored a WT MET8 gene, and strain B-11622 (met8–1), which harbored a PTC-containing MET8 gene, with and without IP as described above. Western blotting revealed that the IPs were specific (Fig. 3 Upper). By using the level of ACT1 mRNA to control for variations in RNA recovery, RT-PCR analysis of samples before and after IP revealed that the level of met8–1 (MET8 PTC) mRNA was the same percentage of the level of MET8 WT mRNA before or after IP with either anti-Cbc1p or anti-eIF4E (Fig. 3 Lower). Therefore, Cbc1p-bound MET8 mRNA, and possibly, eIF4E-bound MET8 mRNA, are targeted for NMD. These data indicate that a Cbc1p-mediated pioneer round of translation is not particular to PGK1 mRNA but applies to at least one other, and possibly, other if not all, mRNAs.

Fig. 3.

Cbc1p-bound MET8 mRNA that harbors a PTC (met8-1) is targeted for NMD. Methods are the same as in Fig. 2B, except that extracts from strains B-11621 (MET8), which harbored a WT MET8 gene, and B-11622 (met8–1), which harbored a PTC-containing MET8 gene, were analyzed. (Upper) Western blotting demonstrated that IP efficiencies were 10–20% for Cbc1p and 5–10% for eIF4E. The three leftmost lanes represent twofold dilutions of protein before IP (–) and demonstrate that the conditions used for Western blotting were semiquantitative. (Lower) RT-PCR demonstrated that the normalized level of Cbc1p-bound MET8 PTC mRNA, when presented as a percentage of the normalized level of Cbc1p-bound MET8 WT mRNA, was comparable to the normalized level of eIF4E-bound MET8 PTC mRNA, when presented as a percentage of the normalized level of eIF4E-bound MET8 WT mRNA. The five leftmost lanes represent twofold dilutions of RNA before IP (–) and demonstrate that the conditions used for RT-PCR were semiquantitative. ACT1 mRNA was amplified by using 17 cycles, whereas MET8 mRNAs were amplified, using 22 cycles.

eIF4E-Bound PGK1 mRNA Is Targeted for NMD in a cbc1-Δ Strain. Given that eIF4E binds to the caps of a fraction of pre-mRNAs, and therefore, the caps of their product mRNAs, and because a PTC reduces the levels of Cbc1p-bound mRNA and eIF4E-bound mRNA to the same percentage of normal, NMD could target not only Cbc1p-bound mRNA but also eIF4E-bound mRNA. To determine whether eIF4E can bind to pre-mRNAs before Cbc1p, and to gain a better understanding of why cbc1-Δ strains are viable and support NMD (11), we transformed strains B-9037 (Normal B) and B-11558 (cbc1-Δ) with the same plasmid that produced PGK1 WT or PGK1 PTC mRNA as described above. As expected, RT-PCR demonstrated that the level of PGK1 PTC mRNA was 26% the level of PGK1 WT mRNA in the Normal B strain, and this percentage was essentially the same in the cbc1-Δ strain (Fig. 4A), which is consistent with previous reports (11). Western blotting revealed that the IP that used anti-eIF4E was successful (Fig. 4B Upper). RT-PCR of samples before and after IP revealed that (i) the ratio of eIF4E-bound ACT1 pre-mRNA to eIF4E-bound ACT1 mRNA manifested no detectable difference in the presence or absence of Cbc1p (Fig. 4B Lower; the lack of a detectable difference may reflect the low abundance of pre-mRNA relative to mRNA as measured by using RT-PCR); and similarly, (ii) the ratio of eIF4E-bound KIN28 pre-mRNA to eIF4E-bound KIN28 mRNA manifested no detectable difference in the presence or absence of Cbc1p (Fig. 4B Bottom). Furthermore, in a separately performed IP using the same strains and anti-eIF4E (Fig. 4C Top), the level of PGK1 PTC mRNA was the same percentage of PGK1 WT mRNA before or after IP, regardless of whether Cbc1p was present (Fig. 4C Middle and Bottom). These data suggest that a fraction of newly synthesized caps normally binds eIF4E without prior binding to Cbc1p. These data also indicate that eIF4E-bound mRNA can be targeted for NMD in the absence of Cbc1p, and most likely, also in the presence of Cbc1p (see Discussion). We conclude that whereas normally a fraction of eIF4E-bound PGK1 mRNA derives from Cbc1p-bound PGK1 mRNA (Fig. 1), eIF4E can bind to PGK1 transcripts without prior binding by Cbc1p.

Fig. 4.

eIF4E-bound PGK1 mRNA that harbors a PTC is targeted for NMD in a cbc1-Δ strain. (A) PGK1 PTC mRNA is targeted for NMD in Normal B and cbc1-Δ strains to the same extent. Methods are the same as in Fig. 2 A, except that B-9037 (Normal B) and B-11558 (cbc1-Δ) strains were analyzed. (B) Analysis of eIF4E-bound transcripts in Normal B and cbc1-Δ strains. (Upper) Western blotting demonstrated that IP efficiencies were 5–10% for eIF4E. The five leftmost lanes represent twofold dilutions of protein before IP (–) and demonstrate that the conditions used for Western blotting were semiquantitative. (Lower) The ratio of ACT1 pre-mRNA and mRNA is essentially the same in the two strains. The same applies to the ratio of KIN28 pre-mRNA to mRNA. (C) eIF4E-bound mRNA is targeted for NMD in a cbc1-Δ strain. IPs, Western blotting, and RT-PCR were as in Fig. 2B except that the anti-Cbc1p IP was omitted.

Discussion

Whether Cbc1p-bound mRNA is translated so that a Cbc1p-mediated pioneer round of translation occurs in S. cerevisiae has been a matter of debate. Cbc1p stimulates translation 2.5-fold when steady-state translation initiation is impaired in extracts prepared from strains harboring a mutated form of eIF4G that interacts only weakly with eIF4E and the poly(A)-binding protein (18). Moreover, eIF4E antagonizes the interaction of Cbc1p with eIF4GI (18). Whereas these data indicate that Cbc1p can direct translation initiation, subsequent studies led to the opposite conclusion. In these studies, growth rate, protein synthesis, and composition of the transcriptome were not affected in strains expressing a point-mutated eIF4G that no longer interacted with Cbc1p (19). However, these studies did not indicate that Cbc1p does not support translation but only that Cbc1p is not essential for translation, as we have previously shown (11), and in experiments reported here (Fig. 4). Furthermore, Cbc1p functions in ways other than those that can be attributed to its interacting with eIF4G (19), which is consistent with our demonstrating a dependence of nuclear mRNA degradation on Cbc1p (20). In this communication, we provide the first indication, to our knowledge, that Cbc1p-bound mRNA can be translated and that normal WT strains of yeast undergo a Cbc1p-mediated pioneer round of translation. However, a Cbc1p-mediated pioneer round of translation is not essential for a level of protein synthesis that supports cell viability or for NMD. Consistent with this conclusion, a pioneer round of translation can alternatively involve eIF4E-bound mRNA.

We also demonstrate that RNA metabolism in S. cerevisiae and in mammalian cells manifests significant differences. First, intron-containing pre-mRNAs in yeast can be bound by either Cbc1p or eIF4E (Fig. 1). In contrast, intron-containing pre-mRNAs in mammalian cells are detectably bound only by CBP80. Second, NMD in yeast targets both Cbc1p- and eIF4E-bound mRNA (Figs. 2, 3, 4; see model in Fig. 5). In contrast, NMD in mammalian cells detectably targets only CBP80-bound mRNA (1, 2, 4). Nevertheless, neither Cbc1p- nor eIF4E-bound mRNAs in yeast were detectably associated with the Upf3p NMD factor under conditions where both types of mRNA were detectably associated with other mRNP proteins such as Pab1p (data not shown). In contrast, CBP80-bound mRNA in mammalian cells is detectably bound by the orthologous Upf3 and Upf3X proteins (1, 4).

Fig. 5.

Model for the translation of Cbc1p- and eIF4E-bound mRNAs in S. cerevisiae. In the nucleus, a yeast gene (thick bar) is transcribed so as to produce mRNA that is bound at the 5′ cap by either Cbc1p, which binds as a heterodimer with Cbc2p, or eIF4E. After mRNA export to the cytoplasm (21), both Cbc1p- and eIF4E-bound mRNAs initially undergo a pioneer round of translation. Cbc1p-bound mRNA is remodeled to eIF4E-bound mRNA during or after the pioneer round. Additionally, both Cbc1p- and eIF4E-bound pioneer translation initiation complexes undergo other steps of mRNP remodeling that include the loss and acquisition of mRNA-binding proteins (22). Data presented here and elsewhere (see Discussion) indicate that NMD can target either Cbc1p- or eIF4E-bound mRNA during any round of translation provided that an mRNA harbors the appropriate nonsense codon.

In support of our finding that NMD targets newly synthesized Cbc1p- or eIF4E-bound mRNA, NMD in yeast has been shown to take place without significant shortening of the mRNA poly(A) tail (23, 24). Furthermore, the RNA-binding protein Hrp1p/Nab4p, which has nuclear roles in transcript processing and export, associates with a putative destabilizing element early in the biogenesis of mini-PGK1 mRNA and reportedly recruits Upf1p (possibly analogously to how the exon–junction complex recruits Upf1) so as to elicit NMD if there is an upstream nonsense codon (25). Nevertheless, other studies of yeast indicate that NMD also targets steady-state mRNA, indicating that mRNP proteins that are required for NMD are not shed during translation or are shed but can reassociate. In support of NMD targeting steady-state mRNA, nonsense-containing transcripts in cells in which RNA polymerase II had been thermally inactivated (i) accumulated on polysomes in the presence of CHX and (ii) continued to be degraded (i.e., were lost from polysomes) once CHX had been washed away (26). Because polysomes consist largely of steady-state mRNA, the resumed disappearance of polysome-associated mRNA after the removal of CHX suggests that NMD targets steady-state mRNA. Also in support of NMD targeting steady-state mRNA, nonsense-containing transcripts underwent decay in cells in which each of the three Upf NMD factors, under the control of a galactose-inducible promoter, was induced after a period of repression (27). Because the repression of Upf NMD factors lasted long after most mRNAs had been synthesized and exported from nuclei, the resumed disappearance of cytoplasmic mRNA after the induction of these factors suggests that NMD targets steady-state mRNA. NMD has also been argued to target steady-state mRNA in experiments that used programmed –1 ribosomal frameshift signals. When these signals were inserted into a reporter mRNA so that ribosomes encountered a downstream nonsense codon at low frequencies (i.e., 1–12%), decay appeared to occur after the transcriptional shutoff of mRNA production to an extent that might not be limited to the newly synthesized mRNA (28).

Future studies will undoubtedly shed additional light on the nature of Cbc1p-bound mRNA translation and its significance to NMD in yeast.