INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Short peptide coding regions in the 5' leaders of prokaryotic mRNAs (leader peptides) and eukaryotic mRNAs (upstream open reading frames [uORFs]) can serve critical regulatory functions. For example, analyses of evolutionarily conserved mechanisms regulating bacterial amino acid biosynthetic gene expression revealed the phenomenon of transcription attenuation, in which the movement of ribosomes over specific leader peptide coding sequences regulates operon expression (14). Studies of fungal genes involved in multiple-pathway control of amino acid biosynthesis, particularly those of Saccharomyces cerevisiae GCN4, have revealed the importance of multiple, cis-acting uORFs in the gene's mRNA 5' leader and of modulating the activity of trans-acting translation factors in regulation (10).

The Neurospora crassa arg-2 gene was one of the first amino acid biosynthetic genes to be identified (22). It encodes the small subunit of arginine-specific carbamoyl phosphate synthetase. It is unique among N. crassa Arg biosynthetic genes in that it is negatively regulated by the concentration of Arg in the cell (4). The arg-2 mRNA contains a uORF specifying a 24-residue peptide (20). The uORF encoding this peptide, henceforth called the arg-2 arginine attenuator peptide (AAP) because of its involvement in Arg-specific translational regulation, is evolutionarily conserved; similar AAP sequences are encoded by uORFs in all of the other fungal genes specifying this enzyme that have been characterized so far (1, 21, 27).

Elimination of the uORF initiation codon in the N. crassa arg-2 or the homologous S. cerevisiae CPA1 transcripts shows that uORF translation is crucial for Arg-specific regulation in vivo (9, 19, 27). In N. crassa, the arg-2 uORF reduces translation of ARG2 in vivo by reducing the average number of ribosomes associated with the arg-2 mRNA when Arg is plentiful in the growth medium (18). Arg-specific translational regulation mediated by the arg-2 uORF has been reconstituted in vitro by using a homologous cell-free translation system (25). A primer extension inhibition assay that allowed mapping of ribosomes on RNA at positions corresponding to rate-limiting steps in translation has been developed (26). Through high-resolution mapping of primer extension products, this toeprint assay enabled the localization of ribosomes on RNAs that are engaged in initiation with AUG codons in their P sites and ribosomes engaged in termination with termination codons in their A sites. Ribosome movement that was slowed during elongation by limitation for specific amino acids became stalled with codons for the limiting amino acid in their A sites (26).

Toeprint assays indicate that a high level of Arg causes ribosomes translating the arg-2 uORF to stall with its termination codon in the ribosomal A site (26). This primary event of Arg-regulated stalling of ribosomes is hypothesized to result in Arg-specific negative regulation because the stalled ribosomes block ribosomal scanning from the 5' end of the mRNA and therefore block trailing ribosomes from reaching the downstream initiation codon. Whether termination is required for Arg-regulated ribosome stalling remains an important question that is not answered by these studies.

Amino acid residues in the arg-2 uORF peptide critical for its function in Arg-specific regulation have been identified. For example, changing Asp-12 of the uORF coding region product to Asn (D12N) eliminates Arg-specific translational control in vivo (9) and in vitro (25, 26). In parallel fashion, mutation of the corresponding Asp residue in the S. cerevisiae CPA1 uORF product eliminates Arg-specific regulation in vivo (5, 27). Other mutations in the N. crassa and S. cerevisiae uORFs that change the predicted amino acid sequence in the evolutionarily conserved region of the uORF peptide also reduce or eliminate Arg-specific translational control (19, 26, 27). When tested in the N. crassa cell-free system, loss of regulation is associated with reduced stalling at the uORF termination codon (26).

In the present study, we investigated the requirements for N. crassa arg-2 uORF function in translational regulation by using the N. crassa cell-free translation system. We first examined the effects of shortening the distance between the uORF termination codon and the downstream initiation codon. Subsequently, we changed the uORF termination codon, which is normally UAA, to UAG and UGA, to determine whether Arg-specific stalling at the termination codon is codon specific. Finally, to test whether negative translational regulation by Arg requires the uORF termination codon and a downstream initiation codon, we fused functional wild-type AAP and nonfunctional mutant AAP peptides directly to the luciferase (LUC) reporter at the LUC N terminus. The results obtained indicate that AAP-mediated, Arg-specific stalling of ribosomes occurs immediately following AAP translation. Translation of the AAP results in a situation in which the movement of ribosomes involved in either termination or elongation is stalled in response to Arg. While stalling requires a specific nascent peptide sequence, it does not appear to require specific RNA sequences distal to the AAP coding region. The results indicate that the nascent peptide encoded by the arg-2 uORF modulates the expression of a downstream gene product by affecting ribosome movement. The regulation of movement of ribosomes involved in either termination or elongation reveals a novel form of eukaryotic translational control.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Construction of templates for RNA synthesis. Plasmids were designed to produce capped and polyadenylated synthetic RNA encoding firefly LUC with wild-type or mutant arg-2 sequences in the RNA 5' leader region (Fig. 1 and Table 1). Mutations were introduced into this region by PCR with mutagenic primers (Table 1) by using procedures described previously (9). PCR products were ligated into the pHLUC+NFS4 vector (Table 1) (25). A plasmid designed to produce capped and polyadenylated synthetic RNA which lacked arg-2 sequences and that encoded sea pansy LUC was constructed by replacing the firefly LUC coding region of pHLUC+NFS4 with the sea pansy LUC coding region of pRL-CMV (Promega).

View larger version (30K): [in this window] [in a new window]   FIG. 1.   The 5' leader regions of arg-2-LUC genes used in this study (see also Table 1). (A) Sequences of wild-type and mutant templates in which the AAP is encoded by a uORF. The sequence shown begins with the T7 RNA polymerase-binding site and ends within the LUC coding region (26). The amino acid sequences of the arg-2 AAP and the N terminus of LUC are indicated. Point mutations are shown below the wild-type sequence. The endpoints of deletion mutations that shorten the intercistronic region are indicated by the boxed nucleotides. All deletions share the 3' endpoint, which is indicated by a horizontal arrow above the sequence. The extent of each deletion is indicated (e.g., 71 removes the greatest number of nucleotides, leaving 5'-CC-3' between the uORF termination and LUC initiation codons). The sequence for which the reverse complement was synthesized and used as primer ZW4 for toeprint analysis is indicated by a horizontal arrow below the sequence. (B) Sequences of templates containing wild-type and mutant AAP-LUC fusion genes. The sequence shown begins with the T7 RNA polymerase-binding site and ends within the LUC coding region; the amino acid sequence of the N terminus of the AAP-LUC fusion polypeptide is indicated. Point mutations are shown below the wild-type sequence. The mutation indicated by  improves the initiation context for uORF translation. The fs mutation is a 1 frameshift in which the first nucleotide of the Gln codon bridging the AAP and LUC coding sequences is deleted.

Cell-free translation of RNA and analyses of translation products. Plasmid DNA templates were purified by equilibrium centrifugation and linearized with Ppu10I. Capped, polyadenylated RNA was synthesized with T7 RNA polymerase from linearized plasmid DNA templates, and the yield of RNA was quantified as described previously (25). The preparation of cell translation extracts from N. crassa and the reaction conditions for in vitro translation were as described previously (25, 26).

Primer extension inhibition (toeprint) assays. The primer extension assays were accomplished as described previously by using primer ZW4 (26); 8 µl of sample instead of 4 µl was loaded onto each gel lane. The gels were dried and exposed to screens of a Molecular Dynamics Phosphorimager for approximately 24 h. All toeprint data shown are representative of multiple experiments.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Effects of reducing the distance between the uORF termination codon and the downstream initiation codon. Wild-type arg-2 mRNA contains 63 nucleotides (nt) between the uORF stop codon and the ARG2 initiation codon. The arg-2 uORF coding sequence and the intercistronic region were placed upstream of the sequence coding for firefly LUC (another 10 nt of intergenic vector sequences are also present) (Fig. 1). This construct was used to produce capped and polyadenylated synthetic RNA. In vitro translation of this RNA in the presence of a high concentration of Arg (500 µM), in contrast to a low concentration of Arg (10 µM), results in an approximately twofold reduction in LUC polypeptide synthesis (Table 1) (see references 25 and 26). The D12N mutation of the AAP coding sequence (Fig. 1) eliminates the regulatory effect of the AAP (Table 1) (25, 26).

View larger version (172K): [in this window] [in a new window]   FIG. 2.   Effects of shortening the distance between the uORF termination codon and the downstream LUC initiation codon on Arg-specific regulation. Equal amounts of synthetic RNA transcripts (120 ng) were translated in 20-µl reaction mixtures for 20 min at 25°C. Reaction mixtures contained 10 µM () or 500 µM (+) Arg and a 10 µM concentration of each of the other 19 amino acids. The transcripts examined are indicated at the top of the lanes. After 20 min of translation, the translation mixtures were toeprinted with primer ZW4 and analyzed next to dideoxynucleotide sequences of the corresponding DNA template. The nucleotide complementary to the dideoxynucleotide added to each reaction mixture is indicated above the corresponding lane so that the sequence of each template can be directly deduced; the 5'-to-3' sequence reads from top to bottom. The asterisks indicate the positions of premature transcription termination products corresponding to ribosomes at the uORF termination codon; brackets indicate ribosomes stalled behind those at the termination codon. The closed arrowheads indicate ribosomes at the uORF initiation codon; the open arrowheads indicate ribosomes at the LUC initiation codon. wt, wild-type.

Effects of altering the uORF termination codon. The uORF termination codon, UAA, appears to be the most frequently used in N. crassa (6). To test whether this specific termination codon was important for Arg-specific regulation, plasmids were constructed to produce synthetic RNA in which the wild-type or D12N mutant uORFs were terminated with UAA, UAG, or UGA codons. With the wild-type uORF, all three termination codons showed similar levels of Arg-specific regulation as determined by luciferase activity assays (Table 1, constructs pPR102, pRF102, and pRF103). Comparisons of toeprint assays of each RNA translated in reaction mixtures containing either low or high Arg concentrations indicated that the behaviors of ribosomes at each of these termination codons were indistinguishable (Fig. 3). Regardless of which uORF termination codon was present, the D12N mutant uORF conferred no Arg-specific regulation (Table 1 and Fig. 3). Therefore, while the sequence of the uORF coding region is important for regulation, the type of termination codon used to stop uORF translation did not appear to be important.

View larger version (158K): [in this window] [in a new window]   FIG. 3.   Effects of terminating uORF peptide synthesis with each of the three termination codons, UAA (lanes 1 to 4), UAG (lanes 5 to 8), and UGA (lanes 9 to 12), on Arg-specific regulation. Equal amounts of synthetic RNA transcripts (120 ng) were translated in 20-µl reaction mixtures for 20 min at 25°C. Reaction mixtures contained 10 µM () or 500 µM (+) Arg and a 10 µM concentration of each of the other 19 amino acids; they were analyzed by toeprinting as described in the legend to Fig. 2. The RNAs are from 71 deletion constructs containing either the wild-type (wt) or D12N mutant uORFs terminated with UAA, UAG, or UGA codons (Fig. 1 and Table 1) as indicated at the top of the lanes. Dideoxynucleotide sequencing reactions for the pPR102 template containing the wild-type uORF terminated with UAA are shown on the left (lanes C', T', A', and G'). Arrows indicate the positions of premature transcription termination products corresponding to ribosomes bound at the uORF initiation codon (AUGAAP) and the uORF termination codon (TermAAP); brackets indicate the positions of ribosomes stalled behind those at the termination codon.

Effects of fusing the arg-2 uORF peptide directly to LUC. The experiments described above indicate that the function of the arg-2 uORF in regulation was retained regardless of the distance between its termination codon and the downstream initiation codon and regardless of which termination codon was used to terminate uORF translation. Therefore, we next investigated whether the termination of uORF peptide synthesis followed by the initiation of new polypeptide synthesis at a downstream start codon was necessary for AAP-mediated translational regulation. Constructs were made in which the normal LUC initiation codon was eliminated and the wild-type AAP coding region was fused directly to the LUC coding region. In such constructs, the AAP initiation codon would be responsible for initiating translation of an AAP-LUC fusion polypeptide.

View larger version (74K): [in this window] [in a new window]   FIG. 4.   Analyses of [35S]methionine-labeled polypeptides produced by translation of synthetic RNA transcripts in N. crassa cell extracts. Micrococcal nuclease-treated N. crassa extracts (20 µl containing 2 µCi of [35S]methionine) were programmed with 120 ng of the indicated RNAs and incubated for 30 min at 25°C. Reactions were stopped by immersing the tube in liquid nitrogen and examined by SDS-PAGE in 10% polyacrylamide gels. Radiolabeled translation products were visualized by phosphorimaging; the positions of molecular mass markers (in kilodaltons) visualized by staining with Coomassie blue are indicated on the left. Lanes: 1, reaction mixture with no added RNA; 2, with RNA encoding firefly LUC; 3, with RNA encoding the AAP-LUC fusion polypeptide in the wild-type initiation context; 4, with RNA encoding the AAP-LUC fusion polypeptide in the improved initiation context; 5, with RNA encoding the AAP coding region frameshifted (fs) with respect to the LUC coding region; 6, with RNA encoding sea pansy (sp) LUC.

View larger version (12K): [in this window] [in a new window]   FIG. 5.   Time courses of translation of RNAs encoding AAP-LUC, D12N AAP-LUC, and sea pansy LUC enzymes in reaction mixtures containing 10 or 500 µM Arg. Translation in reaction mixtures (120 µl) was initiated by using a mixture of RNAs: 150 ng of RNA encoding the wild-type AAP-LUC fusion and 36 ng of RNA encoding sea pansy LUC (A) or 150 ng of RNA encoding the D12N AAP-LUC fusion and 36 ng of RNA encoding sea pansy LUC (B). Reaction mixtures were incubated at 25°C and contained either 10 or 500 µM Arg and a 10 µM concentration of each of the other 19 amino acids. Firefly LUC production (in reaction mixtures containing 10 [] or 500 [] µM Arg) and sea pansy LUC production (in reaction mixtures containing 10 [] or 500 [×] µM Arg) were determined by using the dual luciferase assay as described in Materials and Methods. RLU, relative light units.

View larger version (11K): [in this window] [in a new window]   FIG. 6.   Time courses of translation of RNAs encoding the wild-type AAP as an N-terminal domain and as a uORF product. Reaction conditions were as described in the legend to Fig. 5. (A) Results with 150 ng of RNA encoding the wild-type AAP-LUC fusion and 36 ng of RNA encoding sea pansy LUC; (B) results with 150 ng of RNA encoding the wild-type AAP as a uORF region and LUC as a separate downstream coding region with its own initiation codon and 36 ng of RNA encoding sea pansy LUC. The inset in panel A shows the combined data for translation of the internal control sea pansy LUC RNA in the four reaction mixtures plotted in panels A and B. Firefly LUC production (in reaction mixtures containing 10 [] or 500 [] µM Arg) and sea pansy LUC production were determined by using the dual luciferase assay as described in Materials and Methods. RLU, relative light units.

View larger version (113K): [in this window] [in a new window]   FIG. 7.   Effects of mutations on Arg-specific regulation of AAP-LUC fusion constructs. Equal amounts of synthetic RNA transcripts (120 ng) were translated in reaction mixtures and analyzed by toeprinting as described in the legend to Fig. 2. The transcripts encoded the wild-type (wt) AAP-LUC fusion or the D12N mutant AAP-LUC fusion as indicated in either the wild-type or improved () initiation contexts. Dideoxynucleotide sequencing reactions for the template containing the wild-type AAP-LUC fusion are shown on the left (lanes C', T', A', and G'). The products obtained from primer extension of pure AAP-LUC RNA (18 ng) in the absence of translation reaction mixture (EXT; lane 9) and from a translation reaction mixture not programmed with RNA (RNA; lane 10) are shown for comparison. The arrow indicates the position of the premature transcription termination products corresponding to ribosomes bound at the AAP initiation codon (AUGAAP). The arrowhead indicates the position of premature termination products corresponding to ribosomes stalled in the presence of a high level of Arg at the codon immediately following the 24 codons of the AAP. The bracket indicates the position of premature termination products corresponding to ribosomes stalled in the presence of a high level of Arg in the LUC coding region.

View larger version (79K): [in this window] [in a new window]   FIG. 8.   Effects of puromycin on Arg-specific regulation of AAP-LUC fusion constructs. Equal amounts of synthetic RNA transcripts (120 ng) were translated in reaction mixtures and analyzed by toeprinting as described in the legend to Fig. 2. The transcripts encoded either the D12N mutant AAP-LUC fusion or the wild-type (wt) AAP-LUC fusion as indicated. Puromycin (Pur) was added where indicated, as described in the text. Dideoxynucleotide sequencing reactions for the template containing the wild-type AAP-LUC fusion are shown on the left (lanes C', T', A', and G'). The arrow indicates the position of the premature transcription termination products corresponding to ribosomes bound at the AAP initiation codon (AUGAAP). The arrowhead indicates the position of premature termination products corresponding to ribosomes stalled in the presence of a high level of Arg at the codon immediately following the 24 codons of the AAP. The bracket indicates the position of premature termination products corresponding to ribosomes stalled in the presence of a high level of Arg in the LUC coding region.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The leader region of the N. crassa arg-2 mRNA contains a 24-residue, evolutionarily conserved uORF encoding the AAP. Translation of the AAP coding region appears necessary for Arg-regulated stalling of ribosomes. We observed that translational control through the AAP is cis-acting and does not appear to absolutely require specific downstream translational events to occur. Shortening the distance between the uORF termination codon and the downstream LUC initiation codon did not affect Arg-regulated stalling at the uORF termination codon or the extent of negative regulation conferred on translation initiated at the LUC initiation codon. Changing the uORF termination codon from UAA to UAG or UGA did not affect regulation. Indeed, negative translational regulation was observed when the AAP coding region was fused in-frame directly to the LUC coding region. The data suggest that the nascent AAP acts in cis within the ribosome that has translated it to cause stalling regardless of whether the ribosome is subsequently engaged in termination or elongation.

The regulatory function of the AAP to stall ribosomes was examined in two ways: by examining the kinetics of Arg-specific regulation in the N. crassa cell-free translation system and by mapping the positions of ribosomes on RNA in this system by using a primer extension inhibition (toeprint) assay. When the AAP is encoded by a uORF, the time of initial appearance of functional LUC enzyme (whose synthesis is initiated from a downstream start codon) is the same regardless of whether the reaction mixture contains a low or high concentration of Arg (Fig. 6B). The negative effect of Arg becomes apparent only later, when it reduces the rate of LUC enzyme accumulation. These data are consistent with the proposed model for regulation (26) mediated by the arg-2 uORF in which some of the first scanning ribosomal subunits loaded on the mRNA leak past the codon initiating AAP and initiate instead at the LUC coding region regardless of whether the level of Arg is high or low. In this model, the time at which the first functional LUC polypeptide is translated from such an mRNA pool is not affected by Arg, but Arg reduces LUC polypeptide synthesis subsequently.

In contrast to the situation in which the AAP is encoded by a uORF, for RNA encoding the AAP-LUC fusion polypeptide, ribosomes initiating at the AAP codon are directly responsible for LUC synthesis, as judged by analysis of radiolabeled LUC polypeptide and LUC enzyme activity produced from such RNAs. Thus, all of the ribosomes that initiate LUC synthesis necessarily begin at the AAP start codon. Since each of these ribosomes must translate the AAP, they will be subject to AAP-mediated, Arg-specific ribosome stalling. Therefore, it would be predicted that the time of the first appearance of AAP-LUC fusion polypeptide would be delayed by Arg. This is what is observed (Fig. 6A).

The toeprint data (Fig. 2 and 3) indicate that translation of the wild-type AAP causes ribosomes engaged at termination codons to stall, which has parallels with other uORFs whose peptide coding sequences are important for controlling translation (2, 17, 26). In addition, the AAP also causes ribosomes engaged in elongation to stall (Fig. 7 and 8). The data indicate that stalling can occur in a short region of RNA following the AAP coding region. Previous considerations of how the movement of eukaryotic ribosomes involved in elongation is translationally controlled have focused on the physical structure of the RNA or on the limitation for charged tRNA. Secondary structures in the RNA have been implicated in the blockade of scanning 40S ribosomal subunits (12) and in the slowing of translating 80S ribosomes (23). The presence of pseudoknots or rare codons in the RNA are implicated in ribosome frameshifting (7). It has been hypothesized that rare codons affect ribosome stalling and allow time for the correct folding of nascent domains of a protein containing multiple domains (11, 13, 23).

The relative importance of different cis-acting sequences for N. crassa arg-2 AAP-mediated regulation deduced by using a cell-free translation system from Neurospora correlates with observations on CPA1 regulatory sequence effects on reporter gene expression in vivo in S. cerevisiae (5). In these studies, the activity of the enzyme encoded by the reporter -galactosidase gene was used to measure regulation; neither the distribution of ribosomes on RNA nor the levels of RNA were measured. Therefore, it was not possible to assess the relative contributions of Arg-specific translational control and Arg-specific effects on CPA1 transcription and CPA1 transcript stability (3). Nevertheless, these observations on regulation in vivo are entirely consistent with observations on regulation in the cell-free translation system. Changing the distance between the CPA1 uORF termination codon and the downstream initiation codon did not alter Arg-specific regulation (5). When the full-length CPA1-encoded AAP was fused directly to the -galactosidase reporter gene product, Arg-specific regulation was retained (5). These results are also consistent with the in vitro studies on N. crassa arg-2, although, in addition to the caveats listed above, a deliberately frameshifted construct was not tested in vivo to rule out the possibility that another initiation codon than that predicted was used for translation of the functional reporter.

The results presented here indicate that the movement of ribosomes involved in termination or elongation can be regulated by the nascent peptide being produced. The strong conservation of the AAP amino acid sequence among different fungi, with nucleotide differences in the coding region primarily in silent positions, and the demonstrated importance of the evolutionarily conserved peptide sequence for the regulatory function of stalling ribosomes in the Neurospora system indicate a primary role for the nascent peptide. Specific RNA sequences may also contribute to this regulatory process by other means than their capacity to encode polypeptide sequence. No evidence that identifies such sequences has as yet been obtained.

The observation that translation of the AAP can regulate the movement of ribosomes involved in elongation provides evidence for a novel form of eukaryotic translational control. In bacterial systems, there are precedents for the stalling of ribosomes involved in elongation mediated by nascent peptides (17). Nascent peptides can also have other effects on ribosomes involved in elongation: ribosome jumping in the translation of T4 gene 60 requires a 16-residue region of the nascent peptide (15). In eukaryotes, N-terminal nascent peptide domains can interact with the signal recognition particle (24), which halts ribosome movement until docking with membrane of the endoplasmic reticulum. Interestingly, recent evidence shows that the nascent signal polypeptide interacts with the ribosome before it exits the ribosome (16).

Taken together, all of the available data indicate that translation of the AAP is an evolutionarily conserved event of primary importance for regulation. AAP-mediated translational regulation represents an unusual instance in which a nascent peptide appears to regulate the movement of ribosomes.