|Home | About | Journals | Submit | Contact Us | Français|
In eubacteria, stalled ribosomes are rescued by a conserved quality-control mechanism involving transfer-messenger RNA (tmRNA) and its protein partner, SmpB. Mimicking a tRNA, tmRNA enters stalled ribosomes, adds Ala to the nascent polypeptide, and serves as a template to encode a short peptide that tags the nascent protein for destruction. To further characterize the tagging process, we developed two genetic selections that link tmRNA activity to cell death. These negative selections can be used to identify inhibitors of tagging or to identify mutations in key residues essential for ribosome rescue. Little is known about which ribosomal elements are specifically required for tmRNA activity. Using these selections, we isolated rRNA mutations that block the rescue of ribosomes stalled at rare Arg codons or at the inefficient termination signal Pro-opal. We found that deletion of A1150 in the 16S rRNA blocked tagging regardless of the stalling sequence, suggesting that it inhibits tmRNA activity directly. The C889U mutation in 23S rRNA, however, lowered tagging levels at Pro-opal and rare Arg codons, but not at the 3′ end of an mRNA lacking a stop codon. We concluded that the C889U mutation does not inhibit tmRNA activity per se but interferes with an upstream step intermediate between stalling and tagging. C889 is found in the A-site finger, where it interacts with the S13 protein in the small subunit (forming intersubunit bridge B1a).
Defects in protein synthesis threaten cell viability in eubacteria. mRNA transcripts lacking stop codons (nonstop mRNAs) arise from premature termination of transcription and from mRNA decay by 3′-5′ exonucleases. Since release factors cannot be recruited to nonstop mRNAs, ribosomes stall at the 3′ end and trap additional ribosomes upstream. In Escherichia coli, roughly 1 in every 250 translation reactions leads to ribosome stalling, so that if they were not rescued and recycled, all the ribosomes would be stalled in a single generation (23).
A highly conserved quality control system shields eubacterial cells from the negative effects of ribosome stalling (for a review, see reference 24). All known eubacterial genomes contain genes encoding transfer-messenger RNA (tmRNA) and its protein partner, SmpB (2). As implied by its name, tmRNA has two functions. First, tmRNA mimics a tRNA: it is aminoacylated with alanine, enters the A site of stalled ribosomes together with SmpB, and adds Ala to the nascent peptide chain. tmRNA then serves as an mRNA template, encoding a short peptide tag that is recognized by cellular proteases (15). After this tag is translated, the ribosome is released at a stop codon within tmRNA, and the aborted protein is degraded. Because the ribosome translates a protein from two RNA templates, this tagging process is known as trans-translation.
In addition to rescuing ribosomes on nonstop mRNAs, tmRNA can also act on ribosomes stalled on intact transcripts. Stalling can occur during elongation at clusters of rare codons, such as the Arg codons AGA, AGG, and CGA (9, 18, 30). The ribosome pauses on rare codons because the cognate tRNAs are in low abundance and therefore bind and react slowly. Stalling can also occur during translational termination when release of the nascent polypeptide is inefficient. For example, tmRNA tags the full-length YbeL protein in E. coli; the C-terminal Glu-Pro sequence in YbeL inhibits release factor activity (8). When combined with an inefficient opal stop codon (UGA), a C-terminal Pro residue can lead to levels of stalling so high that 40% of the full-length protein is tagged by tmRNA (8).
We have taken a genetic approach to the study of ribosome stalling and rescue, developing genetic selections for and against tmRNA function. While tmRNA is essential in some bacterial species (11), deletion of the tmRNA gene ssrA in E. coli yields only minor phenotypic changes (14). We previously reported a KanR-based selection that ties the life of the cell to tmRNA function (33). This selection has proved valuable in studying mutants of tmRNA and SmpB to understand how the ribosome resumes translation on tmRNA (33, 36). Here, we report two novel selections in which tmRNA function led to cell death. To our knowledge, these are the first genetic selections developed against tmRNA activity. We applied these negative selections to identify rRNA mutations that inhibit tagging at two different stalling sequences. Our goal was to understand which ribosomal elements are specifically required for trans-translation and the role of the ribosome in the rescue process.
The pbar-R selection plasmid was generated by amplifying the barnase gene with PCR using the primers 3483 (CATATGGCTAGCCAGGTTATCAACACGTTTGAC) and 3600 (GAGCTCGAATTCGTCACCTTGTTTTGTAAATCAGCCAGTCGC) (Table (Table1).1). The PCR product was cloned downstream of the araBAD promoter in pBAD-GFP using the NheI and EcoRI restriction sites. The pbar-P selection plasmid was cloned in a similar manner, except that barnase was amplified with primers 3483 and 3722 (GAGCTCGAATTCGTCAGGGTGAGTAAAGAATCCGGTCTG). The ptmRNA-B1 expression plasmid was generated by amplifying the pKW11 plasmid (29) by inverse PCR using the primers 3486 (GATCCGCTAATAAAGGATTCTTGGTCCTCTCTCCCTAGCCTCC) and 3487 (TTGGTAAAGGTCTGATAATGGTCGACTATTTTTTGCGGCTTTTTAC). Following phosphorylation with polynucleotide kinase, the blunt ends of the PCR product were ligated together. The ptmRNA-B2 plasmid was cloned in a similar manner with primers 3486 and 3721 (TTGGTAAAGGTCTGATAATGGTCTGTTGTTTTGTAAATCAGCCAGT CGACTATTTTTTGCGGCTTTTTA).
To create the ptRNA-BT selection plasmids, the tetR and ssrA genes were amplified from the ptmRNA-B1 or -B2 plasmid described above with the primers TTCGCCTCCGGATTGACAGCTTATCTTCGATTAGCTTTAATGCGGTAGTTTATC and ACTCACTCCGGAAAATAAATCCTGGTGTCCCTG, and the PCR product was inserted into the BspEI site in ptRNA-67 (3). Into the HindIII site of the resulting plasmids we inserted the araC gene and bar-R or bar-P expression cassettes amplified with the primers ACCAGCAAGCTTTTCACCGTCATCACCGAAAC and TAAACACAAGCTTCGCAGAAAGGCCC.
The maltose binding protein (MBP) vectors were derivatives of pMAL-C2G (New England Biolabs) in which the lacI coding sequence had been deleted and MBP was expressed constitutively from the lacI promoter. We combined this MBP expression cassette with a pCDF origin and a TetR marker, creating pCDF-MBP. To add stalling sequences, the 3′ end of the MBP gene was amplified with the primer TAACAAAGATCTGCTGCCGAACC and either GGAAGCCTGCAGTCAGGGTGAAGTCTGCGCGTCTTTCAGGG (Ser-Pro-opal), GGAAGCCTGCAGTCACCTCCTCCTCCTCCTCCTCCTCCTAGTCTGCGCGTCTTTCAGGGC (eight rare AGG Arg codons), GGAAGCCTGCAGTCACCTAGTCTGCGCCTCTTTCAGGGC (Arg-opal), or ACAAGGCTGCAGAAAAAAGCCCGCTCATTAGGCGGGCTGCGGTCTGCGCGTCTTTCAGGG (nonstop, adding the trpA terminator). These PCR products were digested with BglII and PstI and ligated into the expression vector.
The glutathione S-transferase (GST) recoding vectors were created by inverse PCR of pCDF-MBP with the primers CCCTACTGCAGGCAAGCTTGGCACTG and TTGCTAGGTACCATTCACCACCCTGAATTGACTC to add PstI and KpnI sites. GST was amplified from pGEX-3 (GE Healthcare) using the forward primer TTGCTAGGTACCATGTCCCCTATACTAGGTTATTG and one of the following reverse primers: GGAAGCCTGCAGTTATTACTTATCGTCATCGTCCTTATAGTCTTGTCAGGGAGAGAGGATCCCACGACC (readthrough), GGAAGCCTGCAGTTATTACTTATCGTCATCGTCCTTATAGTCTTGGTCAGGGAGAGAGGATCCCACGACC (+1 frameshifting), or GGAAGCCTGCAGTTATTACTTATCGTCATCGTCCTTATAGTCTTGGGTCAGGGAGAGAGGATCCCACGACC (−1 frameshifting). The GST amplicons were digested and ligated into the PstI and KpnI sites. araC and the araBAD promoter were amplified with the primers AGATCCACTAGTACTGACAGATCTTCGCTACGTGACTGGGTC and GGACATGGTACCCATATGTATATCTCCTTCTTAAAGTTAAAC and inserted between SpeI and KpnI sites upstream of GST to drive its expression.
rRNA libraries generously provided by Alexander Mankin (38) were amplified and introduced into the SQ171 strain carrying either the ptRNA-BT1 or -BT2 selection plasmid. The cells were rescued for 2 h at 30°C in 2×YT with 0.2% glucose and then added to 400 ml of 2×YT, 0.2% glucose, and 100 μg/ml ampicillin. The culture was grown for 20 h at 37°C to allow loss of the wild-type ribosome plasmid, ribosome turnover, and the synthesis of the new mutant ribosomes. The culture was then plated onto medium containing ampicillin and 2% arabinose to induce barnase expression. Plasmid DNA was extracted from the pool of surviving cells and resubjected to the selection procedure.
Tagging of the MBP protein was assayed by immunoblotting as described previously (22), except that expression of MBP was constitutive, not induced with IPTG (isopropyl-β-d-thiogalactopyranoside). Analysis of recoding events on GST was performed similarly, except that expression was induced with 2% arabinose for 2 h and readthrough or frameshifting was detected with a monoclonal anti-FLAG antibody (Sigma).
To isolate rRNA mutations that prevent tmRNA function, we created a genetic selection that links tagging to cell death. We altered tmRNA so that instead of tagging proteins for proteolysis, it completes the synthesis of a toxic protein. Barnase is a 110-residue RNase that is highly toxic; it has been used previously in negative selections with excellent results (13, 35). A catalytic base, His102, is required for activity (21). We removed this critical residue by deleting the last 10 amino acids in the barnase protein. The truncated protein was harmless.
We created a tmRNA mutant (tmRNA-B1) that rescues ribosomes stalled during barnase synthesis and completes the protein by encoding the last 10 amino acids, DHYQTFTKIR (Fig. (Fig.1A).1A). To induce stalling at the proper site, we mutated Thr100 to the rare Arg codon AGG, followed by the opal stop codon UGA, yielding the bar-R construct. As shown previously, inefficient release at the opal stop codon traps the cognate Arg-tRNA, leading to further depletion of this rare tRNA and stalling at the Arg codon (9). Upon rescue, the nascent barnase protein is transferred to Ala-tmRNA and the ribosome switches to the tmRNA-B1 open reading frame, completing the toxic protein. Barnase synthesized through the tagging process contains only one mutation, Thr100Ala.
To validate this negative selection, we transformed a strain lacking tmRNA (X90 ssrA::cat) with a plasmid carrying bar-R driven by the arabinose-inducible araBAD promoter. Plating of serial dilutions revealed that bar-R alone is nontoxic, as cells grew equally well on glucose or arabinose (Fig. (Fig.1A).1A). In contrast, fewer than 1 in 104 cells containing both bar-R and tmRNA-B1 survive on media containing arabinose (Fig. (Fig.1A).1A). These results show that stalling and rescue of the barnase protein by tmRNA-B1 lead to cell death.
We also created a second barnase selection in which stalling occurred at Pro-opal rather than Arg-opal. We anticipated that such a selection might identify mutants that inhibit stalling, perhaps by suppressing the inefficient termination at C-terminal Pro residues (8). Like the bar-R selection described above, the bar-P selection relies on stalling at a truncated barnase gene and rescue by a modified tmRNA to complete barnase and kill the cell (Fig. (Fig.1B).1B). We truncated barnase by 18 amino acids by mutating Ser92 to Pro and introducing an opal stop codon at codon 93. The tmRNA template was altered to encode the missing barnase residues DWLIYKTTDHYQTFTKIR (referred to here as tmRNA-B2). Two mutations occurred in the final barnase protein: the Ser92Pro mutation and insertion of Ala between residues 92 and 93. As these changes occurred in a surface-exposed loop between two β-strands (20), this scar was not anticipated to affect barnase function. Analysis of cells containing bar-P and tmRNA-B2 confirmed the toxicity upon induction with arabinose, while bar-P alone was nontoxic. Fewer than 1 in 106 cells survived the bar-P selection on arabinose (Fig. (Fig.1B1B).
The isolation of ribosome mutants that prevent tagging requires that each cell contain only one rRNA operon, since wild-type ribosomes would tag barnase and kill the cell. Squires and coworkers generated E. coli strains in which each of the seven rRNA operons was genetically deleted (SQ171 recA mutant) (3; S. Quan and C. Squires, personal communication). We introduced two plasmids into the SQ171 strain to adapt it to our selection. ptRNA-BT1 encodes tRNAs missing from the genome, as well as the selection genes bar-R and tmRNA-B1. The second plasmid, pTS-rrnC, expresses the rrnC rRNA operon and has a temperature-sensitive pSC101 origin. This plasmid is rapidly lost at 37°C if the cells have another source of rRNA, allowing us to exchange it for plasmids encoding mutant rRNA operons. This approach ensured that mutant ribosomes functioned efficiently at all stages of translation, because they were solely responsible for protein synthesis. In effect, this was a built-in positive selection for ribosome function added to the negative selection against stalling and tagging.
To isolate mutants that inhibit tagging at Arg-opal, we screened a library of 104 to 105 mutant 16S rRNA genes (38) in the SQ171 ptRNA-BT1 selection strain. The small-subunit RNA was chosen because we were initially interested in how tmRNA and SmpB interact with the decoding center upon entering the A site of stalled ribosomes. After three rounds of selection and enrichment, nearly all of the transformed cells survived on selective media. Sequencing of the 16S rRNA in surviving colonies revealed three clones containing one mutation each: the point mutant A1150G or deletion of A1150 or U1123. It was impossible to tell which nucleotide in 1150 to 1152 (AAA) was deleted; the same was true of 1121 to 1123 (UUU). Since these two sequences pair with each other, we believe that all three mutations have the same structural consequence (see Discussion). The activities of all three clones were confirmed by retransformation and testing of single colonies in barnase selection. Serial dilutions demonstrated the robust survival of the representative A1150Δ mutant upon induction of barnase with arabinose (Fig. (Fig.2).2). Growth curves revealed that the A1150Δ mutation caused a mild growth defect (a doubling time of 65 min compared to the wild-type 59 min), consistent with a general defect in translation.
We also isolated rRNA mutations that inhibited tagging at Pro-opal. Initially, we hoped to find rRNA mutants that would restore efficient termination. We therefore chose the 23S rRNA for mutagenesis because it contained the peptidyl-transferase center, the site where the peptide is hydrolyzed from tRNA during the termination reaction. We first tested 23S mutants in the bar-P selection that were reported in the literature to inhibit stalling on the peptides SecM and TnaC. These mutants included A2058G, U2609C, U754A, and an insertion of an adenosine after nucleotide 751 (4, 25). None of them showed an increase in survival compared to wild-type rRNA (data not shown). We therefore set out to identify mutants from random rRNA libraries. A library of 104 to 105 23S mutants (38) was introduced into the SQ171 ptRNA-BT2 selection strain. After two rounds of enrichment, nearly all the transformants on the selection plates survived. Sequencing of the 23S rRNA genes revealed two clones, one with the single C889U mutation and the other with two mutations, C889U and U846C. The activities of these mutants were verified by retransformation and testing of single colonies in the barnase selection (Fig. (Fig.22 shows C889U).
The C889U and U846C mutations are found in helix 38, known as the A-site finger (ASF) (32). This helix forms intersubunit bridge B1a, crossing over the A site and contacting the S13 protein in the small subunit (39). The C889U mutation is found at the tip of the finger in the loop that contacts S13. We thought that perhaps this mutation exerted its effect by disrupting bridge B1a. We tested this hypothesis by performing the bar-P selection on cells containing a 22-nucleotide (nt) deletion in helix 38 (H38Δ22), effectively destroying the B1a interaction (16). No increase in survival was observed, indicating that disruption of the B1a interaction is insufficient to prevent tagging (Fig. (Fig.22).
Previous studies had shown a slight growth defect from truncating helix 38 and disrupting B1a (16, 19, 31). We measured growth rates for SQ171 cells expressing either wild-type or C889U mutant rRNA and found them to be identical (with a 59-min doubling time). Since there is only one rRNA operon in these cells, the mutant ribosomes must be capable of performing essential functions in translation as well as wild-type rRNAs.
To confirm that the selected rRNA mutants rescued cells in the barnase selection by reducing the amount of tagged protein produced, we performed immunoblots to measure tagging levels directly. To test the 16S A1150Δ mutant isolated in the bar-R selection, we expressed the full-length MBP with a rare Arg-opal sequence at the C terminus to induce stalling. We altered the tmRNA template to encode the sequence ANDHHHHHHD so that tagging could be detected by anti-His6 antibodies (22). These changes in the tag also prevented rapid proteolysis of the tagged product (29). Analysis of the Arg-opal construct in SQ171 revealed that tagging was reduced nearly 3-fold in the mutant strain compared to the wild type (Fig. (Fig.3,3, lanes 5 and 6).
Similarly, we tested whether our selected 23S mutants inhibited tagging by using an MBP construct ending in Pro-opal. Using anti-His6 antibodies to visualize the addition of the tmRNA tag, we saw a 4-fold decrease in tagging in the C889U and double-mutant strains compared to the wild type (Fig. (Fig.3,3, lanes 1 to 3). No reduction of tagging was seen in the H38Δ22 mutant (lane 4), as predicted by its inability to survive the bar-P selection. These immunoblot data showed that both the A1150Δ and C889U mutations reduced tagging levels at their stalling sequences, Arg-opal and Pro-opal, respectively, confirming the genetic results described above.
We also tested the mutants at other stalling sequences to determine which step in the stalling and tagging process was defective. We initially anticipated that the bar-P selection might yield 23S mutants that restored efficient termination at the sequence Pro-opal. If the C889U mutant acted in this manner, tagging should be inhibited only at Pro-opal and not when ribosomes were stalled for another reason, such as a string of rare Arg codons. We tested this hypothesis by inducing stalling at the C terminus of MBP with eight rare Arg codons (AGG) and measuring tagging in cells containing either wild-type or mutant rRNA. Tagging was dramatically reduced in the C889U and C889U/U846C mutants compared to the wild-type and H38Δ22 rRNAs (Fig. (Fig.3).3). Since the inhibition of tagging is not specific to Pro-opal, it seemed unlikely that the C889U mutant acts by increasing termination efficiency, which presumably is not relevant to stalling and tagging on the eight-Arg sequence. We were unable to test the 16S A1150Δ mutant with the eight-Arg MBP construct, as its growth was severely inhibited. This increased sensitivity to Arg-tRNA depletion and ribosome stalling was consistent with a defect in the rescue process in the A1150Δ strain.
Since the C889U mutation inhibits tagging at both Pro-opal and a string of rare Arg codons—unrelated stalling sequences—it could be that it globally inhibits all tmRNA function. This would occur if the mutation prevents the ribosome from interacting productively with tmRNA or its associated protein, SmpB. To test this possibility, we measured tagging levels on constructs that stalled at a nonstop mRNA. The trpA transcriptional terminator was cloned following the full-length MBP gene, creating an mRNA of defined length that lacked a stop codon. Analyzing tagging at nonstop mRNAs allowed us to study the trans-translation process alone, apart from stalling and other upstream events.
Analysis with anti-His6 antibodies showed no loss of tagging in either the C889U or the C889U/U846C mutant on the nonstop construct (Fig. (Fig.3).3). In contrast, tagging in the A1150Δ mutant was decreased roughly 3-fold versus the wild-type strain. This level of reduction is similar to that seen with Arg-opal. We concluded that the A1150Δ mutant in 16S inhibits tmRNA function directly, while the selected 23S mutations do not inhibit the tagging process itself but some upstream step.
The C889U mutation reduces tagging at two different stalling sequences (Pro-opal and eight Arg) but does not inhibit tmRNA activity itself on a nonstop mRNA. One explanation of these results could be that it induces higher levels of recoding events. Stalled ribosomes can shift reading frame or read through stop codons, incorporating a suppressor tRNA. High levels of +1 frameshifting have been reported at the Pro-opal sequence CCC-UGA (6, 26). Recoding is an attractive explanation because it is downstream of stalling (i.e., slower than efficient termination) and upstream of tagging. Furthermore, deletion of the A-site finger has been reported to cause an increase in +1 frameshifting levels (16).
To test this hypothesis, we measured readthrough and frameshifting levels for ribosomes stalled at Pro-opal on the full-length GST protein. A sequence encoding the FLAG epitope was cloned downstream of the stop codon so that the GST-FLAG fusion protein was synthesized if and only if a recoding event occurred. Three constructs were created: one tested stop codon readthrough, another tested +1 frameshifting, and the last tested −1 frameshifting (Fig. (Fig.4A).4A). Synthesis of the GST-FLAG fusion was detected by anti-FLAG antibodies on protein extracted from SQ171 cells expressing either wild-type or C889U mutant rRNAs.
As expected, the sequence CCC-UGA led to very high levels of +1 frameshifting; −1 frameshifting was at a far lower level, and readthrough was barely detectable when analyzed on the same blot at the same intensity (Fig. (Fig.4B,4B, bottom). When each recoding event was analyzed separately, we clearly saw no significant changes in the level of readthrough or in the levels of +1 or −1 frameshifting in the C889U mutant strain. These results showed that the reduction in tagging in the C889U mutant strain is not a result of an increase in recoding events.
To study ribosome stalling and rescue in E. coli, we created genetic selections that tie tmRNA activity to cell death. Stalling on a truncated barnase gene leads to rescue by a modified tmRNA and completion of the toxic protein. Stalling in the bar-R selection is induced by a rare Arg codon, while stalling in the bar-P selection is induced by inefficient termination at Pro-opal. These selections are powerful tools for identifying mutations in the translation machinery that prevent tmRNA function. In principle, they could also be used to identify small molecules that inhibit trans-translation. This would be of interest, because the reduction of tmRNA activity sensitizes bacteria to antibiotics that target the ribosome (1).
We identified three separate 16S rRNA mutants that survived the bar-R selection: A1150G and deletion of either A1150 or U1123. The A1150 mutation inhibits tmRNA function, not only at the Arg-opal sequence at which it was selected, but also on a nonstop mRNA. We concluded that this mutation interferes with tmRNA function directly, inhibiting the trans-translation process itself rather than an upstream step. All three likely work via the same mechanism; they map to a single site within helix 39 of 16S rRNA (Fig. (Fig.5,5, right). Despite their distance in the primary sequence, A1150 base pairs with U1123. These mutations are expected to destabilize pairing in the first stem of helix 39 near this base pair. Helix 39 forms a coaxial stack with helices 38, 36, and 35, stretching the whole length of the head of the 30S subunit (37). The S9 and S10 proteins bind helix 39; replacement of the A1152 phosphate with phosphorthioate inhibits 70S ribosome assembly (7), presumably because it interferes with proper S10 binding. It seems possible, then, that these mutations alter the conformation or dynamics of the head of the 30S subunit.
We likewise isolated two 23S rRNA clones, C889U alone or in combination with U846C, that survive the bar-P selection and reduce tagging at Pro-opal. As shown in Fig. Fig.5,5, these mutations are found in helix 38, the ASF, so called because it contacts the A-site-bound tRNA (27, 32). The ASF interacts with the S13 protein in the small subunit to form intersubunit bridge B1a (39). The B1 bridges (B1a, -b, and -c) are the sole links between the head substructure in the 30S subunit and the 50S subunit. During the translation process, the B1a interaction is broken and the ASF changes its binding partner from S13 to S19 (34). This motion is part of the ratchet-like rotation of the head that occurs during translocation upon EF-G binding (34). The fact that the ASF binds A-site tRNA and plays a role in ribosome conformational dynamics may explain the role of C889U in ribosome stalling and rescue.
The C889U mutation lies immediately next to the nucleotides 886 to 888 in the loop of the ASF, which interact with S13 (39). We hypothesized that the C889U mutation might exert its effects by disrupting the B1a interaction. In support of this idea, deletion of 22 nt at the tip of helix 38 was reported to reduce the energetic barrier to translocation (16). We found, however, that this H38Δ22 mutant did not lead to survival in the barnase assay, nor did it lower tagging levels in Pro-opal or at a string of rare Arg codons in the immunoblot assay. If changes in B1a are responsible for the reduction in tagging, it must be more subtle than mere disruption, perhaps favoring one ratchet rotation state over the other (S13/S19). All in all, it seems that B1a plays a minor role in the normal translation process, since ASF truncations and S13 mutations have only minor effects (5, 16, 19, 31).
The ASF has a kink-turn motif near its base that is predicted to play a role in its flexibility and motion (28). As the U846C mutation lies within this kink-turn motif, it is tempting to speculate that it affects the structural dynamics of the ASF. The U846C mutant was identified only together with C889U, however, and the single and double mutants did not show any detectable difference in activity in either the immunoblot assays or barnase selection. It is therefore unclear if this mutation conveys any additional advantage.
How does C889U reduce tagging? It is unlikely that C889U restores high-efficiency termination, thereby reducing stalling. Two findings support this conclusion. First, although the C889U mutant was selected for its ability to inhibit tagging at Pro-opal, immunoblots showed that it also reduced tagging levels at a cluster of eight rare Arg codons. While we have not directly ruled out the possibility that stalling is reduced at Pro-opal, it seems more likely that a step downstream of stalling is inhibited, explaining both the Pro-opal and eight-Arg results. Second, if termination rates increased, the levels of readthrough and frameshifting should decrease in the mutant strain, since these recoding events occur because of inefficient termination. We found, however, that C889U did not alter readthrough or frameshifting levels.
In contrast to the A1150Δ case, the C889U mutant did not inhibit trans-translation itself. Since neither stalling nor tmRNA function appears to be inhibited in the C889U mutant, it seems that some intermediate step between the two must be affected. One possibility is that the mutation may affect processing of the stalled mRNA. Before tmRNA and SmpB can enter the A site of stalled ribosomes, the downstream mRNA must be removed, either by an A-site endonuclease or by the action of 3′-5′ exonucleases (10). mRNA sequences longer than 12 nt downstream of the stalled P site prevent rapid release of ribosomes by tmRNA (12). An effect on mRNA processing is an appealing explanation, because tagging is reduced only when stalling occurs in the middle, not at the 3′ end, of an mRNA. We are currently characterizing the molecular mechanism of action of the C889U and A1150Δ mutants. Determining the effects of these mutations will shed light on how the ribosome interacts with tmRNA and SmpB to bring about the rescue of stalled ribosomes.
We thank Cathy Squires for the SQ171 strain and Alexander Mankin for the rRNA libraries.
This work was supported by grant GM77633 to A.B. from the National Institutes of Health.
Published ahead of print on 6 November 2009.