The β-hairpin of 40S exit channel protein Rps5/uS7 promotes efficient and accurate translation initiation in vivo

Reviewer #1:

[…] The manuscript is well-written and organized but may be improved overall by considering where additional transitional sentences could help readers, particularly when introducing more detailed aspects of initiation, and by breaking complex sentences into separate sentences.

Specific comments are as follows:

Abstract: Include the idea of the relative importance of Rps5 relative to eIFs at the conclusion of the Abstract?

We have included the relative importance of Rps5 in the Abstract and the sentence has now been changed to: “We conclude that the Rps5 β-hairpin is as crucial as soluble initiation factors for efficient and accurate start codon recognition.”

Introduction: The GTP bound to eIF2 […] the transition to this may be too abrupt in the first paragraph of the Introduction. Possibly give a more general overview into start site selection (as also occurs in subsequent paragraphs) and put this later in the Introduction?

To make the transition smoother we have changed the sentence as below:

“During scanning, the GTP bound to eIF2 in the TC is hydrolyzed in the 43S PIC in a manner dependent on the GTPase activating protein eIF5, but P_i release is blocked by eIF1, which also impedes stable binding of Met-tRNA_i in the P site.”

Introduction, first paragraph: “AUG recognition triggers dissociation”. Use instead “start codon recognition triggers dissociation”.

We have replaced AUG with start codon and the sentence now reads as “Start codon recognition triggers dissociation of eIF1”.

Introduction, start of second paragraph: Define this TC-binding as loading so that the reference to loading is explicit and not inferred?

We modified the sentence to: “…to which TC rapidly loads, bound in a state capable of inspecting successive triplets entering the P site…”

In the second paragraph of the subsection headed “Substitutions E144R and R225K impair translation initiation and start codon selection in vivo”: S223A mutation is also 2-fold in Supplementary file 1.

We wanted readers to note that E144R and 3 different substitutions of Arg255 are the only mutations that increased the UUG:AUG ratio substantially, which we defined as greater than two-fold, and the ratio for S223A does not exceed 2-fold.

In the same paragraph: I had trouble following the reasoning that the salt bridge might still be important although reinstating it had little impact on phenotype because the nature of the residues might be more important. Simply say that at this point that there is not direct evidence for the importance of the salt bridge?

We reasoned that combining the two mutations E144R and R225E both of which conferred >2-fold increases in the UUG:AUG ratio, should give an additive effect and increase the UUG:AUG ratio more than either of the single mutants. Since the double mutant did not increase the UUG:AUG ratio more than E144R alone, we inferred that reinstating the salt bridge mitigates to some degree the effects of the E144R single mutant. However, in deference to the referee, we added the sentence: “Additional experiments are needed to establish the importance of the salt bridge for Rps5 function” (Results).

Still in the subsection “Substitutions E144R and R225K impair translation initiation and start codon selection in vivo”: Comment on lacZ reporter ratios or levels for these other mutants that enable UUG initiation from his4-301 that confer Sui^- phenotypes? Then readers could get a sense of comparison for Figure 3B (also relevant for the second paragraph of the subsection headed “E144R and R225K elevate UUG initiation indirectly by exacerbating the effect of poor context of the SUI1 start codon and thereby reducing eIF1 abundance”).

We added the sentence: “For example, the eIF1 mutations sui1-K37A and sui1-R33A increase the UUG:AUG ratio by 4.8- and 7.7-fold, but only the latter suppresses the His^- phenotype of his4-301” (Results).

In the second paragraph of the subsection headed “E144R and R225K elevate UUG initiation indirectly by exacerbating the effect of poor context of the SUI1 start codon and thereby reducing eIF1 abundance”: Is there a Slg^- phenotype for R225K in Figure 4–figure supplement 1? Also, in the lower panel, was the Y-axis of the image accidentally stretched (the colonies all look more oval than round).

R225K has a Slg^- phenotype here when compared to the WT in row 1. To clarify the different extents of suppression by scSUI1, we added the sentence: “Introducing scSUI1 also mitigated their Slg^- phenotypes, slightly for -R225K and substantially for -E144R”. The image has been adjusted to its original state in which the cell spottings and individual colonies appear round.

In the last paragraph of the subsection headed “E144R and R225K confer leaky scanning of an upstream AUG start codon”: The authors appreciate that the effect on initiation of the elongated uORF1 UAUG1 relative to initiation at the GCN4 AUG is complicated to interpret. Perhaps simply note that mechanistic possibilities are considered in the discussion and also move some of this material to the Discussion where these concerns are addressed?

We thank the reviewer for this suggestion and have removed this paragraph from the Results and integrated most of it into the Discussion.

In the last paragraph of the subsection headed “Ssu^- substitution E144R destabilizes the PIN conformation of the 48S PIC in vitro”: It is a minor point, but the connections between Figure 6–figure supplement 1 and Figure 4–figure supplement 2 due to possible specific effects of genotype changes is left open (specifically how GCN4-lacZ levels differ with the RPS5 alleles with/without increased eIF1 dose is not clear).

This is a good point. We added the following sentence: “Derepression of GCN4-lacZ by E144R was evident even in the presence of lcSUI1 (Figure 4–figure supplement 2, E144R/scSUI1 vs. WT, unstarved), indicating that it does not result solely from the reduced eIF1 abundance in this mutant.”

In the first paragraph of the subsection headed “Substitutions in the loop region of the Rps5 β-hairpin increase fidelity of start codon 19 selection independently of context nucleotides”: R157E did not seem to have as substantial an effect (or I missed something).

We agree that R157E did not significantly reduce the UUG:AUG ratio and have removed it from the text. The sentence now reads “Furthermore, R148A/E, R156A/E and R157A suppressed the elevated UUG:AUG ratio”.

Discussion, second paragraph: Impossible to increase or impossible to detect an increase?

We have changed the phrase to “it is impossible to detect an increase in leaky scanning”.

Discussion, end of fourth paragraph: Are there possible ramifications for the tRNA anticodon's interactions with the start codon? This subject is discussed later.

As the reviewer notes, the ramifications for the tRNA anticodon’s interaction with the start codon are discussed later in the Discussion.

Figure 1: Scanning inhibitor SI element in eIF1A NTT is not specifically indicated in the figure (it is in Figure 9).

Figure 1 now has the SI element in eIF1A indicated.

Figure 2D: State in legend that region of gradient where polysomes would have been is not shown (I assume they are not there)?

The experimental conditions used in Figure 3D dissociate polysomes and 80S monosomes and allow separation of the ribosomal subunits. The legend now reads: “Similar to (C) but the cultures were not treated with cycloheximide and lysed in buffers without MgCl₂ to allow separation of the dissociated ribosomal subunits.”

Figure 4–figure supplement 1: As mentioned above, R225K data look stretched for Y axis; if this is the case, then fix.

Addressed above.

Reviewer #2:

[…] My major concern is related to the results describing effects of mutations in uS7 on initiation from near-cognate UUG start codon. Because initiator Met-tRNAMet is expected to form less than three standard base-pairs with this codon, an initiation complex assembled on mRNA with UUG, a priori, is less stable than the complex formed on the AUG codon with complete three base-pairs. Consequently, the results describing, for example, dissociation of the initiation complexes assembled on the UUG codon can be exaggerated because inherently unstable complexes additionally fall apart owing to the experimental procedure (viz. electrophoresis during 45 min). Hence, Figure 6C shows that the initiation complex assembled on the UUG codon with the wild type 40S is very weak compared to the stable complex on the AUG codon. In this regard, it would be more correct to describe relative effects rather than provide exact values of constants. Since the Y-axis of most of the graphs is presented in arbitrary units, it would be also very helpful to show actual dmp/cpm values to demonstrate efficiency of signals and their differences.

The in vitro reconstituted experiments in this study are based on the thorough analysis from the Lorsch lab described in Kolitz et al. (2009), using single and double substitutions in the mRNA start codon to investigate the effects of base-pair mismatches with initiator tRNA on the rates of TC association and dissociation from 43S·mRNA complexes. They found that when UUG replaces AUG in the mRNA, the rate constant for TC binding (k_on) was reduced by ∼40%, while the dissociation rate constant (k_off) was increased ∼4-fold, to yield an increase in K_d of ∼7-fold. Thus, the referee is correct in stating that the UUG complex is inherently less stable than the AUG complex; however, this difference is the key feature of this assay that allows us to discern relatively greater effects of Ssu^- mutations on UUG versus AUG complexes. Based on the measured K_d values in Kolitz et al. (2009), UUG is the near-cognate closest to AUG in forming a stable PIC. With WT ribosomes, the UUG complexes are still quite stable, with <20% dissociating in the first hour of incubation; and even less dissociation occurs when the ß-S264Y variant of eIF2 is employed, as done in Figure 8D where the destabilizing effect of R148E on UUG complexes was established. Note also that the key conclusion that E144R destabilizes the PIC derives in large part from its effect of increasing the k_off for the highly stable AUG complex (Figure 6C). Dissociation of the UUG complexes during electrophoresis is not an issue because this would occur equally for all of the time points in the assay, including the zero-time point from which dissociation during the incubation is measured as a function of time. This assay has been used in several previous papers that we published in collaboration with the Lorsch lab, including the analysis of Sui^- and Ssu^- mutations in initiator tRNA (Dong et al., 2014), and of eIF1 Ssu^- mutations by Martin-Marcos et al. (2014). Hence, we believe that the comparable analyses of rps5 Ssu^- mutations presented here is both appropriate and accurate.

I would also like to add the following suggestions:

1) It is not explained in the figures why eIF3 is not depicted in the schemes (Figures 1, 9) and why in Figure 2B the authors provide a picture of the 48S initiation complex with only a few components while normal 48S contains much more initiation factors bound to the 40S subunit?

We did not include eIF3 in the schematic models of Figures 1 and 9 for several reasons. First, eIF3 is a large, multisubunit complex and accurately displaying even a simplified PIC containing only eIFs 1, 1A, TC and mRNA is already challenging. Second, the binding sites in the PIC for all of eIF3’s subunits are not yet known, and adding the well-established binding site for the PCI domains of eIF3α and -c on the solvent-exposed side of the 40S subunit would add little to visualization of the subunit interface surface depicted in these schematics. Third, the functions of the eIF3 subunits in scanning and AUG selection are not well established and Figures 1 and 9 are used to facilitate descriptions of the known functions of eIFs. Figure 2B (now Figure 2A) is the partial yeast 48S structure from Hussain et al. (2014) where densities visible at high resolution were obtained only for the 40S, eIF1, eIF1A, mRNA, tRNA_i and eIF2α; nevertheless, it still represents the most complete PIC at reasonably high resolution currently available. We have altered Figure 2A to show the gamma subunit of eIF2 to indicate better the orientation of TC in the structure. The Figure 2A-B legend was modified to include: “Depiction of the partial yeast 48S PIC (PDB 3J81) showing Rps5 (gold), mRNA (orange), Met-tRNA_i (green), eIF2α (purple), eIF2γ (yellow), eIF1 (cyan) and eIF1A (blue). For clarity other ribosomal proteins, eIF2β and putative eIF5 densities are not shown.”

2) In Figure 2B I would recommend to use a more conventional orientation of 18S rRNA of the small ribosomal subunit, to clearly depict where exactly uS7 is located, where tRNAMet binds etc. I have the same problem with Figure 2C: normally mRNA is shown with its 5'-nucleotides at the left and P-site codon on the right. In general, figure legends do not contain all necessary information.

We appreciate the reviewer’s point. While it might be non-standard, we chose the view in Figure 2A because it illustrates most clearly the location of uS7 in the mRNA exit channel, the residues identified in this study that affect initiation, and the position of the uS7 hairpin relative to the mRNA nucleotides and eIF2α-D1, which are all depicted similarly in Figure 2B after zooming in on the exit channel. We note that Figure 1 of Devaraj et al., on the effects of S7 hairpin mutations on frameshifting in E. coli, depicts exactly the same view of the bacterial components of the exit channel.

3) The Discussion part is rather narrowed to a specific role of uS7 in eukaryotic initiation. It would be useful to draw some parallels with the role of the same protein in bacterial protein synthesis, whose conserved loop shares the same location relatively to the E- and P-site elements.

We agree with the reviewer and have included the following sentences in the final paragraph: “The β-hairpin of uS7 also protrudes into the mRNA exit channel of bacterial ribosomes in position to interact with mRNA residues just upstream from the P site codon (Jenner et al. 2007). In bacterial elongation complexes, the hairpin is also in proximity to E-site tRNA, and truncation of the hairpin increases the frequency of frameshifting most likely by allowing premature dissociation of the E-site tRNA (Devaraj et al. 2009). Interestingly, in the yeast py48S PIC, eIF2α-D1 occupies the position of E-site tRNA (Hussain et al, 2014), in accordance with our suggestion that altering the β-hairpin of yeast uS7/Rps5 impairs start codon selection by altering the position or flexibility of eIF2α-D1.”

Reviewer #3:

[…] The authors used binding studies to examine the affinity of the ternary complex bound to the Wild-type and rps5 mutant 40S ribosomes. They show that the mutant rps5 destabilizes the PIN/closed state. Interestingly, the transition from POUT to PIN is not affected, however, the stability of the PIN complexes is decreased for the rps5 mutant ribosomes. Overall the findings are interesting and significant in that they suggest a novel role for rps5 in start codon recognition, but the manuscript was a bit difficult to understand and it would benefit from additional clarifications and simplifications. Please see the following comments.

It is unclear why the rps5 mutants promote leaking scanning for the GCN4 reporter, since the phenotype is that these mutations confer a hyperaccurate initiation phenotype one would assume that a start codon in a good sequence context would be as good as WT initiation or better (the structural argument provided is not clear nor satisfying as written). However, it does appear that while the rps5 mutants confer reduced initiation at start codons in a poor or week context that they may not enhance initiation in a favorable context. Could it be possible that a general decrease in frequency of start codon recognition could just have a much greater effect on the weaker or poor context AUGs and relatively less on the AUG in a good sequence context? Would this explain all the data, the binding studies showing reduced complex formation, the decrease in UUG initiation, and SUI1 initiation? As the manuscript reads it appears that there are more than one model/mechanism for how rps5 is affecting translation initiation depending on which reporter or assay is used. For example, when ribosomes with rps5 mutations don't recognize the GCN4 uAUG it is promoting leaky scanning, but when it doesn't recognize the SUI1 or UUG it is hyperaccurate.

We thank the referee for this suggestion. We agree with his/her interpretation and had actually laid the groundwork for advancing it as a likely scenario with the text in the second and third paragraphs of the Discussion that seeks to explain the increased leaky scanning observed for uAUG-1 in optimum context as the result of el.uORF1 being unstructured. However, at the end, we failed to propose this explanation explicitly. To rectify this, we added the following sentence: “Thus, it seems plausible that E144R and R225K decrease the efficiency of start codon recognition only for weak initiation sites, including near-cognate UUG codons and sub-optimal AUG start sites, without reducing recognition of AUG codons in strong context that initiate structured coding sequences.” And later in the Discussion we state: “As summarized in Figure 9, our results indicate that both E144 and R148 promote start codon selection by stabilizing the P_IN state, and the finding that E144R reduces initiation at both UUG and sub-optimal AUG codons, while R148E impairs only UUG recognition, can be explained as the result of a relatively stronger contribution of E144 versus R148 to the stability of the P_IN state”.

It is not clear how rps5 could be affecting initiation at UUG codons given is position, perhaps the authors could discuss this further in the Discussion.

We had presented various possibilities at the end of the Discussion of how mutations in the β-hairpin might affect start codon recognition. Briefly, we suggested a disruption of the interaction of hairpin loop residues with the rRNA residue in close proximity to the -3 context nucleotide, perturbation of the hairpin interaction with Rps16 whose CTT closely approaches the codon-anticodon duplex in the P-site, and altering the interaction with domain-1 of eIF2α that interacts with both tRNA_i and the mRNA 5’ of the start codon. We now suggest that this last mechanism would be analogous to the role of the orthologous hairpin in bacterial S7 in stabilizing the E-site tRNA during elongation, as eIF2α-D1 and E-site tRNA occupy very similar positions.

Is the UUG in good sequence context or could this also be contributing to its initiation? What does moderately strong? Is the -3 position optimal? Perhaps defining these sequences would go a long way towards answering these questions.

The sequence context for the HIS4-lacZ UUG reporter examined in vivo is strong, while that of the model mRNA used for in vitro experiments is weak. However, no effect of context has been observed thus far in the reconstituted system using the model mRNAs. Given the concordance of our findings in vivo with the HIS4-lacZ UUG reporter and those in vitro with the UUG model mRNA, we believe that context plays a minor role in suppression of UUG initiation by the Rps5 substitutions. Moreover, the context for the UUG and AUG HIS4-lacZ reporters are identical, and the same holds for the UUG and AUG model mRNAs, which is now stated explicitly in Results. The sequences of the model mRNAs used in the assays were mentioned in the Materials and methods, and we have now added the context of the HIS4-lacZ reporters to that section as well.

There seems to be some disconnect between the Introduction and the Results where the authors state that rps5 is required for efficient initiation. It is unclear where in the results the rate of initiation was measured. Perhaps the authors can better define what they mean by efficiency.

In Figure 3C, we show that the R225K and E144R mutants exhibit moderate, but statistically significant reductions in the polysome:monosome ratio, which indicates a reduction in the rate of initiation versus elongation. We have cited this result in the first paragraph of the Discussion as evidence that the E144R substitution in β-strand 1 of the hairpin reduces the rate of bulk translation initiation. Combining this result with the increased leaky scanning of GCN4 uAUG-1 conferred by this mutation leads us to conclude that selection of AUG codons by the scanning PIC is less efficient in this Rps5 mutant.

At the end of the second paragraph of the subsection headed “E144R and R225K elevate UUG initiation indirectly by exacerbating the effect of poor context of the SUI1 start codon and thereby reducing eIF1 abundance” and Figure 4. The authors conclude: “the increased recognition of UUG start codons conferred by the rps5 mutations is an indirect consequence of their reduced expression of eIF1.” They show that expressing eIF1 from a plasmid returns UUG/AUG recognition ratio back to WT and eIF1 levels increase (although quantification would be helpful as they appear not to come back to wt level). However, the authors assume but do not show whether the defect in sequence context is still present when eIF1 is expressed from a plasmid. Figure 4B should be repeated with yeast strains expressing eIF1 from a plasmid.

We have quantified the blots and added the results to Figure 4C, and the reviewer is correct that the levels of eIF1 are not fully restored to the WT level when an extra copy of SUI1 under its native promoter is introduced into the mutants on a single-copy plasmid (compare lanes 3-4 and 7-8 to 9-10). Despite this, the increase in eIF1 levels in the mutants achieved with scSUI1 is sufficient to restore the elevated UUG/AUG ratio to the WT level (Figure 4D). This demonstrates that elevated UUG initiation is an indirect consequence of low eIF1 levels in the mutants, which is the main point of this experiment. The fact that the eIF1 levels in the mutants harboring scSUI1 remain well below those seen in the corresponding WT/scSUI1 transformant (lanes 3-4 & 7-8 vs. 11-12) indicates that enhanced discrimination against the poor context of native SUI1 still occurs in the mutants at the boosted levels of eIF1. Similarly, we show later in the paper that the increased leaky scanning of GCN4 uAUG-1 in the mutants is not diminished by increasing eIF1 levels with scSUI1. Thus, we think it is clear that the defects in AUG recognition, whether for the native SUI1 AUG or GCN4 uAUG-1, do not arise from the reduced eIF1 levels in the mutants. In fact, because eIF1 promotes scanning, reduced eIF1 levels per se should diminish discrimination against the poor context of SU1I and boost eIF1 levels; and if anything, should reduce, not increase, leaky scanning of uAUG-1.