tRNA ligase structure reveals kinetic competition between non-conventional mRNA splicing and mRNA decay

Essential revisions:

Both reviewers agree that the paper presents an interesting study, which would warrant publication in eLife if amended in response to the reviewers' critiques. The Reviewing Editor concurs with the reviewer' opinions. As you can see in the comments below both reviewers believe that the key results are shown in Figure 3, but they argue that some issues related to these results should be experimentally addressed.

1) What is the basis of the reduction in splicing efficiency in the trl1-H148Y enzyme? Does the enzyme have a lower affinity for specific RNA fragments? An answer to this question might provide a link to the understanding of the enhanced sensitivity to Xrn1 or the exosome. This experiment could be readily done.

We agree with the reviewers’ assessment that a detailed mechanistic understanding of the reduced splicing (ligation) efficiency of the H148Y mutant will add to our understanding of the Trl1 ligation mechanism. We tested several different in vitro approaches to determine binding affinities for both exons. We measured affinities by monitoring fluorescence changes of labeled RNA oligonucleotides upon binding to scTrl1-LIG. Unexpectedly, these experiments yielded very similar affinities for WT and mutant Trl1-LIG towards 5’ and 3’ exon fragment (differing by only ~1.4 and ~1.1-fold, respectively), respectively). We included these fluorescence binding measurements as new Figure 4—figure supplement 1.

We conclude from this work that, if binding affinities are not significantly impeded, the reduced activity of the mutant enzyme must result from a changed catalytic activity. To confirm the reduced ligation kinetics (efficiency) of the Trl1-H148Y enzyme (Figure 4B), we performed time-resolved in vitro ligation experiments with labeled RNA oligonucleotides. These results confirmed the effect of the His-to-Tyr mutation, causing a ~2-fold reduced apparent rate constants. Interestingly, the mutant ligase is not only a 2x slower enzyme but also plateaued at a >5-fold lower amount of ligation product compared to the wild-type enzyme. These results are depicted as new Figures 4C (PAGE gel) and 4D (quantification and fitting). Taken together, these results strongly support our kinetic competition model.

2) What is the band marked with an asterisk in Figure 1D, lane 3?

The * demarks a circularized tandem HAC1 intron. We now added experimental evidence that this is indeed the case. To this end, we used antisense DNA oligos complementary to either the presumptive ligation (circularization) site or the middle of the intron. After annealing to the RNA products of the in vitro splicing reaction and treatment with RNase H, which specifically hydrolyzes the phosphodiester bonds of RNA in RNA/DNA hybrids, the band in question disappeared and a band at the exact position of the severed intron (252 nt) re-appeared. We conclude that * represents two cleaved introns that were circularized by two intermolecular ligation events at their 5’ and 3’ ends. The results are depicted in Figure 2—figure supplement 1. We added the RNase H assay to the Materials and methods.

3) The results in Figure 4 should be quantified.

We quantified the HAC1 splicing results and added the results to Figure 5 (previously Figure 4).

4) Discussion: Where is it shown that Xrn1 degrades the 3' exon?

We softened the statement by changing “Our data show…” to “Our data suggest…”. Based on our RT-PCR results (Figure 5D), we could show that levels of the 5’ exon are not affected by the presence or absence of Xrn1. Although we did not directly “show” the degradation of the 3’ exon by Xrn1, it is the only HAC1 mRNA component left in the non-conventional splicing reaction. Our RT-PCR data demonstrate that deletion of XRN1 restores splicing, which requires a fully intact 3’ exon as substrate.