New article: Evolution of UGA codon translation in ciliates
A new paper from my former lab is out as a reviewed preprint in eLife. It’s a short manuscript written in response to a study from another lab on stop codon reassignment.
Background - Alternative genetic codes
The genetic code, which specifies which codons in mRNA are translated into which amino acids in proteins, is “universal” in the sense that all cellular life uses a genetic code, but these codes are not all identical. Many organisms use alternative codes, that typically differ in one or a handful of codons that code for different amino acids compared to the “standard” code (egotistically, the “standard code” is the one used by us humans and other vertebrate animals).
The most intriguing alternative codes are the ones where stop codons are repurposed to code for amino acids. If a stop codon is reassigned to code for an amino acid, you’d expect to find a tRNA with the complementary anticodon. However, organisms that recode the stop codon UGA to tryptophan were especially puzzling because the corresponding tRNA was not found in their genomes. UGA->Tryptophan has evolved at least four times. All known cases are in microbial eukaryotes, including twice in the ciliates, which are the group that my former lab’s research was focussed on.
What the earlier study claims
A study published just over a year ago found the likely answer to the question of the missing tRNA: instead of a typical tRNA with a 5 base-pair anticodon stem structure and the expected anticodon UCA, they use a tRNA with a 4 base-pair stem and the anticodon CCA. They showed that this tRNA is found in two different organisms, a tryposomatid species with the apt moniker Blastocrithidia nonstop, and a ciliate Condylostoma magnum. They engineered 4-bp variant tRNAs in other model species and showed that the variants increased the rate of translational “readthrough” of the erstwhile stop codons, evidence that they are indeed binding to UGAs to translate them.
They also reported that a specific point mutation in the protein eRF1 (eukaryotic release factor 1), a key player in translation termination, is associated with UGA->Tryptophan. eRF1 is involved in recognizing the stop codon and releasing the nascent polypeptide chain, as its full name implies. They argued that this mutation was also necessary for UGA reassignment, by reducing the ability of eRF1 to recognize UGA. If so, it would no longer compete as much with the tRNA for UGAs, facilitating their translation.
What we found by looking at more species
It’s the second claim about the mutant eRF1 that our article responds to, because it doesn’t hold up if we look at more species. We dug into the genomic data for different species of ciliates, taking advantage of how incredibly diverse they are in the genetic codes that they use. The data covered species using a number of different codes, some with the UGA->Tryptophan reassignment, others without. However, the mutation in eRF1 was not correlated at all with the UGA reassignment: some had the mutation, others didn’t. In contrast, we found the 4-bp stem variant tRNA in all species with UGA reassignment, lending support to their first point.
So why does this matter? The authors of the earlier study hypothesized that the mutation was mechanistically relevant, that eRF1 was “giving way” to tRNAs, so to speak. We think instead that this mutation is not necessary for this trait, and indeed there are signs in the genomes of some ciliates that there is still competition between tRNA and eRF1 for binding with UGAs, so alternative models for the evolution of this stop codon reassignment are still not ruled out.
Thoughts on the eLife publishing model
This is our first time publishing in eLife, which was founded as a nonprofit life sciences publisher in 2012 by three major scientific organizations. Since then it has been a platform for trying out different models of peer review and scientific publishing. At the end of 2022 they switched away from the traditional “accept/reject” model of publication. All manuscripts submitted have to be already publicly accessible as “preprints”. Editors still decide whether or not to send a manuscript out for review, but if sent for peer review, they will be published alongside their reviews, as “reviewed preprints”. Authors can choose to revise and resubmit a final “version of record”. As a comparison, in most journals, crossing the editor’s desk is the first hurdle, but if reviews are negative (or worse - one positive and one negative…) they can still be rejected for publication.
I think this is a step forward. Yes, editors still act as gatekeepers, but they are there in the traditional model too. The advance here is to save on the time and uncompensated labor when a paper is sent for review but rejected and is then shopped around to other journals, where the process is repeated until it’s either accepted or the authors give up. Each round represents time invested by authors in submitting, waiting and responding to reviews, and also by peer reviewers and editors.
I can see how there would be more pressure on editors to be discerning in what they send out for review. Hypothetically, in the traditional model, an editor could indiscriminately send everything for review, regardless of quality, and just wait for the reviewers to filter out the bad ones. That’s of course a huge waste of everyone’s time, but no one would ever know because the rejected papers never see the light of day, at least in that editor’s journal. In the eLife model, the reviewed preprints are posted alongside the reviews, with a short assessment of their significance and quality of evidence. Editors’ names are publicly linked to the papers, so one could with time see trends in how they perform.
This is not our first time publishing under a “post-publication review” model. In 2022 we published a paper (incidentally also on genetic codes) in Peer Community Journal, part of the larger Peer Community In initiative, which has a very similar approach. They only review publicly available preprints, editors choose which submitted papers will be sent for review, and the reviews are published alongside the manuscript. One point that may be different in their process is that if reviews are negative, authors can decide not to proceed further, and these will then not be published. I’m not sure if eLife has this option, too.
More significantly, PCI is “diamond open access” - free for readers and also free for authors to publish there. eLife charges a fee of USD 2000 (previously USD 3000), although waivers are available. This is on the lower end compared to most conventional journals. Prestige journals like Nature and Science don’t charge publication fees but are subscription only (not open access), but their “companion” journals charge high fees for open access: USD 6790 at Nature Communications, USD 2590 at Scientific Reports. Anything that pushes against this inflationary trend is worth looking into.