Evidence for loss and reacquisition of alcoholic fermentation in a fructophilic yeast lineage

Essential revisions:

The reviewers identified two significant matters requiring attention.

1) The title and the Abstract oversell the paper: this over-promising leaves the reader disappointed, rather than appreciative of the work. For example, the Abstract states that "genes required for alcoholic fermentation were lost and subsequently re-acquired from bacteria through horizontal gene transfer" and later indicates that the fermentation pathway was "reinstated", implying that it was lost initially. However, the order of events cannot be established from the results presented. The authors make this clear in the Discussion, writing that "the events leading to the observed profound remodeling of the alcoholic fermentation pathway in W/S-clade yeasts are impossible to trace with certainty because of their antiquity". The evidence required to establish this order – such as a genome without any version ADH1 – is not currently available. This is a shame, since had the authors been able to establish the order of events, the work would be of exceptional interest.

We agree that in the previous version there was a discrepancy between the emphasis given to the “loss before acquisition hypothesis” in the title and Abstract when compared with the Discussion. This pertinent criticism spurred us to review all aspects of the text. It also encouraged us to inspect genomic data concerning species close to the W/S clade generated very recently in the context of the Y1000+ project (https://y1000plus.wei.wisc.edu), and we were fortunate to find a species, Candida galacta that lacks any version of ADH1. As stated in the review report, this is very important evidence that further strengthens the “loss before acquisition” hypothesis. The finding of this genome was therefore included in the revised version and implied the inclusion of new authors.

In addition, we constructed a new St. bombicola mutant in which the ARO10 gene was deleted. The inability of this mutant to produce ethanol provided experimental confirmation that PDC1 was lost and that ARO10 currently fulfills its role in alcoholic fermentation. Since our data also suggests that ADH1 and PDC1 were lost in quick succession, because no genome was found possessing either only “native” ADH1 or only “native” PDC1 alone, the case for an ancestral loss event is presently soundly supported.

The statement in the previous version quoted in the review report was modified to convey the message that the reconstruction of these kind of ancient events always carries some uncertainty, which was the original intention.

Additionally, more evidence would be required to support the conclusion that the observed changes constitute adaptations to a sugar-rich environment as opposed to some other shared feature of the environments occupied by W/S-clade yeasts and not by their close relatives.

We changed the title and Abstract accordingly. In the remainder of the text we do pinpoint the results consistent with adaptation to a sugar-rich environment, which we still consider to be the most likely driver of the unusual evolutionary events reported. The strong and clear phenotype observed for the aro10 mutant allowed us to put forward an hypothesis stating how the various events may have contributed to improve sugar metabolism, which is now presented at the end of the Discussion.

2) The narrative. The description of the results is difficult to follow and disorderly (different data sets, different comparisons, and with little guidance to the reader as to where the investigation is going). It would be helpful to lay out from the outset a plausible scenario for the evolution of fructophilic yeast and guide the reader so that s/he engages with the central hypothesis. Next, focus on lateral gene transfer of abilities associated with acquisition of fructose degradation. Report the candidates. Report the evolutionary analyses and the experimental data. Finally, having dealt with gain of new function, it would be appropriate to ask about the ancestral function. Avoid presenting the Results in chronological order: present the strongest results first followed by the secondary findings.

We thoroughly remodeled data presentation, in particular in the beginning of the manuscript. The order in which the results are presented and the titles within the Results section were also changed. In the revised version we used as a thread the comparison with fructophilic bacteria, to which we return throughout the entire manuscript, including a discussion on differences and similarities between the fructophilic bacteria and fructophilic yeasts.