Dear Dr. Hittinger
Thank you for submitting the manuscript "Gene duplication and the adaptive evolution of a classic genetic switch" for review by Nature. This manuscript reports the use of molecular and genetic tools to address the contribution of coding changes to the divergence of two paralogue genes. Duplication is a phenomenon that occurs in single genes, chromosomal regions and even entire genomes, and is one of the most important hallmarks in biological complexity. The most common consequence of genomic duplication is the loss of all or part of the duplicated sequences through delection or degeneration. On the other hand, another process can occur, although much less well understood: after duplication, one of the duplicated sequences or both might acquire new functions, since they are under weaker selective pressures and can undergo mutations that lead to divergent functions. One of the most exciting natural experiments in genome evolution is the duplication of whole genomes in Saccharomyces species. Comparative genomic analysis of several sequences from these species has revealed that new functions rarely seem to arise from changes in biochemical properties but are frequently associated with regulatory changes. This study enables the understanding that the analyzed pair of duplicated genes (Gal1 and Gal3) diverged following a combination of differentiation and degeneration in different regions. To determine which evolutionary processes shaped the evolution of these two paralog genes, it was used the duplication-degeneration-complementation model. This model is the theoretical basis in which these experiments rely and predicts that (1) degenerative mutations in regulatory elements can increase rather than reduce the probability of duplicate gene preservation and (2) the usual mechanism of duplicate gene preservation is the partitioning of ancestral functions rather than the evolution of new functions. It is proposed that the adaptive conflict regarding the regulation of these two genes was resolved along the lineage leading to S. cerevisiae by gene duplication and subfunctionalization.
This study was meant to investigate the evolution of Gal1 and Gal3 by testing the fitness differences in an ancestral yeast species Kluyveromyces lactis, which unlike S. cerevisiae has not undergone whole-genome duplication. In K.lactis, the functions of Gal1 and Gal3 are encoded by one single gene, which has one regulatory region. It was found that, in K.lactis, there is an increasing of fitness as the expression of Gal1 increases whereas the expression of Gal3 has the opposite effect. In S.cerevisiae there is no such conflict so the expectable evolutionary path should lead to an optimal regulation of the gene encoding Gal1 without maladaptive regulation of the co-inducer of this pathway, Gal3. The gradual degeneration took nearly 100million years to complete and was capped by Gal3's complete loss of galactokinase activity. Gal1 and Gal3 became integrated into a more complex and arguably more optimal genetic pathway.
This study recreates, in a very simple and elegant fashion, the evolutionary processes that took millions of years and that lead to the divergence of two genes after duplication, by studying an ancestral yeast form and comparing it to the most derived form, S. cerevisiae.
Another recent study from Hickman et al., investigated the duplication of the genes encoding the histone deacetylase enzymes Hst1p and Sir2 and it was proposed a different evolutionary mechanism from those observed for Gal1 and Gal3. In this case, the authors found that, although Hst1p and Sir2 have different functions, they use the same biochemical process for their activity.
In yeasts we have an ideal model organism for this kind of studies, since there are many species that have undergone whole-genome duplications and others that have suffered small scale segmental duplications. These analyses are relevant because other taxa and, in particular, multicellular organisms have also undergone duplication events, which may have created windows of opportunity for the appearance of new and even more complex functions.
Unraveling the processes that underlie the evolution of new functions after gene duplication is one of the most interesting focuses of scientific research, since it allows the better understanding of the basic genetic mechanisms that shaped the biological forms throughout millions of years. The new era of comparative genomics, together with the sequencing of an increasing number of genomes will provide a great deal of data for testing various hypotheses about how these processes occur.
With the increasing technological advances, we now have the opportunity to perform revolutionary experiments of large-scale genome sequencing and systems biology approaches that changes the natural-experiment-based knowledge. We now have the chance to work out the underlying processes that can lead to changes in gene function and to discover principles that will help us to understand the form and function of genes and their evolution.
It is true that complementation assays may demonstrate the effect of a certain protein but, since the optimization of a certain subfunction is expected to be quantitative, natural selection can act on very small fitness differences. Therefore, it was necessary to create a method to assess the question more directly and without any kind of bias. The grow assay performed during these experimental work is a simple but totally new way of detecting the most slight fitness difference among strains. In this assay, genetically manipulated green fluorescent protein (GFP)-tagged strains are competed against an otherwise identical blue fluorescent protein (BFP)-tagged strain in liquid culture, cells are counted by flow cytometry, and the effect on fitness is calculated. Targeted gene replacement allowed to make any desired genetic change and provided the control over genetic background necessary to detect small fitness differences. The group of experiments that were performed having the grow assay as a basis was flawless: first, it was examined the ability of various coding sequences to perform the galactokinase function of Gal1 in vivo; afterwards, it was tested the ability of these coding sequences to perform the co-induction function of Gal3 and it was shown that the peptide SA could be the source of the conflict between the galactokinase and co-induction functions but, after the experimental procedure, it is clear that the SA residues were not a source of adaptive conflict and have opposite effects on the co-induction capabilities of Gal1 and Gal3.
Moreover, it was performed an essay to address the divergence in the upstream regulatory elements in which the promoters were swapped between the two paralogues in order to assess their capabilities to drive proper expression of Gal1 and Gal3.
Other experiment was driven to test whether the helical phasing of Gal4 binding sites might be responsible for some of the difference in expression between the promoters of Gal1 in K.lactis and in S.cerevisiae.
All the experiments performed in order to assess the evolutionary changes that lead to the appearance of two functional genes with major importance in yeast by a duplication phenomenon were perfectly adequate to the questions.
To address the questions related to gene subfunctionalization you first examined the ability of various coding sequences to perform the galactokinase function of Gal1 in vivo and the results suggests that there is no effect of any coding replacement in non-inducing conditions but it were observed strong fitness defects in galactose when other proteins were required to perform the function of Gal1.
The other grow assay performed aimed to test the ability of these coding sequences to perform the co-induction function of Gal3 and the results indicates that, aside from the complete degeneration of the galactokinase activity of Gal3, most of the functional divergence between these paralogues has been regulatory, which might suggest that the upstream regulatory elements are a potential source of adaptive conflict.
To address the divergence in the upstream regulatory elements, the promoters were swapped between the two paralogues in order to assess their capabilities to drive proper expression of Gal1 and Gal3. It was seen that the promoter of Gal3 in S.cerevisiae is unable to drive sufficient levels of galactokinase due to its comparatively low mRNA expression level when fully induced. In contrast, whereas the promoter of Gal1 in S.cerevisiae is strongly induced in the presence of galactose, it is more efficiently repressed in non-inducing conditions than the promoter of Gal3. The Gal1 promoter of K.lactis when placed in S.cerevisiae provides addicional support for the idea that the promoters of Gal1 and Gal3 were subfunctionalized from an ancestral bifunctional state.
Other experiment was driven to test whether the helical phasing of Gal4 binding sites might be responsible for some of the difference in expression between the promoters of Gal1 in K.lactis and in S.cerevisiae. The results strongly suggests that the adaptive conflict was resolved by subdivision of the transcriptional regulation of each function between to specialist genes during the evolution of the lineage that gave rise to S.cerevisiae.
This simple yet of results suggests that an ancestral gene that performed two different functions, by duplication and subfunctionalization optimized the performance of those functions during the evolutionary pathway. It is a flawless manuscript and the experiments are perfectly designed.
After careful reading of your manuscript, I truly believe that it has the potential to be one of the most cited articles regarding gene duplication and that it must be published by Nature.