Patent Application
20180312878
The present application relates to the field of yeast, and specifically to the identification of yeast alleles that are involved in maximal alcohol accumulation and/or in tolerance to high alcohol levels. Preferably, said alcohol is ethanol. The identified alleles can be combined or stacked with each other to construe and/or select high alcohol tolerant yeasts, most notably Saccharomyces species.
The capacity to produce high levels of alcohol is a very rare characteristic in nature. It is most prominent in the yeast Saccharomyces cerevisiae, which is able to accumulate in the absence of cell proliferation, ethanol concentrations in the medium of more than 17% (V/V), a level that kills virtually all competing microorganisms. As a result this property allows this yeast to outcompete all other microorganisms in environments rich enough in sugar to sustain the production of such high ethanol levels (Casey and Ingledew, 1986; D'Amore and Stewart, 1987). Very few other microorganisms, e.g. the yeast Dekkera bruxellensis, have independently evolved a similar but less pronounced ethanol tolerance compared to S. cerevisiae (Rozpedowska et al., 2011). The capacity to accumulate high ethanol levels lies at the basis of the production of nearly all alcoholic beverages as well as bioethanol in industrial fermentations by the yeast S. cerevisiae. Originally, all alcoholic beverages were produced with spontaneous fermentations in which S. cerevisiae gradually increases in abundance, in parallel with the increase in the ethanol level, to finally dominate the fermentation at the end. The ability to survive and proliferate in high levels of ethanol is an ecologically important and industrially relevant trait of yeast cells. The ethanol produced by yeast cells slows down growth of competing microbes, but at higher concentrations, it causes stress for the yeast cells themselves. Different yeast strains show significant differences in their ability to grow in the presence of ethanol, with the more ethanol-tolerant ones likely having a fitness advantage over non-tolerant strains [1-3]. Moreover, ethanol tolerance is a key trait of industrial yeasts that often encounter very high ethanol concentrations, for example during beer and wine making and industrial bio-ethanol production.
The genetic basis of yeast alcohol tolerance, particularly ethanol tolerance has attracted much attention but until recently nearly all research was performed with laboratory yeast strains, which display much lower alcohol tolerance than the natural and industrial yeast strains. This research has pointed to properties like membrane lipid composition, chaperone protein expression and trehalose content, as major requirements for ethanol tolerance of laboratory strains (D'Amore and Stewart, 1987; Ding et al., 2009) but the role played by these factors in other genetic backgrounds and in establishing tolerance to very high ethanol levels has remained unknown. Different experimental approaches have been used, including screening of (deletion) mutants for increased ethanol tolerance, transcriptome analysis of ethanol stressed cells and QTL analyses aimed at identifying mutations that cause differences in ethanol tolerance between different yeast strains [4-10]. A polygenic analysis of the high ethanol tolerance of a Brazilian bioethanol production strain VR1 revealed the involvement of several genes previously never connected to ethanol tolerance and did not identify genes affecting properties classically considered to be required for ethanol tolerance in lab strains (Swinnen et al., 2012a). Together, these studies have linked multiple different genetic loci to ethanol tolerance and identified hundreds of genes, involved in a multitude of cellular processes [11-14].
A shortcoming of most previous studies is the assessment of alcohol tolerance solely by measuring growth on nutrient plates in the presence of increasing alcohol levels. (D'Amore and Stewart, 1987; Ding et al., 2009). This is a convenient assay, which allows hundreds of strains or segregants to be phenotyped simultaneously with little work and manpower. However, a more physiologically and ecologically relevant parameter of alcohol tolerance in S. cerevisiae is its capacity to accumulate by fermentation high alcohol levels in the absence of cell proliferation. This generally happens in an environment with a large excess of sugar compared to other essential nutrients. As a result, a large part of the alcohol in a typical, natural or industrial, yeast fermentation is produced with stationary phase cells in the absence of any cell proliferation. The alcohol tolerance of the yeast under such conditions determines its maximal alcohol accumulation capacity, a specific property of high ecological and industrial importance. In industrial fermentations, a higher maximal alcohol accumulation capacity allows a better attenuation of the residual sugar and therefore results in a higher yield. A higher final alcohol titer reduces the distillation costs and also lowers the liquid volumes in the factory, which has multiple beneficial effects on costs of heating, cooling, pumping and transport of liquid residue. It also lowers microbial contamination and the higher alcohol tolerance of the yeast generally also enhances the rate of fermentation especially in the later stages of the fermentation process. Maximal alcohol accumulation capacity can only be determined in individual yeast fermentations, which are much more laborious to perform than growth tests on plates. In static industrial fermentations, maintenance of the yeast in suspension is due to the strong CO2 bubbling and this can only be mimicked in lab scale with a sufficient amount of cells in a sufficiently large volume.
While it becomes increasingly clear that ethanol is a complex stress that acts on several different processes including increasing fluidity and permeability of cellular membranes, changing activity and solubility of membrane-bound and cytosolic proteins and interfering with the proton motive force (for review, see [15-17], the exact molecular mechanisms and genetic architecture underlying ethanol tolerance are still largely unknown.
Experimental evolution to study adaptation to increased ethanol levels could provide more insight into the molecular mechanisms underlying ethanol tolerance since such experiments could reveal different mutational paths that make a sensitive strain more tolerant. Only a handful of studies have looked at adaptation to ethanol in originally non-ethanol tolerant microbes exposed to gradually increasing levels of ethanol [6,18-22]. These have mostly focused on the physiological adaptations found in the evolved cells and have not performed an extensive analysis of the mechanisms and genetic changes underlying this adaptation. Hence, a comprehensive analysis of the type and number of mutations a non-ethanol tolerant strain can (or needs to) acquire to become more ethanol tolerant is still lacking. Experimental evolution has proven to be a valuable tool to investigate the different mechanisms and pathways important for cells to adapt to specific selective conditions. Seminal papers have increased our understanding of the molecular basis of adaptation to specific stresses, such as heat stress, nutrient limitation and antibiotic treatment [23-28]. Recent advances in DNA sequencing technologies allow affordable and fast sequencing of complete genomes of clones and populations. While sequencing clones yields information on individual lineages within the experiment, population data provides information on the heterogeneity of adaptation. Additionally, sequencing samples isolated at different time points during the evolution experiment makes it possible to capture evolution in action. This has provided valuable information on the rate and types of mutations underlying adaptation, the genetic basis of ‘novel’ phenotypes and the existence of parallel pathways to establish comparable phenotypic outcomes [29-32]. For example, a common strategy observed in populations evolving under nutrient limitation is the amplification of genetic regions encoding transporters responsible for the uptake of the limiting nutrient [24,33]. Other studies using multiple replicate populations have discovered a high degree of parallelism in the adaptive solutions found by different populations. Clonal interference, the competition between lineages carrying different beneficial mutations, is another commonly observed phenomenon in evolution of asexually propagating populations that can increase complexity of mutational dynamics as well as impede the spread of beneficial mutations in a population [32,34,35]. To unravel the molecular mechanisms of adaptation to a specific condition, most studies have used isogenic replicate populations, with all cells having the same initial genome size. Genome size can significantly change during evolution; with both small-scale changes (chromosomal 108 deletions and amplifications) and large-scale changes (increase or decrease in ploidy). Moreover, ploidy shifts have been reported in the evolutionary history of many organisms, including Saccharomyces cerevisiae and as a response to selective pressure [36-38]. Conversely, genome size has also been reported to affect evolution rate: polyploidy has been shown to increase adaptability [39,40]. Multiplying the amount of DNA increases the genetic material available for evolution to tinker with and can alter gene expression [41-43]. These polyploid genomes can be unstable, resulting in loss of chromosomes and thus aneuploid cells [44-46]. Although studies have looked at adaptation of lineages of different ploidy, none have followed the mutational dynamics in these evolving populations over time in detail [36,47-49].
It would be advantageous to identify novel pathways and alleles that contribute to alcohol tolerance and are relevant in industrial yeast strains as well, to further increase yield of alcohol fermentations with such strains. Indeed, even a slight increase in alcohol tolerance or alcohol accumulation capacity would have huge economic benefits.
Experimental evolution of isogenic yeast populations of different initial ploidy was used to study adaptation to increasing levels of ethanol. Evolved lineages showed a significant increase in ethanol tolerance compared to ancestral strains. High coverage whole-genome sequencing of more than 30 evolved populations and over 100 adapted clones isolated throughout a two-year evolution experiment revealed how a complex interplay of different evolutionary mechanisms led to higher tolerance, including de novo single nucleotide mutations, extensive copy number variation and ploidy changes. Although the specific mutations differ between different evolved lineages, application of a novel computational pipeline to identify target pathways reveals shared themes at the level of functional modules. Moreover, by combining an allelic replacement approach with high-throughput fitness measurements, we could identify several SNPs that arose in our adapted cells previously not implicated in ethanol tolerance and that significantly increased ethanol tolerance when introduced into a non-tolerant background. Taken together, our results show how, in contrast to adaptation to some other stresses, adaptation to a complex and severe stress involves an interplay of different evolutionary mechanisms. In addition, our study highlights the potential of experimental evolution to identify mutations that are of industrial importance.
It is an object of the invention to provide alleles that increase alcohol tolerance and/or alcohol accumulation in yeast. Most particularly, the alcohol is ethanol. Most particularly, the yeast is a Saccharomyces species. Increase in alcohol tolerance is an economically relevant property, as a non-limiting example, an increase in alcohol tolerance and/or accumulation may be favourable for bio-ethanol production. The increase in alcohol tolerance and/or accumulation may result in a higher speed of fermentation (i.e. less time needed to reach a particular % of alcohol) and/or a higher final ethanol titer.
According to specific embodiments, such alleles are selected from a MEX67 allele, a PCA1 allele, a PRT1 allele, a YBL059W allele, a HEM13 allele, a HST4 allele, a VPS70 allele. The alleles can also be in intergenic regions. According to these embodiments, the alleles are typically selected from the intergenic region of Chromosome IV (particularly around or at position 1489310) and the intergenic region of Chromosome XII (particularly around or at position 747403).
Particularly envisaged alleles are selected from the group consisting of a MEX67 allele with a G456A mutation, PCA1 with a C1583T mutation, PRT1 with a A1384G mutation, YBL059W with a G479T mutation, HEM13 with a G700C mutation, HST4 with a G262C mutation, VPS70 with a C595A mutation, the intergenic region of Chromosome IV with a A>T substitution at position 1489310, and the intergenic region of Chromosome XII with a C>T substitution at position 747403. It is particularly envisaged that the wild type alleles mentioned are selected from MEX67 as shown in SEQ ID NO: 1, PCA1 as shown in SEQ ID NO: 2, PRT1 as shown in SEQ ID NO: 3, YBL059W as shown in SEQ ID NO:4, HEM13 as shown in SEQ ID NO:5, HST4 as shown in SEQ ID NO:6, and VPS70 as shown in SEQ ID NO:7. Thus, it is particularly envisaged that the allele is selected from the group consisting of SEQ ID NO:1 with a G456A mutation, SEQ ID NO:2 with a C1583T mutation, SEQ ID NO:3 with a A1384G mutation, SEQ ID NO:4 with a G479T mutation, SEQ ID NO:5 with a G700C mutation, SEQ ID NO:6 with a G262C mutation, or SEQ ID NO:7 with a C595A mutation.
The alleles can be used to increase alcohol tolerance and/or alcohol accumulation in yeast either alone or in combination. The latter provides additive or even synergistic effects. Any permutation/combination of the 9 alleles can be used to increase alcohol tolerance and/or alcohol accumulation, and this is explicitly envisaged herein. Any combination of these alleles (including single alleles) may also be combined with known mutations or alleles that increase alcohol tolerance and/or alcohol accumulation. This is particularly the case when the alleles are to be used in industrial yeast strains adapted to have high alcohol tolerance.
Particularly envisaged combinations are those with the MEX67 allele. According to these embodiments, a MEX67 allele is provided to increase alcohol tolerance and/or alcohol accumulation in yeast. This MEX67 allele may be combined with other alcohol tolerance and/or accumulation modulating alleles. This may be known alcohol tolerance alleles. It is also explicitly foreseen that the MEX67 allele may be incorporated in an industrial yeast strain adapted to have high alcohol tolerance (and that thus already has a combination of known alcohol tolerance alleles). However, it is also explicity envisaged that the MEX67 allele is further combined with the new alcohol tolerance and/or accumulation modulating alleles reported herein. According to these embodiments, the use of a MEX67 allele is provided, wherein said MEX67 allele is combined with other alcohol tolerance and/or accumulation modulating alleles. Particularly, said other alcohol tolerance and/or accumulation modulating alleles are selected from the group consisting of PCA1, PRT1, YBL059W, HEM13, HST4, and VPS70. According to alternative embodiments, they may also be selected from the intergenic region of Chromosome IV (particularly around or at position 1489310) and the intergenic region of Chromosome XII (particularly around or at position 747403). According to specific embodiments, the MEX67 allele has a G456A mutation. According to further specific embodiments, the other alcohol tolerance and/or accumulation modulating alleles are selected from the group consisting of PCA1 with a C1583T mutation, PRT1 with a A1384G mutation, YBL059W with a G479T mutation, HEM13 with a G700C mutation, HST4 with a G262C mutation, VPS70 with a C595A mutation, the intergenic region of Chromosome IV with a A>T substitution at position 1489310, and the intergenic region of Chromosome XII with a C>T substitution at position 747403.
Albeit that the combinations with the MEX67 alleles are particularly envisaged, the above applies mutatis mutandis to the other alleles described herein (e.g. the PCA1, PRT1, YBL059W, . . . alleles).
The alleles described herein have not before been linked to increased alcohol tolerance or accumulation. This is also true for the genes and intergenic regions, with exception of VPS70. For VPS70, another allele has recently been shown to also modulate alcohol tolerance (WO2014/170330). Thus, according to particular embodiments, VPS70 is excluded from the envisaged combinations. According to alternative embodiments, when a VPS70 allele is used to increase alcohol tolerance and/or accumulation, it is the VPS70 allele with a C595A mutation.
According to a further aspect, the alleles provided herein, particularly the MEX67 allele, are used for selecting a yeast strain with higher alcohol accumulation and/or resistance. Particularly, said yeast is a Saccharomyces spp. Particularly, the alcohol is ethanol.
The selection of the strain can be carried out with every method known to the person skilled in the art. As a non-limiting example, strains may be selected on the base of an identification of the allele by PCR or hybridization. The selection may be combined by a selection for other alleles, known to be involved in alcohol accumulation and/or alcohol tolerance, such as but not limited to specific alleles of ADE1, VPS70, MKT1, APJ1, SWS2, or KIN3. Said selection may be carried out simultaneously or consecutively. In case of a consecutive selection the sequence of the selection is not important, i.e. the selection using MEX67 may be carried out before or after the other selection rounds.
(A) Experimental evolution of prototrophic, isogenic populations of different ploidy (haploid (VK111), diploid (VK145) and tetraploid (VK202)) for increased ethanol tolerance was performed in a turbidostat. Every 25 generations, the ethanol concentration in the media was increased in a stepwise manner (starting at 6% (v/v) and reaching 12% at 200 generations). Increasing the ethanol concentration from 10% to 11% dramatically reduced growth rate of evolving cells. Therefore, instead of increasing ethanol levels, we first reduced the ethanol level to 10.7% after 100 generations. (B) Red circles represent sampling points (indicated as number of generations) for which whole-genome sequencing was performed. For each circle, heterogeneous populations as well as three evolved, ethanol tolerant clones were sequenced. Sequencing of the population sample of reactor 4 at 200 generations failed, so this data is omitted from the specification. For generation 80 of reactor 1, only population data is available.
Evolved populations from different reactors show increased fitness in EtOH. Fitness was determined for population samples of each reactor after 40 generations (blue) and 200 generations (red). Fitness is expressed relative to the ancestral strain of each reactor (haploid for reactor 1-2; diploid for reactor 3-4 and tetraploid for reactor 5-6). Data represent the average of three independent measurements, error bars represent standard deviation.
Flow cytometry analysis of DNA content (stained by propidium iodide) of evolved populations, compared to ancestral haploid (red) and diploid (blue). For the clonal ancestral samples, two peaks are observed, corresponding to the G1 and G2 phase of the cell cycle. Evolved populations sometimes display three peaks, indicative of both haploid and diploid subpopulations.
Genome view of yeast chromosomes and CNV patterns from a sliding window analysis. The y axis represents log 2 ratios of the coverage observed across 500 bp genomic windows to the coverage expected in a diploid genome without CNVs. Area of the plot located between the red lines (from −0.23 to −1) marks putative CNV loss events, whereas region between the blue lines 270 (from 0.23 to 0.58) marks putative CNV gain events. It should be noted that the amplified region of chromosome XII observed in some of our clones does not correspond to the ribosomal DNA genes.
Mutations (reaching a frequency of least 20% in the evolved population samples) and corresponding frequencies were identified from population sequencing data. Muller diagram represents the hierarchical clustering of these mutations, with each color block representing a specific group of linked mutations. Indels are designated with I, whereas heterozygous mutations are in italics. Mutations present as heterozygous mutations in all clones of a specific time point and present at a frequency of 50% in the population, are depicted as a frequency of 100% in the population, since it is expected that all cells in the population contain this mutation. After 80 generations, a mutator phenotype appeared in this reactor (indicated by arrow under graph), which coincides with the rise in frequency of an indel in the mismatch repair gene MSH2. Frequencies of haplotypes, dynamics and linkage for reactor 1 were calculated but are not shown.
Shown is the sub-network that prioritizes putative adaptive mutations by applying PheNetic on all selected mutations, excluding those originating from the populations with a mutator phenotype i.e. reactor 2 and 6. The nodes in the network correspond to genes and/or their associated gene products. Node borders are colored according to the reactors containing the populations in which these genes were mutated. Nodes are colored according to gene function, for each gene the most enriched term is visualized (grey indicates no enrichment). Cell cycle related processes have been subdivided into DNA replication and interphase. The edge colors indicate different interaction types. Orange lines represent metabolic interactions, green lines represent protein-protein interactions, red lines represent protein-DNA interactions. Sub-networks were extracted by separately analyzing the mutated genes observed in each of the different populations (reactors)—this data is not shown here.
Plots show the average selection coefficient (smut) as a function of ethanol concentration for (A) pca1C1583T, (B) prt1A1384G, (C) ybl059wG479T, (D) intergenic ChrIV A1489310T, (E) hem13G700C, (F) intergenic ChrXII C747403T, (G) hst4G262C, (H) vps70C595A, and (I) mex67G456A. Superscripts denote the exact nucleotide change in each of the mutants tested. YECitrine-tagged mutants were competed with the mCherry-tagged parental strain (orange dots); dye-swap experiments were carried out by competing the mCherry-tagged mutants with the YECitrine parental strain (blue dots), except for (I). Error bars show the S.E.M. from three experimental replicates. Asterisk show P-values from the one-way ANOVA tests of the mean differences in 4-8% ethanol compared to fitness in 0% ethanol: * p<0.05; ** p<0.01; ***p<0.005.
Coding mutations were introduced into the industrial reference biofuel strain Ethanol Red. Ethanol production of these mutants is shown, relative to its proper control. All experiments were repeated 4 times (real biological replicates, using independent transformants where available).
Coding mutations were introduced into the industrial reference biofuel strain Ethanol Red. Fermentation performance in YP+35% glucose (w/v) is shown as the cumulative weight loss—a proxy for CO2 production—during the fermentation. Each line represents the average of 4 replicate fermentations, error bars represent standard deviation.
The present invention will be described with respect to particular embodiments and with reference to certain drawings but the invention is not limited thereto but only by the claims. Any reference signs in the claims shall not be construed as limiting the scope. The drawings described are only schematic and are non-limiting. In the drawings, the size of some of the elements may be exaggerated and not drawn on scale for illustrative purposes. Where the term “comprising” is used in the present description and claims, it does not exclude other elements or steps. Where an indefinite or definite article is used when referring to a singular noun e.g. “a” or “an”, “the”, this includes a plural of that noun unless something else is specifically stated.
Furthermore, the terms first, second, third and the like in the description and in the claims, are used for distinguishing between similar elements and not necessarily for describing a sequential or chronological order. It is to be understood that the terms so used are interchangeable under appropriate circumstances and that the embodiments of the invention described herein are capable of operation in other sequences than described or illustrated herein.
The following terms or definitions are provided solely to aid in the understanding of the invention. Unless specifically defined herein, all terms used herein have the same meaning as they would to one skilled in the art of the present invention. Practitioners are particularly directed to Sambrook et al., Molecular Cloning: A Laboratory Manual, 2nd ed., Cold Spring Harbor Press, Plainsview, N.Y. (1989); and Ausubel et al., Current Protocols in Molecular Biology (Supplement 47), John Wiley & Sons, New York (1999), for definitions and terms of the art. The definitions provided herein should not be construed to have a scope less than understood by a person of ordinary skill in the art.
An allele as used here is a specific form of the gene, which is carrying SNP's or other mutations, either in the coding (reading frame) or the non-coding (promoter region, or 5′ or 3′ non-translated end) part of the gene, wherein said mutations distinguish the specific form from other forms of the gene.
Gene as used here includes both the promoter and terminator region of the gene as well as the coding sequence. It refers both to the genomic sequence (including possible introns) as well as to the cDNA derived from the spliced messenger, operably linked to a promoter sequence.
Coding sequence is a nucleotide sequence, which is transcribed into mRNA and/or translated into a polypeptide when placed under the control of appropriate regulatory sequences. The boundaries of the coding sequence are determined by a translation start codon at the 5′-terminus and a translation stop codon at the 3′-terminus. A coding sequence can include, but is not limited to mRNA, cDNA, recombinant nucleotide sequences or genomic DNA, while introns may be present as well under certain circumstances.
Promoter region of a gene as used here refers to a functional DNA sequence unit that, when operably linked to a coding sequence and possibly a terminator sequence, as well as possibly placed in the appropriate inducing conditions, is sufficient to promote transcription of said coding sequence
Nucleotide sequence”, “DNA sequence” or “nucleic acid molecule(s)” as used herein refers to a polymeric form of nucleotides of any length, either ribonucleotides or deoxyribonucleotides. This term refers only to the primary structure of the molecule. Thus, this term includes double- and single-stranded DNA, and RNA. It also includes known types of modifications, for example, methylation, “caps” substitution of one or more of the naturally occurring nucleotides with an analog.
Modulation of alcohol accumulation and/or tolerance, as used here, means an increase or a decrease of the alcohol concentration, produced by the yeast carrying the specific allele, as compared with the alcohol concentration produced under identical conditions by a yeast that is genetically identical, apart from the specific allele(s)
Alcohol as used here can be any kind of alcohol, including, but not limited to methanol, ethanol, n- and isopropanol, n- and isobutanol. Indeed, several publications indicate that the tolerance to ethanol and other alkanols is determined by the same mechanisms (Carlsen et al., 1991; Casal et al., 1998).
The causes and mechanisms underlying evolutionary adaptation are key issues in biology. However, the mechanisms underlying evolutionary adaptation to complex and severe stress factors (e.g. broadly acting toxins or extreme niches that require immediate, complex adaptive changes) remain understudied. Here, we study how yeast populations adapt to gradually increasing ethanol levels—a broad-acting, ecologically and industrially relevant complex stress that is still poorly understood. Our results reveal how over a two-year evolution period, several evolutionary mechanisms—including mutator phenotypes, changes in ploidy, and complex clonal interference—result in adapted populations capable of surviving in medium containing up to 12% ethanol. In addition, we report several previously unknown adaptive mutations that increase ethanol resistance; which may open new routes to increase the efficiency of industrial fermentations. Specifically, we exposed six initially isogenic yeast populations of different ploidy to increasing levels of ethanol. We combined high-coverage whole-genome sequencing of more than 30 populations and 100 clones isolated throughout this two-year evolution experiment with a novel computational pipeline to reveal the mutational dynamics, molecular mechanisms and networks underlying increased ethanol tolerance of our evolved lineages. High-throughput fitness measurements allowed us to characterize the phenotypic effect of identified mutations in different environments. Our results suggest that adaptation to high ethanol is complex and can be reached through different mutational pathways. We find that adaptation to high ethanol levels involves the appearance of mutator phenotypes, and evolving populations showing strong clonal interference. Evolved cells display extensive variation in genome size; with initially haploid and tetraploid populations showing quick convergence to a diploid state. Our results are the first to attribute a significant fitness advantage of a diploid cell over an isogenic haploid cell under selective conditions. In addition, evolved clones repeatedly gained extra copies of the same chromosomes. By combining an allelic replacement approach with high-throughput fitness measurements, we could identify several mutations previously not implicated in ethanol tolerance that significantly increased ethanol tolerance when introduced into a non-tolerant background. Together, our study yields a detailed view of the molecular evolutionary processes and genetic changes underlying long-term adaptation to a severe and complex stress, and highlights the potential of experimental evolution to identify mutations that are of industrial importance. The methods developed and insights gained serve as a model for adaptation to a complex and lethal stress. Moreover, specific adaptive SNPs can be used to guide engineering strategies aimed at obtaining superior biofuel yeasts. Last but not least, our experimental setup is also interesting because it takes an atypical approach. In many experimental evolution studies, populations are exposed for a limited amount of time (typically a few months) to a defined, constant and ‘simple’ stress (e.g. a fixed concentration of antibiotic or limiting nutrient), so that in the beginning of the experiment, cells are confronted with a very strong selection, while the selection pressure is reduced or even eliminated when cells become resistant. In our experimental set-up, ethanol levels were gradually increased over time so that cells were evolving under constant selection over a long (two-year) period. High concentrations of ethanol slow down growth and actively kill non-tolerant cells. Our work thus combines aspects of traditional evolution studies with principles used in so-called morbidostat experiments, and may therefore offer a realistic picture of adaptation to a complex and severe stress where a population gradually penetrates a new niche.
It is an object of the invention to provide alleles that increase alcohol tolerance and/or alcohol accumulation in yeast. Most particularly, the alcohol is ethanol. Most particularly, the yeast is a Saccharomyces species.
According to specific embodiments, such alleles are selected from a MEX67 allele, a PCA1 allele, a PRT1 allele, a YBL059W allele, a HEM13 allele, a HST4 allele, a VPS70 allele. The alleles can also be in intergenic regions. According to these embodiments, the alleles are typically selected from the intergenic region of Chromosome IV (particularly around or at position 1489310) and the intergenic region of Chromosome XII (particularly around or at position 747403).
Particularly envisaged alleles are selected from the group consisting of a MEX67 allele with a G456A mutation, PCA1 with a C1583T mutation, PRT1 with a A1384G mutation, YBL059W with a G479T mutation, HEM13 with a G700C mutation, HST4 with a G262C mutation, VPS70 with a C595A mutation, the intergenic region of Chromosome IV with a A>T substitution at position 1489310, and the intergenic region of Chromosome XII with a C>T substitution at position 747403.
The alleles can be used to increase alcohol tolerance and/or alcohol accumulation in yeast either alone or in combination. The latter provides additive or even synergistic effects. Any permutation/combination of the 9 alleles can be used to increase alcohol tolerance and/or alcohol accumulation, and this is explicitly envisaged herein. Any combination of these alleles (including single alleles) may also be combined with known mutations or alleles that increase alcohol tolerance and/or alcohol accumulation. This is particularly the case when the alleles are to be used in industrial yeast strains adapted to have high alcohol tolerance.
Particularly envisaged combinations are those with the MEX67 allele. According to these embodiments, a MEX67 allele is provided to increase alcohol tolerance and/or alcohol accumulation in yeast. This MEX67 allele may be combined with other alcohol tolerance and/or accumulation modulating alleles. This may be known alcohol tolerance alleles. It is also explicitly foreseen that the MEX67 allele may be incorporated in an industrial yeast strain adapted to have high alcohol tolerance (and that thus already has a combination of known alcohol tolerance alleles). Non-limiting examples include ADE1, INO1, VPS70, MKT1, APJ1, SWS2, PDR1, or KIN3.
However, it is also explicitly envisaged that the MEX67 allele is further combined with the new alcohol tolerance and/or accumulation modulating alleles reported herein. According to these embodiments, the use of a MEX67 allele is provided, wherein said MEX67 allele is combined with other alcohol tolerance and/or accumulation modulating alleles. Particularly, said other alcohol tolerance and/or accumulation modulating alleles are selected from the group consisting of PCA1, PRT1, YBL059W, HEM13, HST4, and VPS70. According to alternative embodiments, they may also be selected from the intergenic region of Chromosome IV (particularly around or at position 1489310) and the intergenic region of Chromosome XII (particularly around or at position 747403). According to specific embodiments, the MEX67 allele has a G456A mutation. According to further specific embodiments, the other alcohol tolerance and/or accumulation modulating alleles are selected from the group consisting of PCA1 with a C1583T mutation, PRT1 with a A1384G mutation, YBL059W with a G479T mutation, HEM13 with a G700C mutation, HST4 with a G262C mutation, VPS70 with a C595A mutation, the intergenic region of Chromosome IV with a A>T substitution at position 1489310, and the intergenic region of Chromosome XII with a C>T substitution at position 747403.
Albeit that the combinations with the MEX67 alleles are particularly envisaged, the above applies mutatis mutandis to the other alleles described herein (e.g. the PCA1, PRT1, YBL059W, . . . alleles).
The alleles described herein have not before been linked to increased alcohol tolerance or accumulation. This is also true for the genes and intergenic regions, with exception of VPS70. For VPS70, another allele has recently been shown to also modulate alcohol tolerance. Thus, according to particular embodiments, VPS70 is excluded from the envisaged combinations. According to alternative embodiments, when a VPS70 allele is used to increase alcohol tolerance and/or accumulation, it is the VPS70 allele with a C595A mutation
According to a further aspect, the alleles provided herein, particularly the MEX67 allele, are used for selecting a yeast strain with higher alcohol accumulation and/or resistance. Particularly, said yeast is a Saccharomyces spp. Particularly, the alcohol is ethanol.
In a specific embodiment the invention provides a yeast strain, particularly an industrial yeast strain, particularly an industrial yeast strain of Saccharomyces cerevisiae comprising at least one allele selected from the group consisting of a MEX67 allele with a G456A mutation, PCA1 with a C1583T mutation, PRT1 with a A1384G mutation, YBL059W with a G479T mutation, HEM13 with a G700C mutation, HST4 with a G262C mutation, VPS70 with a C595A mutation, the intergenic region of Chromosome IV with a A>T substitution at position 1489310, and the intergenic region of Chromosome XII with a C>T substitution at position 747403.
The selection of the strain can be carried out with every method known to the person skilled in the art. As a non-limiting example, strains may be selected on the base of an identification of the allele by PCR or hybridization. The selection may be combined by a selection for other alleles, known to be involved in alcohol accumulation and/or alcohol tolerance, such as but not limited to specific alleles of ADE1, VPS70, MKT1, APJ1, SWS2, or KIN3. Said selection may be carried out simultaneously or consecutively. In case of a consecutive selection the sequence of the selection is not important, i.e. the selection using MEX67 may be carried out before or after the other selection rounds.
It is to be understood that although particular embodiments, specific configurations as well as materials and/or molecules, have been discussed herein for cells and methods according to the present invention, various changes or modifications in form and detail may be made without departing from the scope and spirit of this invention. The following examples are provided to better illustrate particular embodiments, and they should not be considered limiting the application. The application is limited only by the claims.
In this study, we use experimental evolution to dissect the adaptive mechanisms underlying ethanol tolerance in yeast. Six isogenic Saccharomyces cerevisiae populations of different ploidy (haploid, diploid and tetraploid) were asexually propagated in a turbidostat over a two year period, with ethanol levels gradually increasing during the experiment. This step-wise increase in exogenous ethanol levels resulted in a constant selective pressure for our cells. Whole-genome sequencing of evolved populations and isolated, ethanol-tolerant clones at different times during the experiment allowed us to paint a detailed picture of the mutational dynamics in the different populations. Several common themes in the type of adaptations and evolutionary mechanisms emerge, with all lines showing extensive clonal interference as well as copy number variations. Additionally, the haploid and tetraploid lines showed rapid convergence towards a diploid state. Despite these common themes, we find multiple lineage-specific adaptations, with little overlap in the mutated genes between the different populations. By applying a novel computational pipeline to identify affected pathways, we could reveal overlap between the functional modules affected in the different adapted populations, with both novel and previously established pathways and genes contributing to ethanol tolerance. Importantly, introduction of specific mutated alleles present in adapted populations into the ancestral strain significantly increased its ethanol tolerance, demonstrating the adaptive nature of these mutations and the potential of using experimental evolution to unravel and improve a complex phenotype.
To study the mutational dynamics underlying increased ethanol tolerance, six prototrophic, isogenic S. cerevisiae strains of different ploidy (two haploid, two diploid and two tetraploid lines—with each line of the same initial ploidy started from the same preculture (VK111, VK145 and VK202 respectively) were subjected to increasing levels of ethanol. This was done by gradually increasing ethanol levels from 6% (v/v) to 12% over a two-year period in a continuous turbidostat with glucose (4% (w/v)) as a carbon source. Samples were taken at regular time intervals and subjected to whole-genome sequencing analysis. In addition to sequencing each of the 6 evolving populations, we also sequenced the genomes of three clones isolated from each of the population samples, resulting in a total of 34 population samples and 102 clonal samples that were sequenced.
We determined the relative fitness of the evolved populations, isolated at 40 and 200 generations, and generally observed increases in fitness in high (9% v/v) EtOH (
Propidium iodide staining and flow cytometry analysis of evolved populations showed that diploid cells appeared relatively quickly in the originally haploid populations (
To identify the mutational pathways during adaptation to ethanol, we performed whole-genome sequencing of populations of evolving cells throughout the two-year experiment. Each population sample was sequenced to 500 fold coverage on average, and this for multiple time points (34 samples for all reactors combined, see
After 2 years, yielding around 200 generations, evolved clones (excluding clones from reactor 2 and 6, which acquired a mutator phenotype, see below) contained on average 23 SNPs compared to the ancestral strain (data not shown). This number is higher than what would be expected based on measured rates of spontaneous mutations [50], and could reflect an increased mutation rate under the stressful conditions imposed by the high ethanol levels in our set-up. Across all reactors and clones sequenced, we identified a total of 8932 different sites mutated. The largest fraction are SNPs (6424 out of 8932; 72%); Indels are found mostly in non-coding regions (1830 out of 2508; 73%), whereas SNPs are mostly found inside genes (4971 out of 6424; 77%). Most of these coding SNPs are non-synonymous (3672 out of 4971; 74%). In two reactors (reactor 2 and reactor 6), we noticed a marked increase in the number of mutations found in individual clones (see
While the sequencing of clones yielded valuable information on individual lineages within the evolving populations; sequencing of evolving populations yielded more information on the complex mutational paths and dynamics between sub-populations within each evolving population. We observe distinct pattern of mutations appearing and disappearing over time in each of the six reactors. Some of these mutations remain in the population, eventually reaching high levels or even complete fixation (i.e. presence in 100% of all cells in the population). Other mutations only persist for a short time, until lineages carrying these mutations are outcompeted by others, so called clonal interference. In total, we identified 1637 mutations across all populations and time points. 117 of these mutations are no longer present in the final time points sequenced, and 101 mutations drop more than 10% in frequency after reaching their maximum frequency, indicative of clonal interference. Interestingly, we also identified some overlap between the mutations found in different independently evolving populations (i.e. in different reactors). Specifically, we find 20 genes that are mutated twice in different generations and populations, 3 genes mutated 3 times, 2 genes mutated twice, 2 genes mutated 5 times and 1 gene 6 times (data not shown). This significantly differs from what would be expected by chance (see Materials and Methods). Repeatedly hit genes are, amongst others, involved in stress response, cell cycle and heme biosynthesis. The higher number of sequenced samples from reactors 1 and 2 allowed us to further analyze these population sequences and group mutations based on correlations in the changes in their respective frequencies (based on the pipeline developed by [32]; see also Materials and Methods). This yields a more detailed picture of the different co-evolving sub-populations present in these reactors, which is depicted in the Muller diagrams of
In both reactors, selective sweeps mostly consists of groups of mutations that move through the population together. While these reactors were inoculated with the same strain, the type and dynamics of mutations observed during adaptation appear very different. However, both evolving populations of reactor 1 and reactor 2 are characterized by strong clonal interference. In reactor 1, 4 different subpopulations are present around generation 90, each carrying different mutations. By generation 130, these lineages have been outcompeted by another lineage that almost completely dominates the population by 200 generations (data not shown). In reactor 2, a lineage carrying a mutation in PDE2 (encoding a high-affinity cAMP phosphodiesterase) is driven to extinction by a subpopulation carrying indels in ASG1 and MSH2. ASG1 is a transcriptional regulator involved in the stress response and has been found mutated in evolved populations from different reactors (including reactor 1, see also
The results discussed above revealed extensive variability in the type and number of mutations present in each evolving population. While this could suggest the presence of several, different mutational pathways (and thus lack of parallel evolution), mutations in different genes might affect identical or similar pathways, implying that the physiological adaptation to high ethanol might in fact be more similar than what is immediately apparent from the individual mutations. To gain insight into the affected biological pathways and investigate the possible similarities in adaptation to increased ethanol, different computational approaches were used. First, a term-enrichment analysis was performed on the complete list of mutated genes for all reactors (for enriched clusters, see Table 1 and data not shown). These enrichment methods have been used as one of the standard functional analysis tools and gave us a first insight into potential adaptive pathways present in our evolved lineages.
In a second step, we used a sub-network-based selection method [58,59] (see Materials and Methods) first developed for E. coli expression data. Here, we adapted and extended this method to select the subnetwork from the global yeast interaction network that best connected the mutated genes in the most parsimonious way. This method also identifies the intermediary genes involved in signaling mechanisms, which are not necessarily mutated in our evolved lineages but mediate the cellular response. We excluded mutations obtained from the populations with a mutator phenotype (reactor 2 and 6) because of their low signal to noise ratio. This analysis identifies genes frequently mutated in the different populations (DSK2, ASG1), as well as genes that are closely connected on the interaction graph (HEM3, HEM12, . . . ). This latter set reflects parallelism at the pathway level in the different reactors. From
Our analyses suggest that the cell cycle and DNA replication are affected in our evolved lineages (
PheNetic analysis shows that protoporphyrinogen metabolism is affected in our evolved populations (
The high number of mutations (both SNPs and Indels) precluded an exhaustive analysis of all mutations present in our tolerant clones and populations and their effect on ethanol tolerance. One commonly used approach to investigate putative beneficial mutations is backcrossing of evolved clones to their ancestor, which does not contain any mutation. However, since each of the evolved populations proved unable to form any spores, this strategy was not accessible to us. Hence, we focused on SNPs reaching high frequencies in our 200 generation population samples. In total, nine SNPs were selected for further study (Table 2). Several of the mutated genes belong to or are linked to the processes affected across and/or within specific reactors (heme metabolism, protein transport and cell cycle, see also Table 2). These nine SNPs were subsequently introduced into the ancestral haploid strain. The effect of these mutations was assessed by high-throughput competition experiments [64], in 0, 4, 6 and 8% (v/v) EtOH conditions, with glucose as a carbon source (
Of all mutations introduced into the ancestral strain, vps70C595A provided the largest fitness increase. VPS70 is putatively involved in sorting of vacuolar carboxypeptidase Y to the vacuole [65]. A mutation in VPS70 has been recently identified by members of our team as a determinant of ethanol tolerance in a Brazilian bioethanol strain [66]. Interestingly, the VPS70 mutation in our evolved ethanol-tolerant lineages alters the same amino acid as the mutation present in the industrial ethanol tolerant strain (Goovaerts A. and Thevelein J. M., personal communication). The fact that one of the mutations identified in our evolved lineages was also found in an industrially used bio-ethanol strain underscores the potential of our approach to find biologically relevant mutations for increased ethanol tolerance. Other members of the VPS family have been previously implicated in ethanol tolerance as well, but also here the exact molecular mechanism through which they could increase ethanol tolerance is still unclear [10,12,67]. To our knowledge, none of the other genes investigated in this study have been previously implicated in ethanol tolerance nor have these mutations been found in natural or industrial yeasts so far. This implies that these mutations could be prime candidates to improve the ethanol tolerance and production of existing industrial yeasts [68].
As a next step, the 7 coding mutations shown in Table 2 (identified with a lab yeast strain) were introduced into the industrial reference biofuel strain Ethanol Red, a diploid Saccharomyces cerevisiae strain. For each of the coding mutations, the two wild type alleles in Ethanol Red were replaced with two mutant alleles (more details can be found in the materials and methods section). Next, maximal ethanol accumulation in YP+35% (w/V) glucose was determined for each of these mutant strains.
The relative ethanol production of these mutants is summarised in
All but one of the mutants yielded higher ethanol titers. Statistical significance was reached for the mex67 mutant. This mutation showed the highest relative increase in the industrial strain. Of note, the relative increase in ethanol yield of around 4% is not 4% ABV extra, but an increase of 4% compared to the proper control. In practice, this corresponds to approximately 0.8% alcohol extra. All experiments were repeated 4 times (real biological replicates, using independent transformants where available).
Fermentation capacity in YP+35% glucose (w/v) of each of the mutants was comparable to that of the wild type Ethanol Red strain. Fermentation capacity of these mutants is summarised in
Many fundamental questions on the dynamics and genetics of adaptation to a complex and severe stress such as ethanol are still unanswered. Which type and number of mutations are needed and/or sufficient? To what extent are these mutations and the pathways they affect predictable? To address these questions, we performed high-coverage whole-genome sequencing of clones and populations isolated throughout a two-year evolution experiment and characterized their adaptation to increasing ethanol levels. This allowed us to paint a detailed picture of the mutational dynamics in our evolving lineages.
Our study demonstrates how many different evolutionary mechanisms all come together to provide adaptation to a severe and complex stress. Specifically, we find that adaptation to high ethanol levels involves changes in ploidy, copy number variation and the appearance of mutator phenotypes, with evolving populations showing strong clonal interference. Although these mechanisms have been observed in other studies (see, for example, [24,27,32,69-71], different mechanisms are often observed and/or studied separately. More importantly, in traditional evolution studies, populations are exposed to a fixed concentration of antibiotic or limiting nutrient. Under these conditions, selection pressure is reduced or even eliminated when cells become resistant. In our experimental set-up, ethanol levels were stepwise increased over time so that cells were evolving under constant selection. High concentrations of ethanol also actively kill non-tolerant cells, so that cells not adapted to the higher ethanol concentrations die and/or are washed out of the turbidostat. Taken together, our study combines aspects of traditional evolution studies with principles used in morbidostat experiments [27]. Under such severe stress conditions (near-morbidostats), the number of generations may not be the ultimate way of measuring evolutionary time: cell death and mutations that are not associated with DNA replication become increasingly important, which could result in quick sweeps.
We find extensive variation in genome size in our evolved cells; with aneuploidies in a large number of our evolved clones as well as convergence toward a diploid state in our initially haploid and tetraploid populations. Becoming diploid thus appears to be a frequently used strategy by cells of different ploidy which effectively increases ethanol tolerance. Convergence to diploidy has been observed in other laboratory evolution experiments [25,36,72]. In contrast to these studies, we could demonstrate a clear fitness advantage of our diploid strains in ethanol. Although further research is needed to fully understand this fitness advantage, changes in gene expression and/or cell size could be important factors contributing to the increased fitness of diploid cells [41,73,74]. Unfortunately, this convergent evolution prevented us from performing a detailed analysis on the difference in adaptive strategies employed by haploids, diploids and tetraploids to increase ethanol tolerance.
Aneuploidy and copy-number variation are increasingly recognized as common themes in rapid adaptation [24,28,69,70]. It is believed that changes in the copy number of chromosomes or chromosomal fragments provide a relatively easily accessible way to change expression levels of specific key genes [75,76], and these CNVs can provide large, usually condition-specific, fitness effects [77]. However, such large-scale changes in the genome likely also have some unwanted, detrimental side effects, such as imbalance between gene products and genome destabilization [78-80]. Evolving populations are believed to gradually replace adaptive CNVs with more specific mutations that show fewer pleiotropic effects [28]. Notably, several of our evolved clones, isolated from different reactors, carry an extra copy of chromosome III and/or ChrXII; pointing towards a potential adaptive benefit of this specific aneuploidy. Interestingly, we find that clones isolated at later time points have a specific, smaller region of chromosome XII amplified (data not shown); indicating a more refined solution. GO enrichment and network analyses of the repeatedly amplified region of ChrXII (position 657500-818000) hints at cell wall formation as one of the key processes affected by these amplifications. Previous studies have indeed shown that cell wall stability is a key factor involved in ethanol tolerance [18].
Apart from diploidization of our evolving lineages, another example of parallelism at the phenotypic level is the appearance of a mutator phenotype in two of our six evolving populations. The sweep of the MSH2 mutation in reactor 2 is likely caused by a so-called hitchhiking event, with the high mutation rates in the MSH2 mutant leading to the appearance of one or several beneficial mutations that drive the selective sweep. Because of the lower temporal resolution of sequenced population samples of reactor 6, identifying the allele(s) underlying the mutator phenotype in this reactor has proven to be difficult. Parallelism at the genotypic level is less clear: we find few mutations and mutated genes shared between the different evolved lineages. Applying different types of network and enrichment analyses revealed functional modules affected in several of the adapted populations. These pathways include response to stress, intracellular signal transduction, cell cycle and pathways related to membrane composition and organization (such as isoprenoid metabolism, glycerophospholipid catabolism and fatty-acyl-coA metabolism). For some of these pathways, further work is needed to clarify their exact involvement in ethanol tolerance.
To investigate the phenotypic effect of mutations present in our evolved lineages, we performed high-throughput fitness measurements in different ethanol concentrations. While our evolved lineages contained multiple mutations, single mutations reaching high frequency in the evolved populations could already significantly increase ethanol tolerance when introduced into the ancestral, non-evolved haploid strain. Moreover, several of these single mutations selected for further study also showed a (modest) fitness benefit in conditions with no ethanol, with greater increases as ethanol levels rise. The number of mutations identified in this study was too large to investigate the fitness effect of each individual mutation.
However, our strategy to select mutations that reached fixation in the evolved populations clearly proved successful, identifying as many as 7 mutations (out of 9 tested) that confer a fitness advantage in high ethanol environments. Mutations in genes linked to processes identified as affected across our different reactors—protein transport (VPS70) and heme metabolism (HEM13 and PCA1)—indeed increased fitness in EtOH. This underscores the potential of PheNetic, a sub-network based selection method for identifying adaptive mutations. Two of the mutations tested, ybl059wG479T and hst4G262C, did not increase fitness, although they reached high frequency in our adapted populations. This is indicative of hitch-hiking with other, beneficial mutations; or possible epistatic interactions with other mutations.
Why then would not all feral yeasts show high ethanol tolerance, if it appears so easy to attain? Firstly, it seems plausible that not all yeasts are confronted with selection for high ethanol tolerance. Furthermore, it is important to note that we have not tested the fitness of the mutants under many different conditions that mimic the natural habitats of yeasts. It seems likely that some of the mutations identified in this study would result in lower fitness in other environments [81,82]. Moreover, we have not investigated the effect of combined mutations. While it is possible that combining different mutations could increase ethanol tolerance even further, it also seems likely that some mutations and/or specific combinations of mutations could lead to reduced fitness in different low and/or high ethanol environments. Indeed, while our clones have increased fitness in EtOH, we observe that fitness of several of our evolved clones (containing multiple other mutations apart from the ones investigated in this study) decreased in medium without exogenous EtOH (data not shown). These results are indicative of antagonistic pleiotropy: the specific mutations present in our evolved clones increase fitness in one condition (high ethanol, which was selected for), whereas they reduce fitness in other environments. Ethanol resistance is an important trait for the survival of feral yeasts in nature because the ethanol produced inhibits growth of competing microorganisms, while it serves as a carbon source in later stages of growth, when all fermentable sugars are depleted (the so-called make accumulate-consume strategy [83-85]). Our results suggest that adaptation to high ethanol is complex and can be reached through different mutational pathways. Apart from yielding insight into the evolutionary mechanisms leading to such complex and ecologically important phenotypes, our study is also of considerable industrial importance. Several of the mutations identified in this study may be useful to increase the ethanol tolerance of industrial strains used for the production of alcoholic beverages or biofuels.
Strains Used in this Study
Starting strains for the evolution experiment are all derived from the haploid prototrophic S288cstrain FY5 [86]. To prevent clumping of cells during the evolution experiment, the flocculation genes FLO1, FLO10 and FLO11 were deleted in this strain using deletion cassettes based on pUG6, conferring resistance to G-418 disulfate [87]. Markers were removed through the Cre/LoxP technique using pSH65 [88]. Mating type switching of this strain was then performed, using plasmid pSB283, to create isogenic diploid and tetraploid strains. Fluorescent versions of strains (YECitrine or mCherry tagged) were constructed by integrating fluorescent markers at an intergenic, neutral region of chromosome II.
Populations were founded in 400 mL ethanol containing media. Media contained 10 g/L yeast extract, 20 g/L bactopeptone, 4% (w/v) glucose, 0.001% (v/v) Rhodorsil, Antifoam Silicone 46R, chloramphenicol (50 μg/mL) and increasing concentrations of ethanol. Populations were maintained at an average population size of 1010 cells. After 25 generations, the level of EtOH in the media was increased each time (starting at 6% (v/v) and reaching 12% at 200 generations).
Turbidostat cultures were maintained using Sixfors reactors (Infors) at 30° C., pH was kept constant at 5.0 with continuous mixing at 250 rpm in aerobic conditions. At regular times, a population sample was obtained from each of the cultures for further analyses and stored in glycerol at −80° C. For DNA extraction purposes, a population cell pellet was also frozen down at −80° C.
Fitness for all evolved strains was determined in rich medium (YP, 2% (w/v) glucose) with 9% (v/v) ethanol, by competing strains against a YECitrine labeled ancestral strain. Cultures were pre-grown in YPD 6% ethanol. After 12h, wells of a 96 deep well plate were inoculated with equal numbers of labeled reference and unlabeled strains (˜106 cells of each) and allowed to grow for around 10 generations. Outer wells only contained medium and acted as a buffer to prevent ethanol evaporation. Additionally, plates were closed with an adhesive seal and plastic lid, and parafilm was used to prevent ethanol evaporation. Cultures were regularly transferred to new medium to prevent nutrient depletion. The ratio of the two competitors was quantified at the initial and final time points by flow cytometry. Data analysis was done in FlowJo version 10. Measurements were corrected for the small percentage of labeled, non-fluorescent cells that occurred even when the reference strain was cultured separately as well as for the cost of YFP expression in the labeled reference strain. For each fitness measurement, three independent replicates were performed. The selective advantage, s, of each strain was calculated as s=(ln(Uf/Rf)−ln(Ui/Ri))/T where U and R are the numbers of unlabeled and reference strain respectively, the subscripts refer to final and initial populations and T is the number of generations that reference cells have proliferated during the competition. The fitness of the unlabeled WT strain was designated 1, fitness of the evolved strains as 1+s.
DNA content of evolved populations and evolved clones was determined by staining cells with propidiumiodide (PI) and analyzing 50 000 cells by flow cytometry on a BD Influx. The ancestral haploid and diploid strains used in the evolution experiment were used for calibration.
For evolved populations, genomic DNA was directly extracted from pellets that were frozen at the time of sample taking. Evolved clones were selected from the different population samples by streaking glycerol stocks from the corresponding population samples on YPD plates. Swabs from each population were subsequently grown in YPD 6% EtOH and dilutions were plated on YPD plates with different ethanol concentrations (ranging from 8% to 10%; with a 0.5% stepwise increase in ethanol concentrations). From these plates, ethanol tolerant clones were selected and genomic DNA of these clones was extracted. Genomic DNA was prepared using the Qiagen genomic tip kit. Final DNA concentrations were measured using Qubit. Paired-end sequencing libraries with a mean insert size of 500 bp were prepared and libraries were run on an Illumina HiSeq2000 (EMBL GeneCore facility, Heidelberg). Average sequencing coverage for clone and population sample is 80× and 500× respectively.
Adaptors removal and reads quality control were done by Trim Galore! (http://www.bioinformatics.babraham.ac.uk/projects/trim_galore/) with options (−q 30 −length 50). The clean reads were then mapped reference S. cerevisiae genome (S288C, version genebank64) using Burrows Wheeler Alignment allowing maximum 50 bp gaps [89]. To identify mutations we employed the BROAD Institute Genome Analysis Toolkit 628 (GATK, version 3.1) [90]. We began by performing local realignment of reads around Indels, in order to eliminate false positives due to misalignment of reads, which was followed by a base recalibration step. We then performed SNP and Indel calling according to the GATK best practice recommendations. Population samples and single clone datasets were analyzed independently. For population data, Indel calling was performed using the UnifiedGenotyper tool, and the identified variants were filtered based on the recommended criteria (QD<2.0, FS>200.0, ReadPosRankSum<−20 and InbreedingCoef<−0.8). For the SNPs, we used the UnifiedGenotyper to perform multi-sample SNP calling on all the population samples together. The resulting multi-sample callset was then used as a prior to call SNPs in all individual samples. The called SNPs were then subject to filtering (QD<2.0, FS>60, ReadPosRankSum<−8.0, MQ<40, MQRankSum<−12.5). Furthermore, we filtered out SNPs that were called inside regions containing Indels. We also generated a list of all known repetitive regions and low complexity regions using the program RepeatMasker (http://www.repeatmasker.org/cgibin/WEBRepeatMasker), which were masked out from the assemblies. Finally, variants present in the ancestral strain were filtered out from all samples. A similar approach was used to analyze the clonal samples. Ploidy level of each isolated clone was determined by PI staining and variant calling was performed with UnifiedGenotyper for haploid clones or HaplotypeCaller for diploid clones. Subsequently, SNPs and Indels were filtered based on the GATK best practice recommendations as mentioned above. The inbreeding coefficient filter was excluded in the clonal sample filtering for Indels, as it is a population-level metric. Annotation and effect prediction of all identified variants was performed using snpEff [91]. Final processing of the data was performed using a R script, which involved filtering out variants called within the sub-telomeric regions (15 kbps from the chromosome ends) and—for population samples—variants whose frequency did not change by more than 10% during the experiment. We used the infrastructure of the VSC—Flemish Supercomputer Center for these analyses. For improved performance, CPU multithreading (−nct) capabilities of GATK were utilized, providing up to 20 cores and maximum memory capacity of 64 GB per process.
CNVs in the samples were identified using Nexus Copy Number software, version 7.5 (http://www.biodiscovery.com/software/nexus-copy-nurnber). Reads were accumulated in 500-base bins along all chromosomes, rejecting bins with less than 10 reads. Log 2 ratios of copy numbers were then estimated from the read-depth counts in these intervals. We used the following calling parameters: minimum number of probes per segment of 5, significance threshold of p-value=1×10̂-9, percentage of removed outliers of 5% and the limits for copy gain and loss of +0.25 and −0.25, respectively.
The p-values probabilities of genes being hit by a mutation a specific number of times were calculated using the binomial distribution. The number of draws was set to the total number of mutations found inside coding regions (i.e. 817); the number of successes was set to the number of times a specific gene was hit by a mutation (2 to 6); the probability of success was set to the specific ratio of the length of each gene hit multiple times (in nucleotides) over the entire coding content of the S. cerevisiae genome (9080922 nucleotides).
Haplotype reconstruction was based on the approach described by [32]. Prior to the actual reconstruction, variants identified in the sequencing data were subject to further processing by excluding variants that were multi-allelic, variants that exhibited mixed zygosity in the isolated clones (i.e. homozygous in one clone and heterozygous in another clone) and variants that didn't reach the frequency of 0.2 at any point during the experiment. Haplotypes (or ‘mutational cohorts’) present in the remaining variants were next reconstructed using a Matlab script (kindly provided by the Desai lab) on a local machine running Matlab 2013a. Briefly, variants present in our dataset were clustered into haplotypes based on the Euclidean distance between their frequencies at specific time-points of the experiment. Afterward, frequencies of individual variants assigned to a specific haplotype were averaged to obtain the frequency of the haplotype itself. Frequencies of the identified haplotypes at specific time points (data not shown) were then used as source data to draw their approximate Muller diagram representations with lnkscape 0.48.4.
Functionally meaningful terms enriched in the list of genes hit in our evolution experiment were identified using DAVID Tools [57,92]. Overall, we identified 170 clusters of functionally meaningful terms, of which 10 passed our statistical enrichment criteria. An enrichment score cutoff value of 1.3 was used, equivalent to a p-value of 0.05 for term enrichment, as recommended by the authors [57].
Intergenic mutations were discarded for the analysis. The interaction network used as input for PheNetic was composed of interactomics data obtained from KEGG [93] for metabolic interactions, String for protein-protein interactions [94] and Yeastract for protein-DNA interactions [95]. The total interaction network contains 6592 genes and 135266 interactions. This interaction network was converted to a probabilistic network using the distribution of the out-degrees of the terminal nodes of the network edges. By doing so, edges connecting nodes with a low out-degree will receive a high probability while edges connecting nodes with a high out-degree receive a low probability. Using lists of mutated genes as input, PheNetic will now infer that sub-network of the probabilistic network that best connects the mutated genes in the list over the probabilistic network. As the probabilistic network penalizes hub nodes, the inferred sub-network will therefore preferentially connect the mutated genes through the least connected parts of the network. This results in selecting the most specific parts of the network that can be associated with the mutated genes.
PheNetic was used to connect the mutated genes over the interaction network [58] with the following parameters: 100-best paths with a maximum path length of 4 were sampled between the different mutations in combination with a search tree cutoff of 0.01. As the size of the selected sub-network by PheNetic is dependent on both a cost parameter and the number of mutated genes in the input, different costs were used for the sub-network inference from different sizes of mutated gene lists. For the sub-network inference between all the mutated genes from the non-mutator reactors a cost of 0.25 was used, for the non-mutator reactors (1,3,4,5) a cost of 0.05 was used as they all have a similar amount of mutated genes, and for the mutator reactors (2 and 6) a cost of 0.5 was used.
The resulting networks were visualized using Cytoscape and a functional enrichment using the biological process terms of Gene Ontology in combination with the annotation of SGD of the sub-networks was performed using the Bingo plugin, version 2.44 [96].
Selected mutations identified from the whole-genome sequencing data were introduced into the ancestral genetic background using the following protocol. First, a selectable marker conferring resistance to hygromycine was introduced near the genomic location of interest through homologous recombination. Part of this locus was then amplified together with the selectable marker using a forward primer containing the desired mutation. The resulting PCR product was then transformed into the ancestral strain; and presence of the mutation was verified by Sanger sequencing.
YECitrine- or mCherry-tagged site-directed mutant strains were competed with the parental mCherry or YECitrine tagged strains, respectively, as described [64]. In brief, saturated cultures of mutant and parental strains were mixed in equivalent volumes and inoculated onto 150 μl of YNB-low fluorescent medium in 96-well microtiter plates (Corning 3585). Micro-cultures grew without shaking and were serial-diluted every 24 hrs for approximately 28 generations (7 days) in a fully automated robotic system (Tecan Freedom EVO200) that integrates a plate carrousel (Liconic STX110), a plate reader (Tecan Infinite M1000), a 96-channel pipetting head, an orbital plate shaker, and a robotic manipulator arm. The equipment was maintained in an environmental room at constant temperature (30° C.) and relative humidity (70%). Fluorescence signal (mCherry: Ex 587 nm/5 nm and Em 610 nm/5 nm; YECitrine: Ex 514 nm/5 nm and Em 529 nm/5 nm) and absorbance at 600 nm were monitored every hour during the entire experiment. The YECitrine- or mCherry-tagged parental strains were competed to each other for normalization and monitored individually to determine background fluorescence signal. Fluorescence and absorbance output data was analyzed in Matlab as described [64] to obtain an average selection coefficient, smut, with its S.E.M. from three experimental replicates.
Selected mutations identified from the whole-genome sequencing data were introduced into the diploid strain Ethanol Red using the following protocol. First, a selectable marker conferring resistance to hygromycin was introduced near the genomic location of interest through homologous recombination. Part of this locus was then amplified together with the selectable marker using a primer containing the desired mutation. The resulting PCR product was then transformed into Ethanol Red; and presence of the mutation was first identified by PCR using 3′ mismatch primer pairs, and subsequently verified by Sanger sequencing. Second, a selectable marker conferring resistance to nourseothricin was introduced into Ethanol Red near the genomic location of interest. Part of this locus with the desired mutation was amplified together with the marker. The PCR product was then transformed into the corresponding mutants, which already has one copy of the target gene mutated and linked to hygromycin resistance marker. The selection of successful transformants was performed on double selection, and the mutants were first identified by PCR, then verified by Sanger sequencing.
Lab-scale fermentations under VHG conditions were started with an overnight pre-growth of a single colony into 3 ml YPD, followed by transfer of the entire culture to 30 ml YP+4% (w/v) glucose and additional growth for 48 h to the stationary phase (200 rpm, 30° C.); 250-ml Schott bottles each filled with 150 ml YP+35% (w/v) glucose and a magnetic rod (35×5 mm) were inoculated to a starting OD600=1.0 (approximately 2.0×107 cells/ml). These bottles were sealed with a waterlock and stirred continuously at 150 rpm on a magnetic stirring platform (IKA® RO 15) at 30° C. The bottles were weighed daily to determine the cumulative weight loss, a proxy for CO2 production and fermentations were stopped after 7 days. The ethanol production was determined using Anton Paar Alcolyzer. A Tukey HSD test was performed to check for significant differences in ethanol production between mutants and their respective controls.
| Number | Date | Country | Kind |
|---|---|---|---|
| 15183670.7 | Sep 2015 | EP | regional |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/EP2016/070732 | 9/2/2016 | WO | 00 |
