Abstract
Hybridization between species often leads to non-viable or infertile offspring, yet examples of evolutionarily successful interspecific hybrids have been reported in all kingdoms of life. However, many questions on the ecological circumstances and evolutionary aftermath of interspecific hybridization remain unanswered. In this study, we sequenced and phenotyped a large set of interspecific yeast hybrids isolated from brewing environments to uncover the influence of interspecific hybridization in yeast adaptation and domestication. Our analyses demonstrate that several hybrids between Saccharomyces species originated and diversified in industrial environments by combining key traits of each parental species. Furthermore, posthybridization evolution within each hybrid lineage reflects subspecialization and adaptation to specific beer styles, a process that was accompanied by extensive chimerization between subgenomes. Our results reveal how interspecific hybridization provides an important evolutionary route that allows swift adaptation to novel environments.
Main
A canonical view of Darwinian evolution asserts that new species arise during extended periods of gradual evolution and selection, in combination with reproductive isolation. However, amendments to Darwin’s original theory have been proposed to incorporate other mechanisms for evolution and speciation, including the emergence of new species through interspecific hybridization1,2,3. Hybridization provides a way to rapidly combine distinct phenotypic features from established populations that converge in an unfamiliar ecological niche. Sometimes, unique combination of phenotypes of a hybrid can enable it to thrive in the new environment and outcompete its parental species4,5,6.
Examples of rapid niche specialization via interspecific hybridization are found across all kingdoms of life. A new hybrid lineage of Darwin’s finches, with competitive beak morphology, emerged in only three generations7. Hybridization of the sunflowers Helianthus annuus and Helianthus petiolaris gave rise to three novel species capable of colonizing previously untapped environments8,9,10. The emerging opportunistic fungal pathogen Candida metapsilosis arose from a single hybridization event between two non-pathogenic parental species11.
The Saccharomyces species complex provides numerous examples of interspecific hybridization. Despite the high sequence divergence between the species within this genus (up to 20% (ref. 12)), prezygotic barriers are weak and species can successfully cross-breed13,14. The resulting hybrids are typically infertile, yet viable, and can reproduce asexually by budding14,15,16,17. Wild Saccharomyces interspecific hybrids are occasionally encountered18,19, but the best-known example, Saccharomyces pastorianus, was isolated from an industrial environment20,21,22,23.
Humans have historically used the capacity of Saccharomyces to convert high concentrations of sugars into ethanol and carbon dioxide to produce a wide variety of fermented products24,25,26. This long-term domestication process has resulted in hundreds of different industrial strains, each with characteristics suitable for specific industrial processes27,28,29,30,31,32,33,34,35,36. Strains used in the production of lager (pilsner) beer generally belong to the species S. pastorianus (S.pas), a hybrid of S. cerevisiae (S.cer) and S. eubayanus ( S.eub)20,21,22,23,37,38. Two archetypical S.pas sublineages exist, dubbed ‘Frohberg’ and ‘Saaz’, each with their own distinct fermentation properties39. Several other industrial interspecific hybrids have been isolated, including S.cer × S. uvarum (S.uva) × S.eub triple hybrids in wine and cider, and S.cer × S. kudriavzevii (S.kud) hybrids in ale beers and wine fermentation vessels40,41,42,43,44,45,46,47.
A comprehensive analysis of the prevalence, molecular details and ecological and evolutionary context of interspecific hybridization in yeasts is lacking. Here, we report how whole-genome sequencing of more than 200 industrial yeasts revealed that a surprisingly large fraction (~25%) proved to be interspecific hybrids derived from four parental species: S.cer, S.kud, S.eub and S.uva. The ubiquity of these hybrids, and the defined industrial environments that they were isolated from, make them an excellent model for studying the role of hybridization in microbial evolution and adaptation. By combining large-scale phenotyping with our knowledge of industrial niches and beer brewing history, we provide evidence that these hybrids originated in industrial environments and are highly niche specific. Additionally, we characterized the genomic changes that occurred during their domestication and describe the genetic mechanism leading to a key domestication phenotype, namely, flavour production. Our results demonstrate that interspecific hybridization is an evolutionary strategy that allows swift adaptation to novel niches and opens new avenues for the development of superior industrial yeasts.
Results
Origins and diversity of Saccharomyces interspecific beer hybrids
For several years, we have been collecting and sequencing yeasts isolated from different industrial niches, including beer, wine, bread, sake, chocolate and liquor fermentations. As expected, the majority of isolates were S.cer, the yeast species most commonly associated with the production of fermented foods and beverages28,48,49. However, our analysis revealed that a surprisingly large fraction (~25%) of sequenced isolates were interspecific hybrids. While some were isolated from lager beer fermentations, known to be driven by hybrid yeasts, many were collected from other beer niches, such as Trappist beers, spontaneously fermented ‘Lambic’ beers and old beer bottles or equipment (Supplementary Table 1). We identified three distinct species compositions across the hybrids investigated: S.cer × S.eub, S.cer × S.kud and S.eub × S.uva. Phylogenetic trees were built to position the hybrid subgenomes in the context of their parental species (Fig. 1 and Supplementary Fig. 1).
Each line of the tanglegram represents one hybrid sequenced in this study, coloured by hybrid type. Species subgenomes within the same hybrid strain are connected between the phylogenetic trees of S.cer, S.eub, S.uva and S.kud. Lineages that do not contain hybrid strains are collapsed (diamonds). Saccharomyces paradoxus (S.par) and Saccharomyces castellii (S.cas) are used as outgroup species. Expanded trees including strain origins are reported in Supplementary Fig. 1.
The S.cer subgenomes of the S.cer × S.eub (S.pas) lager hybrids belong to a well-defined monophyletic clade in the Beer 1 lineage, which mainly contains strains isolated from ale beers28,29,34 (Fig. 1 and Supplementary Fig. 1a). Within Beer 1, the lager subclade is most closely related to Hefeweizen (German wheat beer) strains and the Belgium/Germany lineage, indicating a Western European origin of lager yeasts. The lager clade of Beer 1 is further separated into the archetypical ‘Frohberg’ and ‘Saaz’ lineages22,50,51, both hallmarked by very low nucleotide diversity (π = 1.18 × 10−3 and π = 3.65 × 10−4, respectively). The S.eub subgenomes of the S.cer × S.eub hybrids similarly form a monophyletic group within the previously identified Holarctic clade52 (Fig. 1 and Supplementary Fig. 1b), with defined ‘Frohberg’ and ‘Saaz’ subclades. The Saaz lineage further divides in two subclades, each harbouring one of the two first S.pas strains isolated at Carlsberg in the nineteenth century: Unterhefe nr. 1 (CBS1513) and nr. 2 (CBS1503) (Supplementary Table 1 and Supplementary Note). In accordance with previous reports, the S.eub progenitor(s) of lager yeasts appear to be most closely related to Tibetan S.eub strains, which may have reached Europe through Silk Road trading23,50,53. Interestingly, the origin of domesticated barley (another major beer ingredient) in northwestern Europe was also traced to Tibet54. However, considerable outcrossing and incomplete lineage sorting among Holarctic S.eub strains makes determination of the exact geographical origin of the S.eub ancestor(s) of lager yeasts difficult52.
All but one of the S.cer subgenomes of the S.cer × S.kud hybrids belong to a monophyletic clade, closely related to the industrial yeast clade Beer 2 (Fig. 1 and Supplementary Fig. 1a). Similar to lager yeasts, this ‘traditional Belgian beer’ clade further divides into two subgroups mainly containing hybrids isolated from either Lambic beers (‘spontaneous beer fermentations’; ABI1606 and ABI1525) or Trappist beers (‘Trappist ales’; BE105, BE108, BE109, BE116 and ABI1620) (Supplementary Fig. 1a). One hybrid bread strain (BR005) clusters within this clade, probably reflecting the historical relationship between brewers and bakers55 (Supplementary Fig. 1a). A close association between bread and ale beer strains was previously found among pure S.cer strains as well (‘Mixed’ clade Supplementary Fig. 1a)28. The only S.cer × S.kud strain not belonging to the monophyletic clade is the VIN7 wine strain, whose S.cer subgenome clusters within the Wine clade. This indicates that VIN7 originated independently from the ale and bread hybrids. The phylogenetic relationships of the S.kud subgenomes in the S.cer × S.kud hybrids mirror the S.cer subgenome counterparts (Fig. 1 and Supplementary Fig. 1c).
The phylogenetic structure of the hybrid subgenomes offers insight into the origins of these interspecific hybrids. First, the S.cer subgenomes of the lager beer lineage stem from Beer 1, whereas the traditional Belgian beer lineage forms a sister clade to Beer 2, indicating that both major domesticated beer lineages were involved in the emergence of interspecific beer yeast hybrids. Second, the monophyletic clustering of each beer hybrid type suggests that present-day strains are the result of only one hybridization event per hybrid clade, or few events involving very similar strains53,56,57,58,59. This implies that the present-day diversity is largely due to the spread and diversification of existing hybrids rather than multiple, independent emergence and selection events. Third, the S.cer progenitors seem to come from industrial niches closely associated with the ones from which the hybrids were isolated, suggesting that these successful industrial yeast hybrids formed close to the fermentation environments in which they are now found.
The S.eub subgenomes of the S.eub × S.uva hybrids form a monophyletic sister clade to those of the S.cer × S.eub lager strains, indicating that the S.eub parents were closely related. In contrast to the monophyly of the S.eub subgenomes, the S.uva subgenomes are genetically clearly separated, indicating that at least three hybridization events gave rise to the S.eub × S.uva hybrids (Supplementary Figs. 1d and 2). The S.eub subgenomes clearly separate according to geographical isolation (Germany versus Belgium), suggesting that a single or few closely related S.eub strains formed multiple hybrids, which then evolved and diverged independently. The genetic diversity across the S.uva subgenomes is significantly higher than across the corresponding S.eub subgenomes (average nucleotide diversity π = 2.11 × 10−3 and π = 4.66 × 10−4, respectively, one-sided Mann–Whitney U test, P < 2.2 × 10−16) and there is no clear niche substructure, suggesting that S.uva strains move more freely across environments. Indeed, in contrast to S.eub, S.uva has been isolated all over the world from a wide array of niches, natural and human-made60. Moreover, the species is a known contaminant in brewing environments37,61,62. Coupled with the fact that all of the S.eub × S.uva hybrids were isolated from spontaneous fermentations, old bottles or brewing equipment, it is likely that the hybridizations between S.eub strains and S.uva contaminants occurred within the brewing environment. All S.uva hybrid subgenomes are also closely related to European strains (Supplementary Table 1), and thus the hybrids probably originated in Europe. However, as the population structure of S.eub and S.uva is not described as elaborately as that of S.cer, an Asian origin (as suggested in ref. 23) cannot be formally ruled out.
Interspecific hybrid genomes are hallmarked by extensive ploidy variation and chimeric chromosomes
Newly formed hybrids experience extensive genome reorganization resulting in aneuplodies and chimeric chromosomes63,64. We found extensive variation within and between hybrid types in overall ploidy, copy number of large chromosomal fragments and full chromosomes, as well as in the degree of parental species contribution to the hybrid genome (Fig. 2).
a–d, Ploidy profiles of S.cer × S.kud (a), S.uva × S.eub (b), S.cer × S.eub (Saaz) (c) and S.cer × S.eub (Frohberg) (d) hybrids. Strains are sorted on the basis of the phylogeny of the S.cer or S.uva parental species (see Supplementary Fig. 1). Chromosomes are coloured on the basis of calculated ploidy. Density plots to the left of each tree represent the per-species ploidy distribution aggregated across all the strains of each hybrid type (top panels). Detailed representation of the genomic contribution of the two parental species in one selected hybrid (bottom panels).
The S.cer × S.kud hybrids are overall triploid, with diploid S.cer and haploid S.kud subgenomes (Fig. 2a). Notable exceptions are two Belgian spontaneous beer fermentation strains, ABI1606 and ABI1525, with a total ploidy of roughly 4n (triploid S.cer, haploid S.kud).
The S.uva × S.eub strains are divided into several subgroups with respect to subgenome content (Fig. 2b). The first subgroup, containing a subset of Belgian Lambic strains, exhibits a diploid S.uva subgenome and an extremely fragmented S.eub subgenome. The second subgroup, consisting of German brewing contaminant strains, presents a more uniform 1:1 S.uva:S.eub ratio, with multiple partial or complete chromosome deletions and a few duplicated chromosomal segments. The last subgroup contains two Lambic strains ABIC1571 and ABIC1602 that exhibit an intermediate genome composition (Fig. 1 and Supplementary Fig. 2).
In S.cer × S.eub strains, there is a clear distinction in genomic composition between the Saaz and Frohberg lineages (Fig. 2c,d). Saaz strains are typically triploid (haploid S.cer, diploid S.eub)21,53. Frohberg strains are generally tetraploid to pentaploid with a basal 2n:2n ratio of the parental subgenomes. This high ploidy level is also in line with previously reported Frohberg genomes20,33,53,56,59,64,65. Both types demonstrate severe deletions and amplifications of large segments and even full chromosomes; Saaz strains mostly harbour losses in S.cer and amplifications of S.eub, whereas Frohberg strains demonstrate losses in S.eub and amplifications of S.cer.
Most chromosomal regions exhibit integer ploidy changes, but we did find a few instances of strains within each hybrid type with intermediate ploidy changes (for example, regions of ~0.5 ploidy increments). Given that sequencing was performed on pure culture stocks that underwent a single-cell bottleneck, we would have expected relatively isogenic populations. However, intermediate ploidies suggest that the hybrid genomes are unstable and that the populations used for isolating genomic DNA carried unfixed genomic rearrangements. Interestingly, when further investigating the instability using PCRs targeting regions identified as unstable in four strains (ploidy between 0 and 1), we could only confirm instability in one strain, namely BE138 (Supplementary Fig. 3). In this strain, part of chromosome IV of the S.cer subgenome was present in some clones but absent in others, proving that the hybrid genome is indeed unstable. We did not detect instability of the targeted regions in the other three strains using PCR, which may indicate that the patterns of (partial) chromosome losses are not necessarily the same in different strain subpopulations, such as the subpopulations used for PCR and the subpopulations used for genome sequencing.
Within the same hybrid type, we also observed striking differences in the copy number of full chromosomes and large chromosomal fragments. These copy number changes are often shared by some but not all strains originating from the same hybridization event, and copy number differences occur even among closely related strains, indicating that posthybridization genome structural rearrangements are still ongoing (Fig. 2). In many cases, a copy number change in one subgenome is compensated by an opposite copy number change on the homologous portion of the other subgenome, suggesting the large-scale occurrence of genomic rearrangements leading to chimeric chromosomes and often to loss of heterozygosity.
Mapping of these chimeric regions revealed more than 300 breakpoints amongst the hybrids (72 in S.cer × S.kud, 80 in S.cer × S.eub and 150 in S.uva × S.eub; Fig. 3a–c and Supplementary Figs. 4 and 5a–d). The higher occurrence of breakpoints in S.uva × S.eub hybrids could be due to the lower nucleotide divergence between the two subgenomes (average whole-genome nucleotide identity 91.7%) compared with S.cer × S.kud (84.9%) and S.cer × S.eub (84.5%), which leads to a higher frequency of homologous recombination14,58,59. In fact, DNA sequence homology between subgenomes is significantly higher in breakpoint regions compared with non-breakpoint regions within stretches ranging from 50 base pairs (bp) (microhomology) up to 1 kilobase (kb) (Supplementary Fig. 5e–g).
a–c, Presence (red) and absence (white) of specific chimeric breakpoints are shown for S.cer × S.kud (a), S.uva × S.eub (b) and S.cer × S.eub (c) hybrids. Strains (rows) are sorted phylogenetically according to the S.cer, S.uva and S.cer subgenomes, respectively. Breakpoints (columns) are hierarchically clustered on the basis of their presence or absence across strains. Strains for which low sequencing coverage level negatively affected the detection of breakpoints are indicated with an asterisk. Strain origins are colour-coded per panel.
Breakpoint similarities and differences across the hybrid strains also offer an opportunity to further trace back their origin and evolutionary trajectory during diversification (Fig. 3a–c). Given the larger number of strains and higher coverage of the S.cer × S.eub hybrids, we focused our analysis on this group (Fig. 3c). Out of the 80 identified breakpoints, two are found in all strains. An additional four breakpoints are found across all three subclades (Frohberg, Saaz 1 and Saaz 2), four are found in both Saaz 1 and Saaz 2, two in Saaz 1 and Frohberg and two in Saaz 2 and Frohberg, but the majority (82.5%) are subclade-specific (33 Frohberg, 12 Saaz 1, 21 Saaz 2). The sharing of some breakpoints across the Frohberg and Saaz subclades supports a common origin for all lager yeasts. However, the exact trajectory and relationship of the subclades is difficult to disentangle, given that several breakpoints may have been present at one time but may have been obscured by chromosomal losses (for example, the complete loss of S.cer chromosome XII in all Saaz strains). On the other hand, we cannot exclude the possibility that some of the breakpoint sites are more susceptible to rearrangements (fragile sites), and that shared breakpoints might have arisen independently in different hybrid lineages53,56,59,65.
Hybrid beer yeasts exhibit unique phenotypic features that reflect niche adaptation
Large-scale changes in genome content and structure are intrinsically linked to phenotypic changes, which may confer fitness advantages66,67,68. To assess the extent of phenotypic changes in our hybrids, we extensively phenotyped the sequenced isolates and multiple pure species. Assays covered several industrially relevant traits, including stress tolerances and metabolite production.
On the basis of overall phenotypic behaviour, strains cluster into three major groups, each correlated with a distinct genetic origin and industrial niche (Fig. 4a). Group A contains all pure S.cer strains and S.cer × S.kud hybrids plus one S.cer × S.eub hybrid, which form several subgroups with distinct phenotypic profiles. One of the most distinguishing features between these subgroups is the division between beer-like traits and wine-like traits. Specifically, subgroups A3 and A4 (Trappist S.cer × S.kud hybrids and S.cer ale strains) exhibit weaker environmental stress resistance and sporulation efficiency than subgroups A1, A2 and A5 (strains used in wine, sake, spirits, cider, bioethanol and Lambic brewing). This is indicative of a strong domestication signature (genome decay) in A3 and A4, which is common in ale strains28,29. Group B includes the non-cerevisiae pure species, as well as all but two S.uva × S.eub hybrids. Except for the pure S.kud strains in subgroup B1, these strains demonstrate high cold, osmotic and desiccation tolerance. Group C contains all but two lager hybrids and further subdivides into Frohberg (C1) and Saaz (C3). These can tolerate lower temperatures than strains from Group A. Tolerance to extreme cold (4 °C) is limited compared with strains from Group B. Group C exhibits, overall, a lower environmental stress tolerance than yeasts from non-beer or spontaneous fermentation environments (primarily in Groups A and B).
a, Hierarchically clustered heat map of phenotypic diversity within interspecific hybrids and pure species. Phenotypic values are calculated as normalized z-scores. Missing values are shown in grey. Phenotypes (rows) are sorted on the basis of five major categories (labelled on the left). (Sub)genome compositions are indicated at the branch tips and coloured by species. YPD, yeast extract, peptone, dextrose; NA, not available. b, Correlation of cold tolerance versus maltotriose use of hybrids (circles) and pure species (triangles). c, Performance of artificial interspecific hybrids and parental species in beer wort fermentations at cold temperatures. Three independent crosses of the same parents were performed (H1, H2, H3). Bars indicate mean ± s.d. of four biological replicates. Statistical significance determined by ANOVA between hybrids and the inferior pure species: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. d, Acid tolerance of S.cer × S.kud hybrids (circles) and pure species (triangles). Hybrids are coloured according to isolation origin.
Cold tolerance (a common trait in non-cerevisiae pure species) and the ability to efficiently use maltotriose (a beer-specific carbon source fermentable by many S.cer strains but typically not by non-cerevisiae species) are commonly combined in interspecific hybrids (Fig. 4b). This combination has been proposed as the reason why these hybrids persisted and flourished in cold brewing environments39,69,70, and could be either the result of posthybridization adaptation or a direct consequence of the hybridization process itself. To directly evaluate these possibilities, we phenotyped several newly developed artificial S.cer × S.eub hybrids and compared them with their parental strains (Fig. 4c). The artificial hybrids demonstrated improved cold tolerance compared with the S.cer parent (analysis of variance (ANOVA) F = 8.88, P < 0.001) and improved maltotriose use compared with the S.eub parent (ANOVA F = 75.53, P < 0.001), showing that hybridization can generate immediate fitness advantages in niches, such as cold beer fermentation environments.
Although hybridization can impart immediate fitness advantages, the instability of interspecific hybrids probably facilitates further adaptation to specific beer niches, best exemplified by the ‘traditional Belgian beer’ lineage. This lineage probably originated from a single hybridization event (see Fig. 1) and subsequently split into two distinct subgroups, Lambic and Trappist strains, which function in two profoundly different beer production processes. During Lambic beer production, the presence of acid-producing bacteria leads to considerable concentrations of acetic and lactic acid71,72, whereas in Trappist beer production, these bacteria are much less prominent. Despite their shared origin, tolerance to organic acids is observed in the Lambic, but not in the Trappist subgroup, suggesting adaptation after the hybridization event (Fig. 4d).
Structural rearrangements underlie phenotypic convergence to low off-flavour production in lager yeast
Beer flavour and aroma diversity can often be attributed to the yeast (and, in some cases, bacteria) used during the fermentation process73. The style of beer dictates which specific aroma compounds are desirable and has thus influenced selection over time. The presence of 4-vinyl guaiacol (4-VG) is tolerated in some specialty beers, such as wheat and saison beers, but is undesirable in most other beer styles, including lagers. This compound produces a spicy, clove-like aroma referred to as ‘phenolic off-flavour’ (POF). Production of 4-VG depends on functional copies of two subtelomeric genes, phenylacrylic acid decarboxylase (PAD1) and ferulic acid decarboxylase (FDC1)74.
Our phenotype analysis demonstrated that all non-cerevisiae pure strains (S.eub, S.kud and S.uva) produce 4-VG (POF+) (Fig. 5a). As a dominant trait, newly formed hybrids with a non-cerevisiae parent should be POF+ regardless of the phenotype of the S.cer parental strain. However, in contrast to all S.uva × S.eub and S.cer × S.kud hybrids, all S.cer × S.eub hybrids were found to be POF−(Fig. 5a), suggesting that 4-VG production was lost after hybridization in lager strains. To understand the genetic foundation of this loss, we further evaluated the subgenomes of the lager hybrids.
a, Percentage of strains from each species or hybrid type that demonstrate production of 4-vinyl guaiacol (4-VG+). b,c, The genetic basis for loss of 4-VG production is depicted for the S.cer subgenomes (b) and S.eub subgenomes (c) arranged by hybrid subgroup (shown sorted phylogenetically on the left). The subtelomeric position of the PAD1–FDC1 gene cluster (red triangle) is shown on S.cer chromosome IV and S.eub chromosome 13. Black arrows indicate the location of chimeric breakpoints between the homologous chromosomes. Chromosomes are coloured according to the ploidy of one representative strain (Saaz group 1: BE137; Saaz group 2: CBS1538; Frohberg: BE104). The loss-of-function mutation in the Frohberg S.cer subgenome is highlighted in the tan triangle.
The S.cer subgenomes of these strains localize within the Beer 1 clade; the majority of strains from this clade acquired disruptive mutations in PAD1 and/or FDC1 (refs. 28,29,75), suggesting that the hybrids inherited an inactive POF pathway from their S.cer ancestor. Our analysis shows that Frohberg strains do indeed harbour the same disruptive mutation(s) as other Beer 1 S.cer strains (Fig. 5b). Saaz strains, however, harbour complete deletions of the S.cer PAD1–FDC1 gene cluster (Fig. 5b). Nevertheless, the hybrids must also have inherited an inactive POF pathway from the S.eub parent to render them POF−. Three distinct genomic changes resulted in the loss of the PAD1–FDC1 gene cluster in the S.eub subgenomes of the Saaz 1, Saaz 2 and Frohberg lineages (Fig. 5c), each involving a different chimeric breakpoint between the terminal regions of S.cer chromosome XIII and S.eub chromosome 13. Thus, each lineage experienced an independent loss of heterozygosity event, further supporting a posthybridization loss of 4-VG production, probably after the divergence of Frohberg and Saaz. Interestingly, this was the only region in the S.eub subgenome that was lost three times independently in lager strains, signifying the strong selection against the POF phenotype.
Historical context of the origin and evolution of lager yeasts
Our data offer a unique opportunity to map evolutionary events in a historical perspective and to gain insight into the potential driving forces behind yeast hybrid domestication. Absolute dating of evolutionary events in yeasts is complicated due to a lack of solid calibration points (for example, fossils) and the difficulty of tracing historical movements of industrial yeasts. Here, we used the split between UK and US S.cer strains in the Beer 1 clade, which is thought to have occurred between ad 1607 and ad 1637 when British settlers imported beer yeasts to North America28, as a calibration point to date divergences within and between the ale and lager subclades of Beer 1 (see Methods).
This calibration yields a timeline for beer yeast evolution that correlates with known historical events (Fig. 6 and Supplementary Note 1). We estimate that the most recent common ancestor of present-day ale and lager beer strains dates to the mid-sixteenth century (Fig. 6). The lager yeast lineage splits from the Belgium/Germany clade within the same timeframe, shortly after the enactment of the Bavarian Beer Purity Law (Bayerisches Reinheitsgebot) in ad 1516. This law, historically, had a large, reductive impact on the variety of German beer styles, as many local beers and brewing traditions disappeared to conform with the law. The primary objective was to reduce bacterial contaminations by establishing rules regarding ingredients and restricting brewing to the colder winter months. Cold-temperature brewing probably inadvertently selected for cold-tolerant strains—a trait characteristic of S.eub. This German origin of lager yeasts ties up with historical brewing records, which allowed us to trace the origins of Saaz and Frohberg yeasts to Bavaria, a region in the southeast of Germany. A more elaborate discussion of lager beer yeast history can be found in Supplementary Note 1.
The five subclades of Beer 1 are indicated to the right of the tree. Shaded vertical boxes highlight major splits within the S.cer Beer 1 clade: split of present-day ale and lager yeasts (yellow); split of European and US ale yeasts (green); split of Saaz and Frohberg lager yeasts (light blue). Branches are coloured according to their branch-specific average number of substitutions per site per year. Node bars represent 95% highest posterior density intervals. Dates at the top indicate relevant events in lager beer history, which are listed in the historical date box on the right (see Supplementary Note 1).
A notable difference between the lager and ale clades is the sudden, dramatic reduction in evolutionary rates of the lager lineages approximately 150 yr after the split between the Saaz and the Frohberg lineages. The rate decrease correlates with the onset of diversification of both lineages, suggesting that the limited genetic diversity observed amongst today’s commercial lager strains is a consequence of a genetic bottleneck and subsequent slow, local divergence at the turn of the twentieth century. This period coincides with three important events that revolutionized (lager) beer production. First, Louis Pasteur described the importance of yeast in alcoholic fermentations in ad 1857 (ref. 76). Once brewers became aware of the true significance of their sediment and the possible economic implications, greater care was taken to maintain their brewing cultures. Second, mechanical refrigeration was introduced into breweries from ad 1873 onwards55,77. This technological advance allowed year-round lager fermentations, as well as cooled storage of successful yeast cultures. Many brewers adopted the practice of keeping a separate, refrigerated population of their yeast, from which they regularly re-grew a larger population to inoculate a new brew. Additionally, mechanical cooling also allowed brewing at colder temperatures, fostering selection for cold-tolerant strains. Colder fermentation temperatures and cold storage of yeast cultures partly explains the observed reduced evolutionary rate. Third, the isolation of the first pure yeast culture in ad 1883 (which was later shown to be a S.cer × S.eub hybrid) by E. C. Hansen at the Carlsberg brewery (Denmark) gave rise to a wave of brewers isolating and sharing their strains77. The first isolated culture (Unterhefe nr. 1) was disseminated across many breweries in central Europe, and rapidly implemented in their production processes78. Together, these three events led to standardization of industrial lager production and probably only a few closely related lager hybrid yeasts were disseminated across different breweries, where they were mostly preserved and used in cooled environments, slowing down their evolutionary divergence and further enhancing the selection for cold tolerance.
Discussion
Industrial, human-made environments challenge microbes with unique environmental conditions and therefore offer insight into the processes that allow colonization of novel ecological niches. Our results show that interspecific hybridization is an important and common route towards diversification and adaptation to novel niches. Moreover, our genome analyses and phenotyping suggest that adaptation is probably fuelled both by the direct, selective advantages of the new hybrids, as well as their genomic plasticity, allowing for swift adaptation to specific niches. Beer yeasts provide a perfect example of this phenomenon, as interspecific hybridization yielded new variants that combine the fermentation capacity of S.cer with the cold tolerance of other species. The instability and plasticity of the hybrid genomes probably allowed further adaptation and yielded variants that lost undesirable properties, such as the production of 4-vinyl guaiacol.
Our discovery of a large proportion of interspecific hybrids in Belgian specialty beers suggests that interspecific hybridization played an important role in the history of industrial fermentations. Over the past 200 yr, adoption of new brewing technologies and highly controlled single-strain fermentation processes probably contributed to the decline of the natural beer yeast diversity, including interspecific hybrids that may have been present as low-number contaminants of most ale fermentations. The low diversity of present-day lager yeasts exemplifies this trend. Although the first S.cer × S.eub hybrids originated around the sixteenth century, today’s lager yeasts can be traced back to only a few lineages that were pure-cultured, cold-stored and dispersed across multiple breweries in the late nineteenth century. Belgium represents a notable exception to this biodiversity decline. Traditional Belgian beer styles harbour a remarkably diverse array of yeasts, probably as a result of the continued use of old beer brewing practices. For example, the production of Belgian Lambic beers has remained unchanged for centuries. The use of medieval brewing technologies, such as open-air inoculation and long-term fermentation and storage in barrels stored in the brewery’s cellar, promoted the survival of the unique yeast hybrids presented in this study, and could therefore provide a source of new biodiversity for industrial applications.
The study of the genetic and phenotypic makeup of naturally occurring yeast hybrids may further aid in the development of innovative new hybrids for industry. Recent studies have shown that recreation of interspecific hybridization events that occurred in nature, or development of new combinations of species, results in hybrids with a phenotypic landscape beyond that of the strains usually employed in industrial fermentations63,79,80. Some of these hybrids produce new aromas that cater to changing consumer demands and demonstrate superior performance in challenging production environments. Detailed insight into how these hybrids evolved in different niches, and which phenotypic features are retained from the parental species, can aid in selecting suitable strains for development of industrially relevant hybrids.
Methods
Genomic DNA extraction
For strains BE114, BE116–BE130, BE132–BE136 and SP012, genomic DNA was extracted with the MasterPure Yeast DNA Purification Kit (Epicentre). For the other strains, genomic DNA was prepared using the Gentra Puregene Yeast Kit (Qiagen) with some modifications to the recommended protocol. The main modification involves a 2-h treatment of the overnight cell culture with zymolyase to efficiently digest the yeast cell wall. Final DNA concentrations were measured using Qubit (Thermo Fisher Scientific) with absorbance ratios of 260/230 and 260/280 with Nanodrop (Thermo Fisher Scientific).
Library preparation and whole-genome sequencing
For strains BE114, BE116–BE130, BE132–BE136 and SP012, paired-end sequencing libraries were prepared using the Nextera XT DNA Library Preparation Kit. Sequencing was performed on a HiSeq 2500 system (at Illumina). For the other strains paired-end sequencing libraries (MiSeq reagent kit v.3, 600 cycles) with a mean insert size of ~300 bp were prepared and run according to the manufacturer’s instructions on an Illumina MiSeq at the Nucleomics Core facility in Leuven (http://www.nucleomics.be/).
Reference-based alignments and variant calling
Reads were preprocessed by filtering low quality and ambiguous reads, adapters and PhiX contaminations, using Trimmomatic (v.0.30). Clean reads were mapped to seven Saccharomyces species (Saccharomyces species complex): S.cer reference genome S288c (R64-1-1, EF4-Ensemble Release 74), S. paradoxus (S.par) (strain YP138)32, S. mikatae (strain IFO 1815)81, S.kud (IFO 1802)81, S. arboricola (strain H-6, NCBI: txid1160507)82, S.eub (strain FM1318)56 and S.uva (CBS 7001)81, with the Burrows–Wheeler Aligner (BWA v.0.7.17)83 using default parameters except for −q 10. Non-primary alignments were filtered out and duplicate reads were marked using Picard Tools (v.1.56) (http://picard.sourceforge.net). Coverage was estimated on the basis of read depth in non-overlapping 1-kb windows (reported mean coverage per window) using BEDtools (v.2.27.0)84.
De novo assembly
For each library, low quality and ambiguous reads were trimmed using Trimmomatic (v.0.30). Reads were error corrected and subsequently used for de novo assembly with SPAdes (v.3.10.1)85. Next, the Redundans pipeline (v.1.2)86 was used to detect and remove redundant contigs and perform scaffolding and gap closing on the basis of paired reads information. To determine the coordinates of contigs from each newly assembled strain relative to the reference strain, and to obtain pseudo-chromosomes, whole-genome alignments were performed against the species identified in the section ‘Reference-based alignments and variant calling’ using Ragout (v.1.2)87.
Phylogenomic analyses
To infer the origin of the hybrid Saccharomyces genomes and their genetic relationship across species and strains within the Saccharomyces species complex, genes that are orthologues and present in exactly one copy among strains and across species have been identified (single-copy orthologues). The starting set of genes included 4,722 1:1:1:1 orthologues among S.cer S288c, S.par, S.kud IFO 1802 and S.uva CBS 7001 identified by Scannell et al.81. The starting set of genes was reduced to 4,125 genes after including S.eub FM1318 (ref. 56) and the outgroup species S. castellii88. Next, the presence of these genes and their single-copy status was tested within a collection of 420 Saccharomyces isolates. The collection of strains investigated included 283 S.cer isolates, 3 S.kud isolates, 43 S.uva isolates, 21 S.eub isolates, 10 S.cer × S.kud hybrids, 46 S.pas hybrids and 13 S.uva × S.eub hybrids (Supplementary Table 1). From this step onward, the subgenomes of the hybrid isolates were considered as distinct species: for instance, for a S.uva × S.eub hybrid, two sets of orthologues were identified, one for each species, respectively. A local BLAST database was set up for all the genomes on the basis of their de novo assembly (collapsed representation of each species subgenome) and BLASTN searches were performed (1 × 10−4 E-value cut-off, >98% similarity and >85% coverage with BLAST v.2.5.0+)89,90 using, for each species, the set of genes identified in the previous step. For hybrid genomes an additional BLASTN step was implemented to compare the set of genes identified for the distinct parental species and to exclude genes with high similarity between species that cannot be unequivocally assigned to one or other species. Five sets of genes were obtained: (1) 1,389 genes across the S.cer genomes and subgenomes, dubbed the ‘S.cer’ set; (2) 1,571 genes across the S.eub genomes and subgenomes (‘S.eub’set); (3) 1,364 genes across S.uva genomes and subgenomes (‘S.uva’ set); (4) 1,750 genes across the S.kud genomes and subgenomes (‘S.kud’ set). Considering the high level of species-specific subgenome loss and fragmentation observed in some Saccharomyces hybrids, and to maximize the number of isolates included, 3% missing data per gene was allowed. Extreme cases, with >50% missing genes per strain were excluded from the analysis (only ~200 S.eub genes could be annotated for ABI1605 and it was therefore excluded from the S.euba phylogeny). Multiple nucleotide sequence alignments (MSAs) for each gene in each set identified were generated using MAFFT (v.7.187)91, with default settings and 1,000 refinement iterations. The MSAs were concatenated into supermatrices for each species using FASconCAT (v.1.0)92. Quality checks and format conversions were performed using trimAl (v.1.2)93. The final S.cer supermatrix included 337 taxa and 1,556,065 positions, 97.345% nucleotides, 2.655% gaps and 0% ambiguities. The final S.eub supermatrix included 81 taxa and 2,323,546 positions, 88.141% nucleotides, 11.859% gaps and 0% ambiguities. The final S.uva supermatrix included 55 taxa and 1,453,393 positions, 91.772% nucleotides, 8.228% gaps and 0% ambiguities. The final S.kud supermatrix included 14 taxa and 2,635,158 positions, 90.741% nucleotides, 9.259% gaps and 0% ambiguities. Within each supermatrix, each gene was considered as a separate data partition. Twenty-five completely random starting trees for the S.cer supermatrix and 20 random starting trees for the S.eub, S.uva and S.kud supermatrices were obtained using RAxML (v.8.2.8)94. Maximum-likelihood (ML) tree searches were performed on each fully random starting tree under the GTRGAMMA model (four discrete rate categories) using ExaML (v.3.0.17)95 and the rapid hill climbing algorithm (-f d). During the ML search, the alpha parameter of the model of rate heterogeneity and the rates of the GTR model of nucleotide substitutions were optimized independently for each partition. The branch lengths were optimized jointly across all partitions. For each starting tree, the best tree was selected on the basis of the highest log-likelihood score. Parameters and branch lengths were re-optimized on the best topologies with ExaML (-f E) using the median of the four rate categories for the discrete approximation of the GAMMA model of rate heterogeneity (-a). The tree with the best overall log-likelihood score of all tree inferences was considered the final ML tree. Non-parametric bootstrap analysis was performed on the concatenated matrices using RaxML. The a posteriori boot-stopping criterion96 (MR bootstrapping convergence criterion) was applied to define the number of replicates. After every 50 replicates, the set of bootstrapped trees generated so far was repeatedly (1,000 permutations) split in two equal subsets, and the weighted Robinson–Foulds (WRF) distance was calculated between the majority-rule consensus trees of both subsets (for each permutation). Low WRF distances (<3%) for ≥99% of permutations were used to indicate bootstrapping convergence. Convergence was reached after 200 replicates for the ‘S.cer’ phylogeny: average WRF = 2.04%, percentage of permutations in which the WRF was ≤3.00 was 99.8%; 200 replicates for the ‘S.eub’ phylogeny: average WRF = 1.40%, percentage of permutations in which the WRF was ≤3.00 was 99.7%; 600 replicates for the ‘S.uva’ phylogeny: average WRF = 1.36%, percentage of permutations in which the WRF was ≤3.00 was 99.2%; 50 replicates for the ‘S.kud’ phylogeny: average WRF = 0.26%, percentage of permutations in which the WRF was ≤3.00 was 100%. The final trees were visualized and rooted in R (v.3.4.1)97 with the ggtree package (v.1.8.2)98 using S.par as outgroup for the ‘S.cer’ tree and S. castellii for the other trees.
Divergence time estimation
We used BEAST (v.1.10)99 to estimate divergence times in the Beer 1 clade of the S.cer phylogenetic tree, using the topology of best scoring ML S.cer tree obtained for a supermatrix of 1,389 protein-coding genes, as described in the previous section. Mosaic strains, harbouring mixed genetic backgrounds (for example, Hefeweizen isolates), were excluded from the analysis (population structure analysed with fastStructure v.1.0 (ref. 100)). To date the phylogeny we used a calibration prior on the split between North American and British beer yeasts, using a normal prior with a 99% confidence interval falling between ad 1607 and ad 1637, based on historical events28. We assessed the performance of several molecular clock models using BEAST (v.1.10)99 in combination with BEAGLE 2.1.2 (ref. 101). Specifically, we analysed the data using a strict clock, an uncorrelated relaxed clock with an underlying log-normal distribution102, a random local clock103 and a fixed local clock with predefined clades (Lager–Frohberg, Lager–Saaz, Ale–Britain, Ale–Belgium/Germany, Ale–US)104. A pure-birth Yule speciation prior and a random starting tree were used for the Bayesian inference analyses through Markov chain Monte Carlo. Each analysis was run until effective sample size values of at least 100 could be obtained for all relevant parameters, as computed by Tracer v.1.7 (ref. 105). Of these models, the random local clock provided a significantly better fit to the data than the three competing models, as estimated using generalized stepping-stone sampling106. The uncorrelated relaxed clock model yielded the lowest model fit to the data of the clock models tested, indicative of evolutionary rate shifts having occurred in the S.cer tree, which the single rate distribution in the uncorrelated relaxed clock model is not able to account for. Additionally, the relaxed clock assigns a unique rate to every branch of a tree, but changes in the rate of evolution do not necessarily occur smoothly nor on every branch of a tree. Assuming strict clock rates within predefined clades allows capture of (major) shifts in evolutionary rates between those clades, but does not allow for any rate variation within each clade. The random local clock on the other hand allows sampling of the state space of all possible (strict) local clock models on all possible rooted trees, and concluded that an estimated 50 rate changes occurred throughout the tree. Given that 50% of the prior probability within the random local clock model assumes no rate changes (and over 95% prior probability of less than three rate changes occurring), this shows that there is a strong signal in the data in favour of a large number of rate changes (more than those at the predefined clades in the fixed local clock model), which provides an additional argument for the random local clock significantly outperforming all other models. Given these findings, we present results for the random local clock model only, by summarizing its divergence time estimates in a maximum clade credibility tree using TreeAnnotator, which is part of the BEAST software package99.
Estimation of ploidy and identification of chimeric regions across species
From the alignment of paired-end reads on the multispecies reference genome, a new alignment file was generated for each strain by retaining paired reads with high mapping quality (q > 20), for which the two mates are mapping on chromosomes of the different subgenomes. For example, for read pair A, one read is mapped on S.cer chromosome I and its mate on S.eub chromosome 1. We dubbed these reads ‘discordant reads’. The absolute number of discordant reads over the total number of reads was calculated in non-overlapping 1-kb windows over the full multispecies reference genome, excluding unplaced contigs and mitochondrial contigs. To identify windows in the genome supporting the presence of a chimeric event, we selected windows with at least 15 reads and a minimum of 3 discordant reads. This very conservative threshold allows the identification of potential chimeric events across areas of the genome with varying coverage levels within the same hybrid type and across hybrids with different genome size, and hence varying coverage levels. Windows with a minimum of 3 discordant reads were defined as ‘breakpoint’ windows; breakpoint windows preceded and/or followed by another breakpoint window were defined as ‘major breakpoint’ windows, because of the presence of consecutive windows supporting the chimeric event. Since breakpoint windows often coincided with changes in ploidy, we simultaneously calculated the ploidy level and the occurrence of discordant reads in each 1-kb window along the genome. First, the mean raw coverage per 1 kb non-overlapping window calculated with BEDtools (v.2.27.0)84 was smoothed using a running median function on windows of 11 consecutive 1-kb windows using the CaTools package (v.1.6, https://CRAN.R-project.org/package=caTools) in R. Second, the density of the smoothed coverage was plotted per subgenome and for the full hybrid genome. To identify the mean value of smoothed coverage corresponding to the haploid ploidy level, a Gaussian mixture model was fitted to the density distribution and the mean and standard deviation were calculated for each peak (peaks are a proxy for ploidy levels detected in the hybrid genome). Third, the smoothed raw coverage values were divided by the mean value of the haploid peak to obtain estimated ploidy values for each precomputed window in the genome. Due to the presence of coverage depth noise and potential copy number heterogeneity in the population of sequenced cells, ‘in-between’ (noninteger) ploidy levels were detected. To define integer ploidies and detect ploidy shifts, a second Gaussian mixture model was fitted on the distribution of estimated ploidy values for each strain. This allowed the identification of the mean and standard deviation of each ploidy peak and the definition of ‘ploidy shift boundaries’ on the basis of the intersection points between the ploidy modes in the mixture model. The identification of breakpoint windows and their ploidy context was then followed by identification of the exact mapping location of the reads within the breakpoint window across the two subgenomes for each hybrid strain, on the basis of the following steps: (1) extract the reads from the selected breakpoint windows; (2) retain reads that are still paired after the identification and selection of breakpoint windows; (3) calculate for each pair their summed edit distance from the corresponding reference sequence, normalize it by the summed length of the two reads; (4) select reads with a percentage identity ≥95% against the corresponding reference location; (5) intersect the position of the reads with annotated features from the corresponding reference sequence (Supplementary Table 2). To calculate nucleotide percentage identity (%) between subgenomes for breakpoint windows and non-breakpoint windows, pairwise whole-genome alignments were obtained between Saccharomyces species corresponding to the species combination identified in our set of hybrids using Mugsy (v.1.2.3)107. Next, the 1-kb interval coordinates from the alignments against the reference sequences were mapped on the whole-genome alignments, and nucleotide percentage identity was calculated as number of matches on the total amount of bases in the window, excluding gaps. We repeated the calculations for 500, 100 and 50-bp windows to identify microhomology regions within the starting 1-kb windows. Then, we compared the identity distribution across breakpoint windows and non-breakpoint windows. The nucleotide percentage identity distribution across breakpoint windows and non-breakpoint windows was compared using the Wilcoxon signed rank test implemented in the MASS package (v.7.3-47)108 in R. Twenty-one strains were excluded from the ploidy estimation analysis due to a bias in the read depth profile already observed and described in Gallone et al. and referred to as ‘smiley pattern’: for these samples, coverage follows a convex trend with high depth at the terminal regions of the chromosomes that gradually decreases toward the centre (see Supplementary Table 1 and ref. 28).
Investigation of population heterogeneity using PCR
To investigate whether the population heterogeneity observed in the ploidy profiles of some of the hybrids is caused by unfixed genomic rearrangements and losses in the populations concerned, we monitored the presence or absence of unstable regions, that is, regions that exhibit a calculated ploidy level between 0 and 1, in populations of four strains (BE137, BE138, ABI1620, BR005; 1 region per strain). For each unstable region, we also tested a stable region in its proximity for which no heterogeneity was observed. For each strain, we assessed 45 individual, randomly picked colonies. First, strains were streaked from the −80 °C stock to standard agar plates (YPGlu 2% agar; yeast extract 1% w/v, peptone 2% w/v, glucose 2% w/v, agar 2% w/v) to single colonies. After a 2-d incubation at room temperature, the 45 random colonies were selected, and genomic DNA was extracted using a 10-min boil in NaOH. PCR to assess the absence or presence of the target regions was performed (all primers are provided in Supplementary Dataset 1). The absence of a PCR product is indicative of loss of the region.
Flavour production and flocculation in fermentation conditions
To assess the metabolite production of the yeasts, laboratory-scale fermentation experiments were performed. These fermentations were performed in rich growth medium (YPGlu 10%; peptone 2% w/v, yeast extract 1% w/v, glucose 10% w/v) and beer wort (13 °P, 8 EBC Brewferm). Precultures were inoculated in test tubes containing 5 ml of yeast extract (1% w/v), peptone (2% w/v) and glucose (4% w/v) medium (YPGlu 4%) and incubated overnight at 30 °C (shaking). After 16 h, the preculture was diluted tenfold in 50 ml of YPGlu 4% medium and transferred to 250-ml Erlenmeyer flasks. This second preculture was incubated for 16 h at 30 °C (shaking). Next, the preculture was used to inoculate the growth medium at an initial optical density at 600 nm (OD600) of 0.5 (roughly equivalent to 107 cells ml−1). The fermentations were performed in 250-ml Schott bottles with a water lock placed on each bottle. They were incubated for 7 d at 20 °C (YPGlu 10%) or 16 °C (beer wort), statically. In addition, H2S production during the fermentation was tracked using a lead acetate strip, which was scored from 0 (no colour reaction, white strip) to 3 (intense colour reaction, black strip) after the fermentation, to quantify H2S formation. To estimate fermentation progress, weight loss was measured daily. After 7 d, the fermentations were stopped, filtered (using 0.15 mm paper filter) and samples for chromatographic, density, spectrophotometric and ethanol measurements were taken. Maltotriose use (%) in beer wort was calculated by comparing the total weight loss of the fermentation to the theoretical maximum (calculated on the basis of total fermentable sugar concentrations). Additionally, after fermentation, the flocculation character of each strain was scored visually, using a score ranging from 1 (not flocculent) to 6 (extremely flocculent, big flocs).
Headspace gas chromatography coupled with flame ionization detection (Agilent Technologies) was performed as described previously28.
Acetic acid, sulfite, pH and glycerol production were analysed via the Gallery Plus Beermaster Discrete Analyzer (Thermo Fisher Scientific), according to the manufacturer’s recommendations.
Ethanol accumulation capacity
The maximal ethanol accumulation capacity of all strains was assessed as described previously28.
Screening for environmental and nutrient stress tolerance
All strains were tested in several conditions using robot-assisted spotting assays. All strains were evaluated on YPGlu 2% agar for (1) temperature tolerance (4 °C, 12 °C, 16 °C, 30 °C, 37 °C, 39 °C), (2) sugar and/or osmotic tolerance using increasing concentrations of glucose (final osmolyte concentration of 44, 46, 48% w/v), (3) acid tolerance using increasing concentrations of acetic acid (12.5, 25, 50, 75 mM), (4) ethanol tolerance using increasing concentrations of ethanol (5, 7, 9, 10, 11% v/v), and (5) copper tolerance using 0.050, 0.075, 0.100 mM of copper. For each of these experiments, growth on YPGlu 2% agar at 20 °C was used as a reference condition.
Spotting assays and image analyses were performed as described previously28. Heat maps were obtained using the R function heatmap.2 from the gplots package (v.3.0.1)109. Strains were hierarchically clustered on the basis of phenotypic behaviour using the ward.D2 method110 on Euclidean distances.
Investigation of the sexual life cycle of the yeast
Sporulation was induced on minimal sporulation medium (1% (w/v) KAc, 0.05% (w/v) amino acids, 2% (w/v) agar). After pregrowth in 5 ml YPGlu 2% (overnight at 20 °C, shaking), strains were incubated at 23 °C for 10 d. Dissection of four tetrads of each strain was carried out using a micromanipulator (Singer Instruments), and mating-type determination of all segregants was performed by mating-type PCR.
Yeast survival in beer
To assess survival of all strains, the viability of cultures ageing in Duvel Green (blond ale, 7% (v/v) ethanol) was tracked over a 1-month period. Yeast precultures were shaken for 48 h at 30 °C in test tubes containing 5 ml YPGlu 2%. This sample was then transferred to 15 ml Falcon tubes and centrifuged for 3 min at 3000 r.p.m. After discarding the supernatant, the samples were resuspended in sterile water to reach an initial cell count of 4 × 107 cells ml−1. This sample (0.5 ml) was used to inoculate a sterile gas chromatography vial containing 10 ml of filter-sterilized Duvel Green supplemented with 0.2% w/w glucose to reach a final concentration of 2 × 106 cells ml−1. The headspace of the samples was flushed with CO2 before capping the vials. The vials were incubated statically at 30 °C. After 30 d, samples were taken, and viability was assessed using a methylene blue (0.1% (w/v), Sigma-Aldrich) staining. Automated cell counting was performed with the TC20 automated cell counter (Bio-Rad).
Desiccation tolerance
Desiccation tolerance was measured by a modified version of the assay described in Calahan et al.111. All strains were streaked on YPGlu 2% agar. Strains were pregrown in YPGlu 2%, supplemented with 0.5% v/v Tween 80 and 20 µg ml−1 ergosterol in deep-well plates at 16 °C for 3 d. Cells were harvested by centrifugation (3000 r.p.m., 3 min) and supernatant was removed. Cells (108) were resuspended in 1 ml assay buffer (8× dilution of phosphate buffered saline: 17.1 mM NaCl, 0.338 mM KCl, 1.25 mM Na2HPO4, 0.220 mM KH2PO4, pH 7.4), after which viability was checked using plating. From this cell culture 100 µl was transferred to a 96-well microtitre plate. The lid of the 96-well microtitre plate was lifted 1.7 cm (using four small pieces of cardboard) during the experiment, to allow sufficient air flow. Plates were incubated for 7 d at 8 °C. Afterwards, cells were resuspended in 200 µl assay buffer and viability was assayed using plating. Next, the ratio between the viability after desiccation and the viability before desiccation was calculated. These data were subsequently binned to score desiccation tolerance (viability < 1% scored as ‘1’, viability of 1–20% scored as ‘2’, viability 20–50% scored as ‘3’, viability >50% scored as ‘4’). All strains were tested in biological duplicates.
Biomass production in cold wort
To assess biomass production in beer wort at very low temperatures strains were pregrown in 100 µl of YPGlu + Mal 2% (yeast extract 1% w/v, peptone 2% w/v, glucose 1% w/v, maltose 1% w/v) at 16 °C for 2 d. Next, 5 µl of these cultures was transferred to 95 µl wort (8 °P, 8 EBC Brewferm) in a 96-well microtitre plate and incubated at 8 °C for 7 d on a microtitre plate shaking platform (Heidolph Instruments) at 600 r.p.m. Optical density of the strains was assessed after 0, 3 and 7 d. Experiments were performed in duplicate.
Development and phenotypic evaluation of artificial hybrids
Hybridization was induced by placing single spores from both parental strains together with a micromanipulator (Singer Instruments) on YPGlu 2% agar, followed by visual inspection of zygote formation after 6–8 h of incubation at room temperature. Candidate interspecific hybrids were purified by streaking on wort agar medium (12% w/v malt extract (8 EBC Brewferm) and 1.5% w/v agar). Hybrids were confirmed through a species PCR79,112,113. PCR-confirmed interspecific hybrids were streaked another three consecutive times on 12 °P wort agar before long-term storage at −80 °C to ensure strain purity. Three independent hybrids were developed from the same two parents (BE014 and WL024). Ploidy investigation of the hybrids showed that all hybrids were triploid, which is expected given the ploidy of the parental strains, which were diploid (WL024) and tetraploid (BE014). This results in segregants that are haploid and diploid, respectively.
These hybrids were tested for biomass during wort fermentation at cold temperature as described in the previous section. The maltotriose that was still present after 7 d of fermentation was measured using the Dionex system (ICS 5000+, Thermo Fisher Scientific). Fermentation samples were diluted in dH2O, filtered (0.2 µm), and 10 µl was injected into the system. A Carbopac PA20 column, kept at 30 °C, was used, with a column flow of 0.3 µl min−1. NaOH (250 mM, Eluent 1) and 500 mM CH3COONa + 100 mM NaOH (Eluent 2) were used as eluents.
Reporting Summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
Data availability
All raw sequencing reads generated in this study have been deposited to the NCBI Short Read Archive (BioProject accession: PRJNA516557).
References
Mallet, J., Besansky, N. & Hahn, M. W. How reticulated are species? BioEssays 38, 140–149 (2016).
Abbott, R. et al. Hybridization and speciation. J. Evol. Biol. 26, 229–246 (2013).
Shapiro, B. J., Leducq, J.-B. & Mallet, J. What is speciation? PLoS Genet. 12, e1005860 (2016).
Nolte, A. W. & Tautz, D. Understanding the onset of hybrid speciation. Trends Genet. 26, 54–58 (2010).
Baack, E. J. & Rieseberg, L. H. A genomic view of introgression and hybrid speciation. Curr. Opin. Genet. Dev. 17, 513–518 (2007).
Rieseberg, L. H., Archer, M. A. & Wayne, R. K. Transgressive segregation, adaptation and speciation. Heredity 83, 363–372 (1999).
Lamichhaney, S. et al. Rapid hybrid speciation in Darwin’s finches. Science 359, 224–228 (2018).
Rieseberg, L. H. et al. Hybridization and the colonization of novel habitats by annual sunflowers. Genetica 129, 149–165 (2007).
Rieseberg, L. H. Homoploid reticulate evolution in Helianthus (Asteraceae): evidence from ribosomal genes. Am. J. Bot. 78, 1218–1237 (1991).
Schwarzbach, A. E. & Rieseberg, L. H. Likely multiple origins of a diploid hybrid sunflower species. Mol. Ecol. 11, 1703–1715 (2002).
Pryszcz, L. P. et al. The genomic aftermath of hybridization in the opportunistic pathogen Candida metapsilosis. PLoS Genet. 11, e1005626 (2015).
Kellis, M., Patterson, N., Endrizzi, M., Birren, B. & Lander, E. S. Sequencing and comparison of yeast species to identify genes and regulatory elements. Nature 423, 241–254 (2003).
Greig, D. Reproductive isolation in Saccharomyces. Heredity 102, 39–44 (2009).
Liti, G., Barton, D. B. H. & Louis, E. J. Sequence diversity, reproductive isolation and species concepts in Saccharomyces. Genetics 174, 839–850 (2006).
Hou, J., Friedrich, A., De Montigny, J. & Schacherer, J. Chromosomal rearrangements as a major mechanism in the onset of reproductive isolation in Saccharomyces cerevisiae. Curr. Biol. 24, 1153–1159 (2014).
Morales, L. & Dujon, B. Evolutionary role of interspecies hybridization and genetic exchanges in yeasts. Microbiol. Mol. Biol. Rev. 76, 721–739 (2012).
Naumov, G. I. et al. Natural polyploidization of some cultured yeast Saccharomyces sensu stricto: auto- and allotetraploidy. Syst. Appl. Microbiol. 23, 442–449 (2000).
Zhang, H., Skelton, A., Gardner, R. C. & Goddard, M. R. Saccharomyces paradoxus and Saccharomyces cerevisiae reside on oak trees in New Zealand: evidence for migration from Europe and interspecies hybrids. FEMS Yeast Res. 10, 941–947 (2010).
Stefanini, I. et al. Social wasps are a Saccharomyces mating nest. Proc. Natl Acad. Sci. USA 113, 2247–2251 (2016).
Nakao, Y. et al. Genome sequence of the lager brewing yeast, an interspecies hybrid. DNA Res. 16, 115–129 (2009).
Walther, A., Hesselbart, A. & Wendland, J. Genome sequence of Saccharomyces carlsbergensis, the world’s first pure culture lager yeast. G3 4, 783–793 (2014).
Dunn, B. & Sherlock, G. Reconstruction of the genome origins and evolution of the hybrid lager yeast Saccharomyces pastorianus. Genome Res. 18, 1610–1623 (2008).
Bing, J., Han, P.-J., Liu, W.-Q., Wang, Q.-M. & Bai, F.-Y. Evidence for a Far East Asian origin of lager beer yeast. Curr. Biol. 24, R380–R381 (2014).
Cavalieri, D., McGovern, P. E., Hartl, D. L., Mortimer, R. & Polsinelli, M. Evidence for S. cerevisiae fermentation in ancient wine. J. Mol. Evol. 57 S226–S232 (2003).
Michel, R. H., McGovern, P. E. & Badler, V. R. Chemical evidence for ancient beer. Nature 360, 24 (1992).
McGovern, P. E. et al. Fermented beverages of pre- and proto-historic china. Proc. Natl Acad. Sci. USA 101, 17593–17598 (2004).
Borneman, A. R. et al. Whole-genome comparison reveals novel genetic elements that characterize the genome of industrial strains of Saccharomyces cerevisiae. PLoS Genet. 7, e1001287 (2011).
Gallone, B. et al. Domestication and divergence of Saccharomyces cerevisiae beer yeasts. Cell 166, 1397–1410.e16 (2016).
Gonçalves, M. et al. Distinct domestication trajectories in top-fermenting beer yeasts and wine yeasts. Curr. Biol. 26, 2750–2761 (2016).
Legras, J.-L. et al. Adaptation of S. cerevisiae to fermented food environments reveals remarkable genome plasticity and the footprints of domestication. Mol. Biol. Evol. 35, 1712–1727 (2018).
Peter, J. et al. Genome evolution across 1,011 Saccharomyces cerevisiae isolates. Nature 556, 339–344 (2018).
Yue, J.-X. et al. Contrasting evolutionary genome dynamics between domesticated and wild yeasts. Nat. Genet. 49, 913–924 (2017).
Dunn, B., Richter, C., Kvitek, D. J., Pugh, T. & Sherlock, G. Analysis of the Saccharomyces cerevisiae pan-genome reveals a pool of copy number variants distributed in diverse yeast strains from differing industrial environments. Genome Res. 22, 908–924 (2012).
Duan, S. F. et al. The origin and adaptive evolution of domesticated populations of yeast from Far East Asia. Nat. Commun. 9, 2690 (2018).
Ludlow, C. L. et al. Independent origins of yeast associated with coffee and cacao fermentation. Curr. Biol. 26, 965–971 (2016).
Barbosa, R. et al. Multiple rounds of artificial selection promote microbe secondary domestication—the case of cachaça yeasts. Genome Biol. Evol. 10, 1939–1955 (2018).
Libkind, D. et al. Microbe domestication and the identification of the wild genetic stock of lager-brewing yeast. Proc. Natl Acad. Sci. USA 108, 14539–14544 (2011).
Eizaguirre, J. I. et al. Phylogeography of the wild lager-brewing ancestor (Saccharomyces eubayanus) in Patagonia. Environ. Microbiol. 20, 3732–3743 (2018).
Gibson, B. R., Storgårds, E., Krogerus, K. & Vidgren, V. Comparative physiology and fermentation performance of Saaz and Frohberg lager yeast strains and the parental species Saccharomyces eubayanus. Yeast 30, 255–266 (2013).
González, S. S., Barrio, E. & Querol, A. Molecular characterization of new natural hybrids of Saccharomyces cerevisiae and S. kudriavzevii in brewing. Appl. Environ. Microbiol. 74, 2314–2320 (2008).
González, S. S., Barrio, E., Gafner, J. & Querol, A. Natural hybrids from Saccharomyces cerevisiae, Saccharomyces bayanus and Saccharomyces kudriavzevii in wine fermentations. FEMS Yeast Res. 6, 1221–1234 (2006).
Masneuf, I., Hansen, J., Groth, C., Piskur, J. & Dubourdieu, D. New hybrids between Saccharomyces sensu stricto yeast species found among wine and cider production strains. Appl. Environ. Microbiol. 64, 3887–3892 (1998).
Peris, D., Lopes, Ca, Belloch, C., Querol, A. & Barrio, E. Comparative genomics among Saccharomyces cerevisiae × Saccharomyces kudriavzevii natural hybrid strains isolated from wine and beer reveals different origins. BMC Genomics 13, 407 (2012).
Peris, D., Pérez-Torrado, R., Hittinger, C. T., Barrio, E. & Querol, A. On the origins and industrial applications of Saccharomyces cerevisiae × Saccharomyces kudriavzevii hybrids. Yeast 35, 51–69 (2018).
Lopes, C. A., Barrio, E. & Querol, A. Natural hybrids of S. cerevisiae × S. kudriavzevii share alleles with European wild populations of Saccharomyces kudriavzevii. FEMS Yeast Res. 10, 412–421 (2010).
Peris, D. et al. The molecular characterization of new types of Saccharomyces cerevisiae × S. kudriavzevii hybrid yeasts unveils a high genetic diversity. Yeast 29, 81–91 (2012).
Le Jeune, C. et al. Characterization of natural hybrids of Saccharomyces cerevisiae and Saccharomyces bayanus var. uvarum. FEMS Yeast Res. 7, 540–549 (2007).
Gallone, B. et al. Origins, evolution, domestication and diversity of Saccharomyces beer yeasts. Curr. Opin. Biotechnol. 49, 148–155 (2018).
Steensels, J., Gallone, B., Voordeckers, K. & Verstrepen, K. J. Domestication of industrial microbes. Curr. Biol. 29, R381–R393 (2019).
Wendland, J. Lager yeast comes of age. Eukaryot. Cell 13, 1256–1265 (2014).
G Gibson, B. & Liti, G. Saccharomyces pastorianus: genomic insights inspiring innovation for industry. Yeast 32, 17–27 (2015).
Peris, D. et al. Complex ancestries of lager-brewing hybrids were shaped by standing variation in the wild yeast Saccharomyces eubayanus. PLoS Genet. 12, e1006155 (2016).
Okuno, M. et al. Next-generation sequencing analysis of lager brewing yeast strains reveals the evolutionary history of interspecies hybridization. DNA Res. 23, dsv037 (2016).
Dai, F. et al. Tibet is one of the centers of domestication of cultivated barley. Proc. Natl Acad. Sci. USA 109, 16969–16973 (2012).
Meußdoerffer, F. & Zarnkow, M. Das Bier: Eine Geschichte von Hopfen und Malz (Beck, 2014).
Baker, E. et al. The genome sequence of Saccharomyces eubayanus and the domestication of lager-brewing yeasts. Mol. Biol. Evol. 32, 2818–2831 (2015).
Monerawela, C. & Bond, U. Recombination sites on hybrid chromosomes in Saccharomyces pastorianus share common sequence motifs and define a complex evolutionary relationship between group I and II lager yeasts. FEMS Yeast Res. 17, fox047 (2017).
Belloch, C. et al. Chimeric genomes of natural hybrids of Saccharomyces cerevisiae and Saccharomyces kudriavzevii. Appl. Environ. Microbiol. 75, 2534–2544 (2009).
Hewitt, S. K., Donaldson, I. J., Lovell, S. C. & Delneri, D. Sequencing and characterisation of rearrangements in three S. pastorianus strains reveals the presence of chimeric genes and gives evidence of breakpoint reuse. PLoS ONE 9, e92203 (2014).
Almeida, P. et al. A Gondwanan imprint on global diversity and domestication of wine and cider yeast Saccharomyces uvarum. Nat. Commun. 5, 4044 (2014).
Peŕez-Través, L., Lopes, C. A., Querol, A. & Barrio, E. On the complexity of the Saccharomyces bayanus taxon: hybridization and potential hybrid speciation. PLoS ONE 9, e93729 (2014).
Back, W. Colour Atlas and Handbook of Beverage Biology (Fachverlag Hans Carl, 2005).
Krogerus, K., Magalhães, F., Vidgren, V. & Gibson, B. Novel brewing yeast hybrids: creation and application. Appl. Microbiol. Biotechnol. 101, 65–78 (2017).
van den Broek, M. et al. Chromosomal copy number variation in Saccharomyces pastorianus is evidence for extensive genome dynamics in industrial lager brewing strains. Appl. Environ. Microbiol. 81, 6253–6267 (2015).
Monerawela, C. & Bond, U. The hybrid genomes of Saccharomyces pastorianus: a current perspective. Yeast 35, 39–50 (2017).
Fisher, K. J., Buskirk, S. W., Vignogna, R. C., Marad, D. A. & Lang, G. I. Adaptive genome duplication affects patterns of molecular evolution in Saccharomyces cerevisiae. PLoS Genet. 14, e1007396 (2018).
Selmecki, A. M. et al. Polyploidy can drive rapid adaptation in yeast. Nature 519, 349–352 (2015).
Todd, R. T., Forche, A. & Selmecki, A. in The Fungal Kingdom (eds Heitman, J. et al.) Ch. 28 (ASM Press, 2017).
Baker, E. C. P. et al. Mitochondrial DNA and temperature tolerance in lager yeasts. Sci. Adv. 5, eaav1869 (2019).
Li, X. C., Peris, D., Hittinger, C. T., Sia, E. A. & Fay, J. C. Mitochondria-encoded genes contribute to evolution of heat and cold tolerance in yeast. Sci. Adv. 5, eaav1848 (2019).
De Roos, J. & De Vuyst, L. Microbial acidification, alcoholization, and aroma production during spontaneous lambic beer production. J. Sci. Food Agric. 99, 25–38 (2019).
Steensels, J. & Verstrepen, K. J. Taming wild yeast: potential of conventional and nonconventional yeasts in industrial fermentations. Annu. Rev. Microbiol. 68, 61–80 (2014).
Dzialo, M. C., Park, R., Steensels, J., Lievens, B. & Verstrepen, K. J. Physiology, ecology and industrial applications of aroma formation in yeast. FEMS Microbiol. Rev. 41, S95–S128 (2017).
Mukai, N., Masaki, K., Fujii, T., Kawamukai, M. & Iefuji, H. PAD1 and FDC1 are essential for the decarboxylation of phenylacrylic acids in Saccharomyces cerevisiae. J. Biosci. Bioeng. 109, 564–569 (2010).
Mukai, N., Masaki, K., Fujii, T. & Iefuji, H. Single nucleotide polymorphisms of PAD1 and FDC1 show a positive relationship with ferulic acid decarboxylation ability among industrial yeasts used in alcoholic beverage production. J. Biosci. Bioeng. 118, 50–55 (2014).
Pasteur, L. Etudes sur la bière, ses maladies, causes qui les provoquent, procédé pour la rendre inaltérable, avec une théorie nouvelle de la fermentation (Gauthier-Villars, 1876).
Teich, M. Bier, Wissenschaft und Wirtschaft in Deutschland 1800–1914. Ein Beitrag zur Deutschen Industrialisierungsgeschichte (Böhlau, 2000).
Hansen, E. Recherches sur la physiologie et la morphologie des ferments alcooliques V. methodes pour obtenir des cultures pures de Saccharomyces et de mikroorganismes analogues. C. R. Trav. Lab. Carlsberg 2, 92–105 (1883).
Mertens, S. et al. A large set of newly created interspecific Saccharomyces hybrids increases aromatic diversity in lager beers. Appl. Environ. Microbiol. 81, 8202–8214 (2015).
Peris, D. et al. Hybridization and adaptive evolution of diverse Saccharomyces species for cellulosic biofuel production. Biotechnol. Biofuels 10, 78 (2017).
Scannell, D. R. et al. The awesome power of yeast evolutionary genetics: new genome sequences and strain resources for the Saccharomyces sensu stricto genus. G3 1, 11–25 (2011).
Liti, G. et al. High quality de novo sequencing and assembly of the Saccharomyces arboricolus genome. BMC Genomics 14, 69 (2013).
Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows–Wheeler transform. Bioinformatics 25, 1754–1760 (2009).
Quinlan, A. R. & Hall, I. M. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 26, 841–842 (2010).
Bankevich, A. et al. SPAdes: a new genome assembly algorithm and its applications to single-cell sequencing. J. Comput. Biol. 19, 455–477 (2012).
Pryszcz, L. P. & Gabaldón, T. Redundans: an assembly pipeline for highly heterozygous genomes. Nucleic Acids Res. 44, e113 (2016).
Kolmogorov, M., Raney, B., Paten, B. & Pham, S. Ragout—a reference-assisted assembly tool for bacterial genomes. Bioinformatics 30, i302–i309 (2014).
Wapinski, I., Pfeffer, A., Friedman, N. & Regev, A. Natural history and evolutionary principles of gene duplication in fungi. Nature 449, 54–61 (2007).
Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J. Basic local alignment search tool. J. Mol. Biol. 215, 403–410 (1990).
Camacho, C. et al. BLAST+: architecture and applications. BMC Bioinformatics 10, 421 (2009).
Katoh, K. & Standley, D. M. MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol. Biol. Evol. 30, 772–780 (2013).
Kück, P. & Meusemann, K. FASconCAT: convenient handling of data matrices. Mol. Phylogenet. Evol. 56, 1115–1118 (2010).
Capella-Gutiérrez, S., Silla-Martínez, J. M. & Gabaldón, T. trimAl: a tool for automated alignment trimming in large-scale phylogenetic analyses. Bioinformatics 25, 1972–1973 (2009).
Stamatakis, A. RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics 30, 1312–1313 (2014).
Kozlov, A. M., Aberer, A. J. & Stamatakis, A. ExaML version 3: a tool for phylogenomic analyses on supercomputers. Bioinformatics 31, 2577–2579 (2015).
Pattengale, N. D., Alipour, M., Bininda-Emonds, O. R. P., Moret, B. M. E. & Stamatakis, A. How many bootstrap replicates are necessary? J. Comput. Biol. 17, 337–354 (2010).
R Core Team R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, 2011).
Yu, G., Smith, D. K., Zhu, H., Guan, Y. & Lam, T. T. Y. ggtree: an R package for visualization and annotation of phylogenetic trees with their covariates and other associated data. Methods Ecol. Evol. 8, 28–36 (2017).
Suchard, M. A. et al. Bayesian phylogenetic and phylodynamic data integration using BEAST 1.10. Virus Evol. 4, vey016 (2018).
Raj, A., Stephens, M. & Pritchard, J. K. fastSTRUCTURE: variational inference of population structure in large SNP data sets. Genetics 197, 573–589 (2014).
Ayres, D. L. et al. BEAGLE: an application programming interface and high-performance computing library for statistical phylogenetics. Syst. Biol. 61, 170–173 (2012).
Drummond, A. J., Ho, S. Y., Phillips, M. J. & Rambaut, A. Relaxed phylogenetics and dating with confidence. PLoS Biol. 4, e88 (2006).
Drummond, A. J. & Suchard, M. A. Bayesian random local clocks, or one rate to rule them all. BMC Biol. 8, 114 (2010).
Worobey, M., Han, G.-Z. & Rambaut, A. A synchronized global sweep of the internal genes of modern avian influenza virus. Nature 508, 254–257 (2014).
Rambaut, A., Drummond, A. J., Xie, D., Baele, G. & Suchard, M. A. Posterior summarization in Bayesian phylogenetics using tracer 1.7. Syst. Biol. 67, 901–904 (2018).
Baele, G., Lemey, P. & Suchard, M. A. Genealogical working distributions for Bayesian model testing with phylogenetic uncertainty. Syst. Biol. 65, 250–264 (2016).
Angiuoli, S. V. & Salzberg, S. L. Mugsy: fast multiple alignment of closely related whole genomes. Bioinformatics 27, 334–342 (2011).
Venables, W. N. & Ripley, B. D. Modern applied statistics with S. Technometrics 45, 111–111 (2003).
Warnes, G. R. et al. gplots: Various R programming tools for plotting data. R version 3.0.0 http://CRAN.R-project.org/package=gplots (2016).
Murtagh, F. & Legendre, P. Ward’s hierarchical agglomerative clustering method: which algorithms implement Ward’s criterion? J. Classif. 31, 274–295 (2014).
Calahan, D., Dunham, M., Desevo, C. & Koshland, D. E. Genetic analysis of desiccation tolerance in Saccharomyces cerevisiae. Genetics 189, 507–519 (2011).
Muir, A., Harrison, E. & Wheals, A. A multiplex set of species-specific primers for rapid identification of members of the genus Saccharomyces. FEMS Yeast Res. 11, 552–563 (2011).
Pengelly, R. J. & Wheals, A. E. Rapid identification of Saccharomyces eubayanus and its hybrids. FEMS Yeast Res. 13, 156–161 (2013).
Acknowledgements
We thank all K.J.V. and S.Maere laboratory members for their help and suggestions. We thank AB InBev, White Labs Inc. and R. S. Wagner (Central Washington University) for isolating and sharing various hybrid yeasts from the beer environment. We thank M. Zarnkow (Research Center Weihenstephan for Brewing and Food Quality) for sharing historical facts on lager brewing. B.G. acknowledges funding from the VIB International PhD Program in Life Sciences. J.S. acknowledges funding from Research Foundation-Flanders (FWO) (grant no. 12W3918N). Research in the laboratory of K.J.V. is supported by KU Leuven Program Financing, European Research Council Consolidator Grant (no. CoG682009), Human Frontier Science program grant (no. RGP0050/2013), VIB, FWO (grant no. G.0C59.14N) and Vlaams Agentschap Innoveren en Ondernemen (grant no. HBC.2018.0097). Research in the laboratory of S.Maere is supported by VIB, Ghent University and FWO (grant no. G018915N).
Author information
Authors and Affiliations
Contributions
B.G., J.S., S.Maere and K.J.V. conceived the project. B.G. and J.S. performed the data collection. B.G., J.S. and G.B. carried out the formal analysis. B.G., J.S., S.Mertens, J.L.G., R.W., F.A.T., F.B., V.S., B.H.-M. and G.B. carried out the investigations. B.G. and J.S. performed data visualization. M.H., F.M., P.M., B.S., L.D., T.P., C.W. and K.J.V. provided resources. B.G., J.S., M.C.D., S.Maere and K.J.V. wrote the paper. S.Maere and K.J.V. supervised the project.
Corresponding authors
Ethics declarations
Competing interests
AB InBev partially funded the experiments performed in this study. AB InBev had no role in conceptualization, design, data collection, analysis, decision to publish or preparation of the manuscript. B.S., P.M. and L.D. are employed by AB InBev.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Supplementary Information
Supplementary Figs. 1–5, Tables 1–3, Note 1 and Dataset 1.
Supplementary Table 1
List of strains included in this study.
Supplementary Table 2
Chimeric breakpoints identified in Saccharomyces interspecific hybrids.
Supplementary Table 3
Phenotypic variation of Saccharomyces interspecific hybrids and Saccharomyces pure species.
Rights and permissions
About this article
Cite this article
Gallone, B., Steensels, J., Mertens, S. et al. Interspecific hybridization facilitates niche adaptation in beer yeast. Nat Ecol Evol 3, 1562–1575 (2019). https://doi.org/10.1038/s41559-019-0997-9
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/s41559-019-0997-9
This article is cited by
-
Ancient and recent origins of shared polymorphisms in yeast
Nature Ecology & Evolution (2024)
-
phyBWT2: phylogeny reconstruction via eBWT positional clustering
Algorithms for Molecular Biology (2023)
-
Increased volatile thiol release during beer fermentation using constructed interspecies yeast hybrids
European Food Research and Technology (2023)
-
Macroevolutionary diversity of traits and genomes in the model yeast genus Saccharomyces
Nature Communications (2023)
-
Unlocking the functional potential of polyploid yeasts
Nature Communications (2022)
