No evidence for extrinsic post-zygotic isolation in a wild Saccharomyces yeast system

Abstract

Although microorganisms account for the largest fraction of Earth's biodiversity, we know little about how their reproductive barriers evolve. Sexual microorganisms such as Saccharomyces yeasts rapidly develop strong intrinsic post-zygotic isolation, but the role of extrinsic isolation in the early speciation process remains to be investigated. We measured the growth of F₁ hybrids between two incipient species of Saccharomyces paradoxus to assess the presence of extrinsic post-zygotic isolation across 32 environments. More than 80% of hybrids showed either partial dominance of the best parent or over-dominance for growth, revealing no fitness defects in F₁ hybrids. Extrinsic reproductive isolation therefore likely plays little role in limiting gene flow between incipient yeast species and is not a requirement for speciation.

1. Background

Restriction to gene flow between populations and incipient species can be mediated through mechanisms that act before or after mating. Studies in microorganisms have mainly focused on post-zygotic intrinsic isolation (hybrid unviability or sterility) [1,2]. In fungi, the evolution of reproductive isolation is influenced by life-history traits that would favour pre-zygotic isolation in the highly dispersing basidiomycetes but not in the substrate resident ascomycetes, which are mostly isolated by post-zygotic mechanisms [3]. In the ascomycetes Saccharomyces genus, adaptation to climatic conditions could be a driver of divergence and speciation [4]. This ecologically based divergence could contribute to the interruption of gene flow between populations and eventual incipient species through either selection against migrants that create a geographical barrier or extrinsic post-zygotic isolation [5], which arises when the hybrids between two species show inappropriate phenotypes in parental environments, leading to fitness reduction and their elimination [6].

Although ecological barriers to gene flow have been understudied in fungi [7], hybrid phenotypes have been studied in Saccharomyces yeast. Analyses show extensive hybrid vigour in both inter- and intra-specific crosses [8,9]. However, extending these results to natural systems is difficult since the crosses tested involved either strains that would otherwise rarely interact in nature, or uncommon conditions in the yeasts environment. Natural systems such as the North-American Saccharomyces paradoxus incipient species can be used to examine this mechanism. The three endemic incipient species of S. paradoxus in North America (SpB, SpC and SpC*, an introgressed incipient species resulting from an ancient hybridization event between SpB and SpC) thrive on the bark and soil associated with deciduous trees and are mainly composed of highly homozygous isolates (figure 1a,b) [2,10,11].

Figure 1.

Incipient speciation in Saccharomyces paradoxus. (a) Biogeographical scenario inferred from genomic and phenotypic data (figure adapted from [11]) and currently known discriminating phenotypes. (b) SpC (blue area) is found along the St Lawrence valley in the northeast of the S. paradoxus distribution whereas SpB (red area) can be found in the south of Québec Province and in the northeastern USA. Each of the parental species location is represented on the map by red squares (SpB strains) and blue circles (SpC strains). The black circle represents the area where the hybrid species SpC* can be isolated.

The two main lineages have different but partially overlapping distributions, which could have been partially shaped by selection against migrants. Despite this and the presence of partial intrinsic reproductive isolation through reduced hybrid fertility (figure 1a,b), no F₁ hybrids have been recovered from more than 3000 samples representing more than 392 strains ([11,12] and G. Charron 2016, unpublished data). This implies that either hybrid formation is prevented in nature through pre-zygotic isolation or selection against hybrids (extrinsic reproductive isolation). We tested this last mechanism by measuring and comparing the growth of 21 SpB/SpC F1 hybrids, their wild diploid parents and 20 intra-lineage crosses in 32 different growth conditions.

2. Material and methods

Strains were isolated across their range (electronic supplementary material, table S1) [12]. Hybrids were constructed by mating heterothallic haploid strains (electronic supplementary material, table S1). Solid media were selected based on previous experiments [10,11] (table 1). Laboratory conditions were assayed to test if the isolation method was biased against hybrids. Growth was assessed using colony sizes. Sixty-two strains (10 SpB and 11 SpC (wild diploid isolates), 21 SpB/SpC, 7 SpC/SpC and 13 SpB/SpB hybrids) in seven replicates were randomly arrayed on OmniTray plates using a BM3-BC robot (S&P Robotics, Inc., North York, Canada). The two outer borders were filled with S. cerevisiae strain BY4743 to prevent border effects. Photos were taken every 24 h for 4 days after which they were replicated on fresh media and the process repeated. Images were analysed using ImageJ 1.44o (National Institutes of Health) [13]. Colony pixel intensities after two days were log₂-transformed and used for downstream analyses.

Table 1.

List of media tested in this study. YE, yeast extract; YNB, yeast nitrogen base; Pep, peptones; AS, ammonium sulfate; Glc, d-glucose.

medium	condition type	composition (+2% agar)
ethanol	carbon	1% YE, 2% Pep, 3% ethanol
ethanol 37°C		1% YE, 2% Pep, 3% ethanol, incubation 37°C
fructose		0.174% YNB, 2% fructose, 0.5% AS
galactose		0.174% YNB, 2% galactose, 0.5% AS
glucose		0.174% YNB, 2% Glc, 0.5% AS
maltose		0.174% YNB, 2% maltose, 0.5% AS
mannose		0.174% YNB, 2% mannose, 0.5% AS
sucrose		0.174% YNB, 2% sucrose, 0.5% AS
α-methyl-d-glucopyranoside		0.174% YNB, 2% α-methyl-d-glucopyranoside, 0.5% AS
SOE		0.1% YE, 0.15% Pep, 0.5% Glc, 0.5% fructose, 1% sucrose
25°C	climatic	1% YE, 2% Pep, 2% Glc, incubation 25°C
30°C		1% YE, 2% Pep, 2% Glc
35°C		1% YE, 2% Pep, 2% Glc, incubation 35°C
37°C		1% YE, 2% Pep, 2% Glc, incubation 37°C
freeze–thaw cycle		see methods
Sniegowski I (solid)	laboratory	0.3% YE, 0.3% malt extract, 0.5% Pep, 1% sucrose, 5% ethanol, 1 mM HCl, 1 mg chloramphenicol
Sniegowski II		0.174% YNB, 2% Glc, 0.5% AS, 0.134% complete amino acid drop-out, 4 mM HCl
allantoin	nitrogen	0.174% YNB, 2% Glc, 0.5% allantoin
asparagine		0.174% YNB, 2% Glc, 0.5% asparagine
glutamine		0.174% YNB, 2% Glc, 0.5% glutamine
glycine		0.174% YNB, 2% Glc, 0.5% glycine
histidine		0.174% YNB, 2% Glc, 0.5% histidine
isoleucine		0.174% YNB, 2% Glc, 0.5% isoleucine
lysine		0.174% YNB, 2% Glc, 0.5% lysine
proline		0.174% YNB, 2% Glc, 0.5% proline
tyrosine		0.174% YNB, 2% Glc, 0.5% tyrosine
H2O2	stress	1% YE, 2% Pep, 2% Glc, 1 mM hydrogen peroxide
NaCl		1% YE, 2% Pep, 2% Glc, 1 M NaCl
pH 3		1% YE, 2% Pep, 2% Glc, pH adjusted with HCl
pH 8		1% YE, 2% Pep, 2% Glc, pH adjusted with NaOH
rapamycin		1% YE, 2% Pep, 2% Glc, 200 nM rapamycin
sorbitol		1% YE, 2% Pep, 2% Glc, 1 M sorbitol

Resistance to freezing was measured by growing cells overnight at 30°C in a 96 deep-well plate, diluted to an OD_600nm of 0.1 and grown at 30°C until they reached an OD_600nm of 0.3–0.4. Samples of 200 µl were transferred to PCR plates, one incubated in an ethanol/dry ice bath for 2 h, the other tested for cell viability as a control. Cells were centrifuged in a 5810R Centrifuge (Eppendorf) at 3000 r.p.m. for 10 min and resuspended in 200 µl of sterile water. Volumes of 25 μl were transferred into 225 µl of a 30 nM aqueous solution of SYTOX^® Green (ThermoFisher). The plate was incubated in the dark for 20 min with a resuspension step after 10 min. Fluorescence was measured on a Guava EasyCyte 14HT (EMD Millipore) on 5000 cells in biological replicates (Pearson's r = 0.992) (excitation: 488 nm, detection: 525/530 nm).

Using the median colony size values of the wild diploid parental strains for each hybrid, we calculated mid-parent values as m ((P1 + P2)/2) [8], additive genetic deviation as a (|P1 or P2−m|) and dominant genetic deviation as d (hybrid−m). The degree of dominance (d/a) was used to assign mode of inheritance: values between −1 and 1 indicate partial dominance of either the smaller or bigger colony parent; values lower than −1 indicate under-dominance and higher than 1 indicate over-dominance. Values of −1 and 1 indicate full dominance and a value of 0 indicates co-dominance. Analyses were performed with R v. 3.0 [14] (electronic supplementary material).

3. Results

SpB/SpC hybrids do not show growth defects but rather show over-dominance (47.02%) and partial dominance (35.71%) across conditions (figure 2a,b). Examples are shown in figure 2c–e. The proportions of over-dominant and partially dominant phenotypes (degree of dominance higher than 0) in hybrids is significantly greater than what is observed in intragroup crosses in one case (BC versus BB p = 2.2 × 10⁻¹⁶, Fisher's exact test, one-tailed) but not the other (BC versus CC p = 0.9997). The proportion of under-dominant hybrids is not significantly different between the laboratory media and all other conditions (p = 0.0535, Fisher's exact test, two-tailed). More than 95% (20/21) of parental strains showed median degree of dominance values above 0 in SpB/SpC crosses (electronic supplementary material, figure S1).

Figure 2.

Hybrid strains show dominance (d/a) values indicating heterosis and intermediate phenotypes. (a) Frequency distribution of d/a values for each condition type. Median values are indicated by horizontal lines and colours indicate the type of crosses (SpB/SpC in purple, SpB/SpB in red and SpC/SpC in blue). (b) Counts of d/a values by mode of inheritance. Representative conditions showing partial dominance (c), over-dominance (d) and a mix of over- and partial dominance (e) closer to full dominance.

4. Discussion

We examined whether F₁ hybrid growth defects could contribute to the reproductive isolation of two incipient yeast species. Most hybrids show no fitness defects over the array of conditions tested but rather show over-dominance or intermediate phenotypes. Therefore, it is unlikely that post-zygotic extrinsic reproductive isolation prevents gene flow between the incipient species SpB and SpC. We also note that heterosis appears to be prevalent in the SpC intragroup crosses. Because isolation data suggest that this group displays smaller effective population sizes and population density [10,11], drift could lead to the accumulation of deleterious recessive mutations locally, which would then be masked when outcrossing with other SpC strains. The segregation of deleterious mutations in SpC would also lead to dominance of the SpB phenotype in crosses with SpC. The dominance of the SpB phenotype over SpC in the SpB/SpC hybrids could therefore partly be explained by the masking of SpC deleterious alleles.

Our results are overall in line with the data reported by both Shapira et al. [8] and Bernardes et al. [9] which showed growth advantage in hybrids for both intra- (in S. cerevisiae/S. cerevisiae, S. paradoxus/S. paradoxus) and interspecific (S. cerevisiae/S. paradoxus) hybrid crosses. In natural systems where closely related species are in sympatry, this growth advantage would severely limit the role of extrinsic post-zygotic isolation or would even promote hybridization.

We observed that the hybrids have a non-significant tendency of being under-dominant in our laboratory media used for the yeast isolation. However, because the success rate in strain isolation is less than 10%, there is limited opportunity for inter-strain competition during experiments, which would limit these effects in terms of disfavouring hybrids. One limitation of our study is that we still lack a clear understanding of yeast ecology and what their fitness determinants are in nature [15], which prevents recreating the exact conditions they face in the wild. However, progress was recently made in identifying their preferred habitats, climate and interactions with other microbial species [16–18]. One environmental parameter that has been shown to play a key role in yeast ecology is high temperature [4,10,19] and we observed the highest median d/a value for conditions implicating this parameter. Finally, the observation of frequent heterosis brings the possibility that newly formed SpB/SpC hybrids have the potential to occupy new environments unavailable to their parents and would thus have not been sampled and thus be underestimated [20]. This was seen in Saccharomyces cerevisiae where hybrids between oak tree and vineyard lineages have colonized and adapted to the cherry tree environment [21].

Studies on Fungi suggest that barriers to gene flow in ascomycetes are dominated by post-zygotic mechanisms [7]. Our investigation of ecologically based post-zygotic reproductive isolation did not show any defects in hybrids. As there is only partial post-zygotic intrinsic isolation, other barriers to gene flow should exist. It has been proposed that adaptation to substrate could lead directly to reproductive isolation in ascomycetes [3]. As the two hosts where we most commonly find the S. paradoxus lineages are oak and maple trees, a specialization of each lineage for a different host tree would then restrict gene flow. There are also some evidences that reinforcement could have been evolving in other S. paradoxus yeasts [22]. The study of wild incipient species systems such as the one used here will allow a better understanding of the contributions of environment to reproductive barriers and the evolution of reproductive isolation between incipient species.

Ethics

We were not required to complete an ethical assessment prior to conducting the research.

Data accessibility

Data and scripts needed to reproduce results are available in Dryad: http://dx.doi.org/10.5061/dryad.tv5mb [23]. Supplementary figures and tables are provided as the electronic supplementary material.

Authors' contributions

G.C. and C.R.L. designed the experiments; G.C. performed the experiments and analysed the results, and drafted the manuscript with help from C.R.L. All authors gave final approval for publication and agree to be held accountable for the content herein.

Competing interests

We have no competing interests.

Funding

This project was funded by an NSERC Discovery grant to C.R.L. and an NSERC AG Bell Scholarship to G.C. C.R.L. holds the Canada Research Chair in Evolutionary Cell and Systems Biology.