Spontaneous whole-genome duplication restores fertility in interspecific hybrids

Abstract

Interspecies hybrids often show some advantages over parents but also frequently suffer from reduced fertility, which can sometimes be overcome through sexual reproduction that sorts out genetic incompatibilities. Sex is however inefficient due to the low viability or fertility of hybrid offspring and thus limits their evolutionary potential. Mitotic cell division could be an alternative to fertility recovery in species such as fungi that can also propagate asexually. Here, to test this, we evolve in parallel and under relaxed selection more than 600 diploid yeast inter-specific hybrids that span from 100,000 to 15 M years of divergence. We find that hybrids can recover fertility spontaneously and rapidly through whole-genome duplication. These events occur in both hybrids between young and well-established species. Our results show that the instability of ploidy in hybrid is an accessible path to spontaneous fertility recovery.

Fig. 1

Neutral evolution of yeast hybrids and its impact on viability and sporulation ability. a Evolution of fertility during mitotic proliferation of hybrids between species: Potential scenarios: (1) Gradual recovery over time, (2) rapid sudden recovery, (3) no significant recovery, (4) decline over time, (5) rapid and sudden decline. b Crosses were performed among S. paradoxus lineages (VL_div, L_div, and M_div) and with S. cerevisiae (H_div). Each type of cross involves two biological replicates, i.e., involving independent strains, and each individual cross was performed independently to represent independent hybridization events. c The 672 hybrids were evolved in conditions of weak selection to examine the neutral spontaneous changes of fertility. Mitotic propagation was performed through repeated bottlenecks of single cells. Fertility was measured by estimating spore viability after meiosis at T_ini, T_mid, and T_end. d Survival rates vary among lines. Timing of each glycerol stock indicated by vertical grey lines. Logrank test pairwise comparisons FDR corrected P-values are shown. e Fraction of lines that lost their sporulation capacity at T_ini, T_mid, and T_end. The number of strains tested per cross type is indicated over the corresponding bars

Fig. 2

Contrast in fertility changes between mitotic and meiotic propagation. a Example of spore survival from lines showing a reduction of fertility (top), no change of fertility (middle) and recovery of fertility (bottom). Spores considered as viable are circled in black. Spores from sporulation at the initial time point are shown on the left and from the final time point on the right. b Fertility recovery scores (FRS), which measure the change in spore viability, show no clear directionality in mitotically propagated lines (VL_div, L_div, M_div, and H_div). P-values given for exact binomial tests. FRS for intra-tetrad crosses (ITC) indicate large and systematic increase in fertility. Bins containing strains with high fertility by the midpoint of the experiment highlighted with red arrows. c Autodiploidization was performed on a random set of 16 haploid viable spores dissected from the L_div and M_div crosses at T_ini, T_mid and T_end. Fertility of initial hybrids and evolved lines (parental) increases to above 80% after spore autodiploidization. Numbers in parentheses represent sample sizes. For all boxplots the bold center line corresponds to the median value, the box boundaries correspond to the 25th and the 75th percentile, the whiskers correspond to 1.5 times the inter-quartile range and the dots to individual data points

Fig. 3

Ploidy varies among hybrids and evolves through time. a Ploidy of a subset of 12 hybrid lines after hybridization (T_ini) and after ~770 (T_end) of mitotic generations. The c gray panel corresponds to controls and black circles indicate the highest cell count whose variability suggest that aneuploidy is prevalent in hybrid lines. b Ploidy at the three tested timepoints (T_ini, T_mid, and T_end). Connected dots represent independent lines (24). The c gray panel corresponds to controls. c Frequency of 171 markers corresponding to the SpC parent alleles across the genome of a subset (six diploids, six triploids, and all the five tetraploids) of L_div lines show around 50% of SpC alleles in the diploid and tetraploid strains and 66% in triploid strains. d Allele frequencies estimated by whole-genome sequencing across the 16 chromosomes. Because diploid hybrids have one copy of each genome, allele frequencies are centered around 50% at T_ini. This frequency is preserved in most tetraploids, showing the all chromosomes were doubled at T_end. Loss of heterozygosity is detectable as alternative peaks for instance at 100% (e.g., L1_87) in tetraploids when it involves the loss of one of the two parental alleles. Aneuploidies are also detectable as alternative peaks at 80% (e.g., L1_51) or 30% (e.g. L1_87). M1_40 shows one peak at about 30% suggesting that the sequenced clone of that line is a triploid. The allele frequencies correspond to the hybrid parent 2 alleles (SpC for L lines, SpA for M lines, and S. cerevisiae for the H line)

Fig. 4

Tetraploidization leads to sudden fertility recovery over time. Fertility trajectories at the three timepoints (T_ini, T_mid, and T_end). Each connected set of dots represents an independent line. The colors correspond to ploidy. Dotted and full lines represent the two independent crosses within each genetic divergence class. Bolder lines correspond to the lines with significantly different proportions of viable spores between T_ini and T_end. In the case of VL1 and VL2, the two pairs of strains show different level of fertility, which is frequent in yeast intra-species crosses