Figure 3.

Cytoplasmic–nuclear incompatibility is widespread within C. nouraguensis. (A) A map depicting the two major sites where the strains used in this study were collected in French Guiana. GPS coordinates for NIC54 were obtained from C. Braendle (personal communication), while the other six were obtained from Kiontke et al. (2011) and Félix et al. (2013). Strains in the southern collection site were collected from distinct rotten fruits or flowers within 2 km of each other and are represented as a single point. (B) To determine whether a particular cytoplasmic–nuclear combination is incompatible, we tested for statistical differences in viability between the F1 female backcross that combines heterotypic cytoplasmic and nuclear genotypes (top cross) and the backcross that combines homotypic cytoplasmic and nuclear genotypes (bottom cross, see Materials and Methods). We also calculated the relative viability of the first cross to the second. (C) A scatter plot depicting all the cytoplasmic–nuclear compatibility tests performed. Each point corresponds to a single replicate of a certain cytoplasmic–nuclear combination. Points above the horizontal dashed gray line indicate statistically significant differences in viability between the two types of crosses mentioned in (B) (P < 0.0006 after Bonferroni correction, Fisher’s exact test). Points above the horizontal dashed gray line that have a relative viability <1 are considered statistically significant cytoplasmic–nuclear incompatibilities. The color of a point corresponds to the cytoplasmic genotype being tested. All cytoplasmic genotypes tested show an incompatibility with one or more heterotypic nuclear genotypes. See Figure S4 for separate graphs of all combinations. Above the scatterplot are boxplots depicting the relative viabilities of statistically significant cytoplasmic–nuclear incompatibilities. The color corresponds to the cytoplasmic genotype tested. Incompatibilities involving the NIC59 cytoplasmic genotype have reduced viability compared to those involving the JU1837 and JU1854 cytoplasmic genotypes (P < 0.001, Kruskal–Wallis test followed by Dunn’s test). (D) A heatmap depicting the median relative viability for each cytoplasmic–nuclear combination. Each cytoplasmic–nuclear combination shows the proportion of replicates that exhibit significant incompatibilities (for example, three out of three replicates exhibit significant incompatibilities for the NIC59 cytoplasm–JU1854 nuclear combination, while only one out of three replicates exhibit significant incompatibilities for the JU1837 cytoplasm–JU1854 nuclear combination). Each cytoplasmic genotype is consistently incompatible with at least one heterotypic nuclear genotype. The NIC59 cytoplasm has a more distinct response to hybridization than the others tested.