Paternal outcrossing success differs among faba bean genotypes and impacts breeding of synthetic cultivars

Abstract

Faba bean genotypes showed significant and marked genetic differences in their success as pollen donors to cross-fertilized seeds. The findings may improve exploitation of heterosis in synthetic cultivars. In partially allogamous crops such as faba bean (Vicia faba L.), increasing the share of heterosis in a synthetic cultivar can improve yield and yield stability. The share of heterosis in such synthetic cultivars is increased by higher degrees of cross-fertilization. This trait is defined as percentage of cross-fertilized seeds among all seeds and is a crucial parameter in breeders' yield predictions. Current approaches use degree of cross-fertilization to predict inbreeding and share of heterosis, they even consider genotype-specific degrees; yet, all genotypes are assumed to contribute equally to the cross-fertilized seeds. Here, we expect faba bean genotypes to differ in their success rates as pollen donors, i.e. in paternal outcrossing success. To quantify the variation of both, the degree of cross-fertilization and the paternal outcrossing success, we assessed these parameters in inbred lines and F1 hybrids, grown in four polycrosses composed of eight genotypes each. We identified the paternal genotype of 500 to 800 seeds per genotype and polycross using SNP markers. In both traits, we found marked and significant variation among inbred lines and among F1 hybrids, as well as between inbred lines and F1. Based on our findings, we discuss how differential paternal outcrossing success influences the amount of inbreeding in synthetic cultivars. Our findings offer the potential for a better management and exploitation of heterotic yield increase in faba bean.

Fig. 1

Schematic representation of the partial allogamy in faba bean. From a flowering plant (a), a section is shown in further detail (b): the ovules inside the flowers, ready for fertilization (top), flowers already developed into pods (bottom). The completely white seeds in these pods illustrate self-fertilized seeds. Seeds that are half white and half-filled illustrate cross-fertilized seeds. In this example, the genotype has a degree of cross-fertilization of 4/9, i.e. about 44%

Fig. 2

Schematic representation of paternal outcrossing success in faba bean. Genotype A (compare Fig. 1) is shown with its seeds (black frame), focussing on the composition of the cross-fertilized seeds. The striped halves of the seeds indicate cross-fertilization by three different pollen donors (arrows from genotypes A, B and C). Four seeds of genotype A originate from inter-genotype cross-fertilization: one from genotype B and three from genotype C. In this example, genotype C has a higher paternal outcrossing success than genotype B. In addition, two seeds originate from intra-genotype cross-fertilization (with the same genotype A as female and male parent). The three white seeds of genotype A originate from self-fertilization

Fig. 3

Degree of cross-fertilization of the faba bean genotypes in different sets of genotypes: a set 0 in one year (2014) and at one location (GAR), b set A in two years (2015, 2016) and at one location (DRA), and c set B in one year (2016) and at one location (DEP). For each set, different letters show significant differences among least square means (p = 0.05, multiple comparisons of means with user-defined contrasts and Bonferroni correction). Vertical lines show 95-percent confidence intervals. Least square mean values for each environment (combinations of year × location) are shown in Supplementary Table 6

Fig. 4

Paternal outcrossing success of the faba bean genotypes in different sets of genotypes: a set 0 in one year (2014) and at one location (GAR), b set A in two years (2015, 2016) and at one location (DRA), and c set B in one year (2016) and at one location (DEP), sorted by mean degree of cross-fertilization of the same set. For each set, different letters show significant differences among least square means (p = 0.05, multiple comparisons of means with user-defined contrasts and Bonferroni correction). Vertical lines show 95-percent confidence intervals. The horizontal line indicates a paternal outcrossing success of 12.5%. Least square mean values for each environment (combinations of year × location) are shown in Supplementary Table 7

Fig. 5

Correlation of degree of cross-fertilization C and paternal outcrossing success P in different sets of genotypes: a set 0 in one year (2014) and at one location (GAR), b set A in two years (2015, 2016) and at one location (DRA), and c set B in one year (2016) and at one location (DEP). Error bars show 95-percent confidence intervals. Correlation coefficients are calculated based on Pearson's product-moment correlation coefficient for (i) all genotypes, (ii) only inbred lines