Outcrossing Rates in Annatto Determined by Sequence-Related Amplified Polymorphism Agronomy Journal 102 :1340–1345 , 2010 Ruby Valdez-Ojeda, Carlos F. Quiros, Margarita de Loudes Aguilar-Espinosa, and Renata Rivera-Madrid

Annatto (Bixa orellana L.) Annatto is grown for its unique capability to produce bixin, a natural pigment found predominantly in the outer coatings of its seeds and widely used in the food industry (Bittencourt et al., 2005). However, annatto dye and grain yields vary from sample to sample and from plant to plant (Michelangeli et al., 2002; Valdez-Ojeda et al., 2008). The annatto improvement is actually incipient … Annatto presents different colors in flowers and fruits showing high morphological diversity

In genetic Improvement … (Sleper and Poehlman, 2006) Critical decisions in breeding programs include: The choice of mating. The intensity of selection applied and proper management of inbreeding. All of these require the consideration of the mating system. To allow the breeder to select the most effective breeding scheme.

Sistema de reproducción SRAP markers have been used here to obtain quantitative estimates of mating system parameters and to make appropriate estimates of outcrossing rates, by using MLDT, multilocus estimation of outcrossing with dominant markers procedure. Before this study we generated information concerning the mating system in B. orellana through artificial pollination. We found 57% outcrossing and 31.4% selfing (Rivera-Madrid et al., 2006). To support the existence of both reproduction systems in B. orellana, and because SRAP markers (Li and Quiros, 2001) have been successfully used in Cyperus diformis (Merotto et al., 2009).

MATERIALS AND METHODS Bixa orellana populations bloom from September to November in the yields of Yucatan, Mexico. Maternal trees showing different traits in flowers and fruits. White flowers are indicated by numbers 16, 19, 58, and 87. Pink flowers are indicated by numbers 54, 55, 56, 67, 68, 77, and 79. The plant material was collected from the agronomic region of Yucatan located at latitude 20º13´ to 20º29´ and longitude 88º55´ to 89º54´ named Yaxcaba. The selected trees were sufficiently distant from each other to allow the assumption of no relatedness among the individuals. Eleven trees subjected to open-pollination (OP) were randomly selected for progeny testing in 2005 in a total population of 100 individuals.

From each tree, seeds were randomly collected, and grown under greenhouse conditions Sequence-Related Ampliied Polymorphism Procedure Total genomic DNA was extracted using a DNeasy Tissue Kit (Qiagen). The samples were run in the sequencer IR2 4200, LI-Cor, Biosciences, (Lincoln, NE), in triplicate to ensure reproducibility.

The bands were scored considering two possible alleles: band presence or absence. Data Analysis Nineteen primer combinations were used to detect 50 polymorphic bands, which were used for this study. Sequence-related amplified polymorphism (SRAP) fragments generated with primer combinations SA–12 and EM–8. The families 1, 2, and 3 are indicated. The arrows indicate the markers detected. Size standard (in base pairs) is shown on the left.

The mating system was analyzed using Multilocus estimation of outcrossing with dominant markers procedure (MLDT) (Ritland and Jain, 1981; Ritland,1990; and Ritland 2002). This MLDT specifies that both selfing and outcrossing occur in a population (Shaw and Allard, 1982). From progeny array data, the MLDT simultaneously estimated: tm by the Newton-Raphson method (which includes biparental inbreeding) and, Average ts (which excludes most biparental inbreeding). The difference between these estimates was considered to be the rate of biparental inbreeding. One thousand bootstraps of among the families sampled were used to estimate the standard error for all outcrossing rates (including tm – ts) and allele frequencies. The average single-locus inbreeding coefficient (also known as Wright’s fixation index) of maternal parents (F) via the Newton-Raphson method and the average gene frequencies via the expectation maximization method were also estimated (Ritland and Jain,1981; Ritland, 2002). Data Analysis

RESULTS This indicates that B. orellana is predominantly an outcrossing species. This data show that some selfing also occurs The difference suggests the existence of mating among relatives (biparental inbreeding). This indicate high homozygosity.

In our study, we used four maternal plants with white flowers and with pink flowers; therefore it was possible to test whether different outcrossing rates applied to these morphological types. To test this hypothesis we analyzed white and pink variants separately. We found high outcrossing rates tm, for the white and pink flower variants with no siginificant differences between them. However, of particular interest was the Wrights’ fixation index. This indicated for the white flower variants high values demostrating an excess of homozygotes, and for pink variants an excess of heterozygotes

In summary … We found that the mating system of annatto is predominantly outcrossing. We also found differences in F between different flower color variants. High F values demonstrated an excess of homozygotes in the white lower types. This is probably explained by few pollinators visiting white lower phenotypes, as a result of a reduced preference for this flower color. The annatto variants analyzed in this study present agronomical characteristics desirable for the genetic improvement of this crop, favoring the attainment of high contents of bixin. In the case of the white variants, these are characterized by low bixin contents and indehiscent fruit, a characteristic that may enhance crop yields by protecting the seeds and pigment quality; while the pink variant presents high bixin contents but dehiscent fruits that expose the seeds to sunlight, leading to bixin photooxidation and decreased bixin content (Valdez-Ojeda et al., 2008)

This investigation also strongly indicates a need for in-depth studies concerning the nature of pollination (floral structure and its relation to the mating system; and the contribution of pollinating agents and seed dispersal) since evidence suggests that the white variant can be both self and cross pollinated and that pollinators may play an important role in these processes. Perspectives … Knowledge of all of the above aspects, in conjunction with the results of this study, need to be taken into account in breeding program to improve the annatto crop.