Mating system analysis of Açaí-do-Amazonas (Euterpe precatoria Mart.) using molecular markers

Euterpe precatoria (Açaí-do-Amazonas) produces fruits of which the fresh pulp is consumed. It is almost exclusively collected by extractivist farmers, because no selected genotypes are available for the establishment of plantations. For the domestication and breeding of the species, mating system studies are needed for strategy formulation. This study evaluated the mating system of a natural population of E. precatoria. Thirteen progenies were genotyped with 13 microsatellite loci by capillary electrophoresis in an automated DNA sequencer. Estimates of single-locus and multilocus outcrossing rates were 1.0, and paternity correlation was low (rp̂(m) = 0.293). Euterpe precatoria families consist mainly of half-sibs and the reproductive strategy of the species is allogamy.


INTRODUCTION
The palm Euterpe precatoria Mart.(acaí-do-Amazonas) belongs to the family Arecaceae, occurring in the western and central Brazilian Amazon and to within borders of the Amazon of Peru, Brazil, Colombia (Kahn 1991), and Bolivia (Bussmann and Zambrana 2012).The species is exploited by an extractive production chain, maintained by the local population since several decades and currently intensified due to the interest of different industries.Fruit for consumption in the form of beverage is the main product of E. precatoria (Noda 2012).
When correlating the total production of açaí fruits exploited in the different states of Brazil (IBGE 2018) with the information of geographic distribution and habitat (Lorenzi et al. 2010), it becomes clear that açaí from the states of Amazonas, Acre and Rondônia is extracted from fruits of E. precatoria.However, in the states of Pará, Maranhão and Amapá, it is extracted from Euterpe oleracea, another açaí species.Euterpe precatoria is single-stemmed, while E. oleracea has tillers (multi-stemmed).One of the advantages of E. precatoria is that the fruit pulp has better anti-oxidant and anti-inflammatory properties than E. oleracea (Kang et al. 2012).Moreover, the evolutionary adaptation to nutrient-poor and well-drained lowland, Latosolos and Argisolos is noteworthy (FAO 1987).The adaptation of E. precatoria to the Amazonian ecosystem indicates it for exploitation in the recovery of degraded areas, although no seeds of improved varieties recommended for the production of seedlings and planting are available, since to date, E. precatoria has been neglected in research.
The inflorescences of E. precatoria have numerous male (4.5 x 2.7 mm) and female (3.2 x 2.6 mm) flowers.The male flowers open and release pollen before the female flowers are receptive (Küchmeister et al. 1997, Lorenzi et al. 2010).However, this observation of protandry is no guarantee that the species is autogamous, since nonetheless some crosses may occur.Thus, research must be carried out to fill gaps in the knowledge about the gene flow and mating system to interpret its behavior within tropical forest ecosystems in in situ or ex situ conservation areas.
Euterpe precatoria is a highly important species for the Amazon region.As the genetic structure of tree populations is partially determined by the mating system and strongly by the gene flow among populations (Ramos et al. 2016a), this study will provide valuable information for the domestication and rational management of the species (Ramos et al. 2011).Studies of the mating system can be based on microsatellite markers or simple sequence repeats (SSRs), which are an adequate tool for this purpose, due to the high polymorphism in terms of number of alleles, co-dominant inheritance and low cost of the method (Wadt et al. 2015).Simple sequence repeat-based studies to determine the mating system were carried out for different Amazonian and tropical tree species (Ramos et al. 2011, Medina-Macedo et al. 2015, Picanço-Rodrigues et al. 2015;Moraes et al. 2018).
The objective of this study was to investigate the mating system of a natural E. precatoria population, to identify the levels of outcrossing, mating among relatives and correlated matings to better understand the genetic structure of open-pollinated progenies for ex situ conservation and breeding plans.Moreover, we tried to determine the coancestry coefficient, effective population size within progeny and the estimated number of trees to ensure enough seeds for conservation.

Study area and sampling
The study investigated a natural E. precatoria population in the community of Nossa Senhora de Fatima do Açaí, in the rural area of Vila Amazônia (lat 02º 36' 52.09" S, long 56º 33' 29.13" W, and alt around 40 m), in Parintins, state of Amazonas, Brazil, where the climate is tropical monsoon (Am) (Peel et al. 2007).All 13 trees bearing fruit within 10 hectares of climax forest were used for the study.Plant material (leaflets) and 100 fruits per mother tree were collected, i.e., a total sample of 1300 fruits.The fruits were transported to the seed laboratory of the experimental field Caldeirão of Embrapa Western Amazon, located at Rodovia Manoel Urbano, Km 13, Estrada do Caldeirão, Iranduba-AM, for seed germination, emergence and seedling production.To accelerate germination, fruits with pulp and seeds were immersed in water at 60 °C and removed when the water reached room temperature (Nogueira et al. 1995).After three months, seedlings of all progenies with at least two bifid leaves were sampled for DNA extraction and microsatellite analysis, resulting in 3 to 25 seedlings per progeny, totaling 227 seedlings (Table 2).One leaflet of each seedling (227 samples) and mother tree (13 samples) was collected and packed separately in a previously labelled zip lock plastic bag containing silica gel.These samples were transported to the Laboratory of Genetics and Plant Breeding of the Department of Agronomy of the Federal University of Amazonas (UFAM) for storage at -20 °C for further genomic DNA isolation.

RESULTS AND DISCUSSION
The maternal fixation index (F ̂m) was not different from zero, indicating no inbreeding level of the mother trees.A similar result was found for Malpighia emarginata (Lopes et al. 2002).The multilocus outcrossing rate (t̂m) and singlelocus outcrossing rate (t̂s) were equal to 1 (Table 1), indicating that the progenies were originated by outcrossing and that the species is allogamous.A similar finding was reported for the palm Astrocaryum aculeatum (Ramos et al. 2011).The difference between the multilocus and the single-locus outcrossing rate measured as the mating rate among related individuals (t̂m-t̂s), was zero (0.001), evidencing the absence of mating between relatives of the sampled seedlings.
The estimated selfing correlation (r̂s) among progenies was low (0.11), indicating low variation in t̂m among mother trees.Accordingly, the t̂m among seed trees ranged from 0.79 to 1, in other words, it was significantly lower than 1.0 in 11 of the 13 families (Table 2), showing that some seedlings were originated by inbreeding.This result also suggests that the species is self-incompatible (Moraes et al. 2018).The individual variation in the outcrossing rate may be attributed to flowering asynchrony.
The paternity correlation (r̂p (m) ) was moderate (0.293) and indicated that about 3.4 effective pollen donors (N ̂e (p) ) fertilized the seed trees in the investigated reproductive event (Table 1).These results show that the families are mainly composed of half-sibs (71%) and that mating was not random.The highest N ̂e (p) was reported for the palm A. aculeatum (5.7, Ramos et al. 2011).Correlated matings can be caused by the behavior of pollinators that systematically visit nearby trees (Spoladore et al. 2017).Non-random matings were also evidenced by the variation among families for mating among relatives (t̂m-t̂s: between 0.05 -0.35), paternity correlation rates (r̂p (m) : 0.08 -0.58) and effective number of pollen-donor trees (N ̂e (p) : 1.7 -11.9, Table 2).The mean coancestry coefficient within progenies (Θ = 0.162) was higher than expected in half-sib progenies (0.125).Thus, estimates of addictive genetic variance and heritability must be calculated using a relatedness coefficient (Sobierajski et al. 2006) of 0.324 (2Θ) instead of 0.25.Knowledge about the coancestry coefficient is also important when estimating the variance effective size (N ̂e), which was lower (N ̂e = 2.76) than expected in the random mating populations (4, Furlani et al. 2005).In infinite samples of progeny structures, the variance effective size varies from 1 to 4, where value 1 indicates selfed progenies, 2 full-sibs and 4 half-sib progenies (Sebbenn 2006).Due to the estimated effective population size, the seeds must be collected from at least 54 trees to retain the effective reference population size of 150 in progeny array samples.These analyses are important for estimating sample sizes in breeding, genetic conservation and seed collection programs addressing environmental recovery, as well as for the monitoring of genetic diversity in manipulated populations (Moraes et al. 2018).

CONCLUSION
The mating system indices estimated for E. precatoria in this study indicated that the species is allogamous but self-compatible.The studied progenies were mainly represented by half-sibs, but matings were not random due to the occurrence of some correlated mating, resulting in a few full-sibs within progenies.

Table 1 .
Mating system at the population level