Genetic variability and homozygosity in a F 4 castor bean population by microsatellite markers

Bragantia, Campinas, Ahead of print, 2016 Resumo: Os objetivos deste trabalho foram identificar a variabilidade genética e estimar o nível de homozigose em uma população F4 de mamoneira usando marcadores microssatélites (SSR). Para tanto, foi realizada a genotipagem da população por meio de 53 pares de iniciadores SSR, e as frequências alélicas, número de alelos por loco, heterozigosidade esperada (He), heterozigosidade observada (Ho) e conteúdo informativo de polimorfismo (PIC) foram estimados. Uma matriz de dissimilaridade genética foi gerada pelo índice de Nei e Li, e a análise de agrupamento hierárquico foi realizada por meio do método Unweighted Pair-Group Method Averages (UPGMA). Foi detectado polimorfismo em um total de oito loci (15,09%) dos 53 avaliados, com presença de dois alelos por loco. As frequências alélicas variaram entre 0,71 e 0,53, e o PIC, entre 0,32 e 0,37. A heterozigosidade média observada Ho (0,30) foi menor que a heterozigosidade esperada He (0,47). Houve formação de cinco grupos dissimilares, mostrando que há variabilidade genética entre os genótipos avaliados. A maior dissimilaridade genética foi de 0,708 e a menor, de 0,00. As porcentagens de homozigose entre os genótipos variaram de 25 a 75%. Esses resultados mostram que a autofecundação controlada em mamoneira eleva o nível de homozigose, importante para o programa de melhoramento genético. Palavras-chave: Ricinus communis L., melhoramento genético, seleção. Variabilidade genética e homozigose em uma população F4 de mamoneira por meio de marcadores microssatélites Genetic variability and homozygosity in a F4 castor bean population by microsatellite markers Edna Lobo Machado*, Simone Alves Silva, Luciel dos Santos Fernandes, Helison Santos Brasileiro Universidade Federal do Recôncavo da Bahia Centro de Ciências Agrárias, Ambientais e Biológicas Cruz das Almas (BA), Brasil.


ABSTRACT:
The objectives of this study were to identify the genetic variability and estimate the level of homozygosity in a castor bean F4 population using microsatellite markers (SSR).
To this end, it was performed the genotyping of the population through 53 pairs of SSR primers.Allele frequencies were estimated by number of alleles per locus, expected heterozygosity (H e ), observed heterozygosity (H o ) and polymorphic information content (PIC).An array of genetic dissimilarity was generated by Nei and Li index, and hierarchical cluster analysis was performed using the Unweighted Pair-Group Method Averages (UPGMA) method.
Polymorphism was detected in a total of eight loci (15.09%) of

INTRODUCTION
Genetic improvement of plants is a lengthy process that takes, on average, 12 years for annual species and 30 years for perennial species (Bespalhok et al. 1999).Therefore, the breeder should try to anticipate future needs.One of the challenges is to search for renewable energy sources.Given the above, several breeding programs have worked with species that can be used to produce alternative fuels such as ethanol (sugarcane) or biodiesel (rapeseed, sunflower and canola) (Bespalhok et al. 1999).The castor bean is an oleaginous of great social and economic importance in Brazil, especially for the Northeastern region, where the culture generates jobs and income for small and medium farmers (Freitas et al. 2010).
Ricinus communis L. is a monotypic species belonging to the Euforbeaceae family and considered self-pollinated (Távora 1982) because of its reproductive system.However, due to the high rate of outcrossing, it is a better fit for the mixed or intermediate plants classification . Vencovsky et al. (2001) stated that a mixed population has a crossing rate between 5 and 95%; these values represent averages on population, not on the individual.Thus, castor bean is an interesting plant for breeding programs since it is easy to obtain crossings and controlled self-fertilization.
The fundamental condition to implement a breeding program is the existence of genetic variability.In the search for variability, crossings are made to maximize the genetic diversity of the population, and, once this variability is achieved in a segregating population, it is necessary to identify the superior homozygous genotypes obtained through selfing in advanced populations.Selffertilization is the process that leads to more intense degree of inbreeding (Maia et al. 2008).It is easier to predict the behavior of the progenies in these populations since segregation of alleles in homozygous does not occur.Thus, the identification of variability and level of homozygosity achieved in the pairs of alleles in genotypes derived from an F4 population gives an indication of how many heterozygous loci are present in the study population and how many generations of self-fertilization are needed to achieve the expected homozygosity.
The detection of genetic variability and the level of homozygosity of an advanced population such as the F4 castor bean can be done using molecular markers.Microsatellite (SSR) markers are informative due to their codominance, high degree of polymorphism, reproducibility and easy interpretation (Borém and Miranda 2009).It is important to point out that genetic variability studies and degree of homozygosity using SSR markers in an advanced castor bean population are still unprecedented.Given the above, the objectives of this study were to identify the genetic variability and estimate the level of homozygosity of an F4 castor bean population using SSR markers to monitor the species breeding.

MATERIAL AND METHODS
The F4 population with 32 individuals used resulted from the crossing between the BRS Nordestina and Sipeal 28 cultivars, followed by selfing.These cultivars were selected because they were different and suitable for the Northeastern region (Bahia et al. 2008).
The seeds of the genotypes were germinated in washed sand inside plastic basins.After 15 days, healthy leaves were collected for DNA extraction according to the protocol described by Doyle and Doyle (1990).The quantity and quality of DNA were evaluated by comparative analysis with known concentrations of lambda DNA (Invitrogen, Carlsbad, CA, USA) in 0.8% agarose gel, stained with ethidium bromide (0.5 mg•mL -1 ).The samples were diluted in Tris-EDTA to adjust concentration (5 ng•μL -1 ).
The amplifications were carried out in thermocycler Biocycler MJ96+/MJ96G (Applied Biosystems of Brazil Ltda., São Paulo, SP) at initial denaturation temperature of 94 °C for 1 min followed by 35 amplification cycles [94 °C for 1 min, annealing temperature ranging from 62 to 66 °C, depending on the primer pair (Table 1), for 1 min and 72 °C for 1 min] and final extension at 72 °C for 10 min (Bajay et al. 2011).The amplified products were subjected to electrophoresis in 7% denaturing polyacrylamide gel, following the protocol proposed by Litt et al. (1993).The gel was stained with silver nitrate (Creste et al. 2001) and, after dried, was scanned (HP ScanJet N8460, Hewlett-Packard Company, Palo Alto, CA, USA).The amplitude of the fragments was estimated by comparison with a standard molecular weight of 50 bp (Invitrogen, São Paulo, SP), and the loci were characterized as either polymorphic or monomorphic due to the existence of different alleles between the genotypes.
Estimates of allele frequencies, number of alleles per locus, high allele frequency (HAF), expected heterozygosity (H e ), observed heterozygosity (H o ) and polymorphic information content (PIC) were obtained using the PowerMarker software version 3.25 (Liu and Muse 2005); the genetic dissimilarity matrix was obtained using the Nei and Li index (1973).
A hierarchical cluster analysis was performed using the Unweighted Pair-Group Method Averages (UPGMA) method (Sneath and Sokal 1973).The dendrogram was constructed using the MEGA 5 software (Tamura et al. 2011).
The level of homozygosity of the evaluated genotypes was estimated following the procedure adopted by Mendes et al. (2006) where: P is the proportion of completely-homozygous genotypes; m represents the number of generations; n is the number of alleles.
The chi-square test (χ 2 ) at 5% was conducted to check the differences between the expected and observed values using the Genes software (Cruz e Carneiro 2003).

RESULTS AND DISCUSSION
In the studied population, polymorphism was detected in eight loci (15.09%) of the 53 evaluated ones.The polymorphic SSR loci and the sequences of primers are shown in Table 1. Figure 1 shows the polymorphism of the Rco23 locus in the 32 evaluated genotypes.A smaller percentage of polymorphism (9.72%) was observed by Pranavi et al. (2011) in nine commercial castor bean hybrids using 92 EST-SSR markers.However, a higher polymorphism (31.13%) was detected in 24 castor bean samples collected in different countries (Qiu et al. 2010).These authors analyzed 379 EST-SSR loci.
Two alleles were identified for each SSR locus (Table 1).This result was expected because, in a segregating population that originated from the cross between two parental diploids, four alleles at most could be found.This population resulted from the crossing between two cultivars.In general, commercial cultivars have a high degree of homozygosity that resulted from the breeding process.Bajay et al. (2009Bajay et al. ( , 2011) ) identified, in 12 and 11 polymorphic SSR loci, two to five alleles per locus, for the 38 and 76 castor bean genotypes, respectively.The higher number of alleles found by the authors is explained by the fact that it was not a self-fertilization population, such as the castor bean F4 one analyzed in this study.
For the evaluated loci, the PIC ranged from 0.32 (Rco11 and RcoM28 loci) to 0.37 (Table 1).The highest PIC (0.37) was detected for most loci: Rco2, Rco23, Rco29, RcoM2 and RcoM11 (Table 1).Bajay et al. (2009) found PIC value for SSR markers in castor bean ranging from 0.1703 to 0.6597.This greater variation in PIC value can be explained by the greater number of genotypes evaluated by these authors.
The average PIC was 0.36.In general, the loci were moderately informative (Table 1).According to Botstein et al. (1980), PIC values are grouped into three levels: highly informative (PIC > 0.5); moderately informative (0.25 < PIC < 0.5) and little informative (PIC < 0.25).The PIC estimates the discriminatory power of the locus, taking into account not only the number of identified alleles but also their relative frequencies.
The average observed heterozygosity H o (0.30) was lower than the expected heterozygosity H e (0.47).This result agrees with what is expected for a segregating F4 population, which characteristically has a higher homozygosity level with allele fixation.Low H o values (0.416 and 0.0922) were also found by Bajay et al. (2009Bajay et al. ( , 2011)), respectively.
Table 3 shows the homozygosity analysis of 32 genotypes.The percentage of homozygosity among genotypes ranged between 25 and 75%, with emphasis on genotypes 13, 20, 23 and 28, which showed the highest value (75%), while genotypes 2, 10, 16 and 26 had the lowest value (25%).The analysis of homozygosity also resulted in the formation of five groups (Table 3).
According to Allard (1971) and Borém and Miranda (2009), the proportion of homozygous plants is a function of the number of allelic pairs and the number of selfing generations.Over five generations of selfing, in five separate/independent genes, 85% of the population genotypes are homozygous in all five loci.However, considering three generations of selfing in the F4 castor bean population, in which the ratio of homozygotes increases and of heterozygotes decreases, and eight loci are involved, it is expected that 34% of the genotypes are in complete homozygosity.Thus, the average value obtained for the homozygosity level of the genotype was high (52.7%),with quick setting of homozygous alleles and possible genetic gain from selection, considering another generation of selfing.
The detection of heterozygous alleles with molecular markers is more sensitive than the assessment of the genetic constitution using phenotypic data.So, for the proportion of five independent loci, after three generations of selfing, 51% of the population genotypes would be expected to be in homozygosity in all eight loci (Allard 1971;Borém and Miranda 2009), thus corroborating the homozygosity value obtained in this work.
The c 2 test to analyze the differences between the expected and observed homozygosity level for genotypes can be found in Table 4.In this case, the calculated values of the c 2 test at 5% probability for RcoM27 (18.892) and RcoM11 (15.621) initiators were significantly different.The probabilities associated with these values are 7.90 and 4.05%, respectively.This indicates that some factor other than chance is responsible for the differences between the observed and expected values.
The value calculated for the other initiators did not differ significantly by the c 2 test at 5%.The highest probabilities were found for the primers Rco23 (99.75%) and RcoM28 (91.66%), indicating a high probability that the difference between observed and expected values happened by chance.
The analysis of genotypes of F4 castor bean population using SSR markers showed that most of the alleles are  homozygous, suggesting a strong inbreeding, which may have resulted from selfing processes.These studies demonstrate the efficiency of using self-fecundation in a population to achieve high level of homozygosity, important for the breeding programs of castor bean.

CONCLUSION
The homozygosity is high in F4 castor bean population, with good fixation of homozygous alleles over generations

Figure 2 .
Figure 2. Genetic relationships among the 32 genotypes of castor bean based on SSR markers generated by the UPGMA method.

Table 1 .
: Loci , SSR primer sequence, reason (microsatellite repeat), amplitude of fragments, high allele frequency, number of alleles per locus, expected heterozygosity, observed heterozygosity, polymorphic information content in eight polymorphic loci identified in a F4 Ricinus communis L. population, Cruz das Almas, Bahia, Brazil, 2015.T a = Optimal annealing temperature of the initiators; HAF = High allele frequency; NA = Number of alleles per locus; H e = Expected heterozygosity; H o = Observed heterozygosity; PIC = polymorphic information content.

Table 2 .
Genetic dissimilarity matrix among 32 genotypes of an F4 castor bean population obtained from the Nei and Li dissimilarity coefficient(1973).

Table 3 .
Percentage of homozygous genotypes of the F4 castor bean population from microsatellite markers.

Table 4 .
c 2 test for the SSR loci in the F4 castor bean population that originated from the crossing between the BRS Nordestina and Sipeal 28 cultivars.
*Significantly different.DoF = Degrees of freedom