Genetic diversity trends in sugarcane germplasm : Analysis in the germplasm bank of the RB varieties

Brazil is the largest sugarcane producer in the world and, the main varieties grown in Brazil, known as RB cultivars, were developed by the Interinstitutional Network for the Development of the Sugar and Alcohol Sector (RIDESA) and are used in 58.9% of the planted area in Brazil. These varieties were obtained through intercrosses between genotypes from the Serra do Ouro germplasm bank and successive crosses with related genotypes may have increased the level of genetic similarity. The aim of the present study was to analyze the genetic base of the Serra do Ouro germplasm bank over the past decades using microsatellite molecular markers. The genetic similarity among varieties using all the markers ranged from 0.166 to 0.823, and regression analysis showed an increase in genetic similarity in the 1970s; however, a narrowing of the genetic base over the last five decades was not observed.


INTRODUCTION
In the early twentieth century, researchers in India and Java developed interspecific hybrids between the polyploid species Saccharum officinarum and S. spontaneum while maintaining noble and rustic features, respectively (Grivet et al. 2004).After successive backcrosses, modern hybrids have approximately 80% of the S. officinarum genome, 10% from S. spontaneum, and 10% recombinant between the two genomes (D'Hont et al. 1996).Sugarcane currently is very important to the world economy because of the production of sugar, ethanol, and energy from its biomass.In Brazil, sugarcane cultivation occupies a total area of 9.8 million hectares, with an output of 739 million tons (FAO 2013).The development of new varieties has contributed to expanding cultivation because the new genotypes are resistant to biotic and abiotic factors.The Interinstitutional Network for the Development of the Sugar and Alcohol Sector (RIDESA) and the Sugarcane Technology Centre (CTC) are currently the main centers for the development of new varieties of sugarcane in Brazil.They are responsible for developing the cultivars RB (República do Brasil) and SP (São Paulo), respectively.According to Barbosa et al. (2012), RB cultivars comprise 58.9% of the varieties that are planted and cultivated in Brazil, while the varieties developed by the Sugarcane Technology Centre comprise 35.8%.
Cultivars developed in the latter half of the twentieth century by RIDESA and CTC programs, although genetically distinct, have very high parental similarity because of the small number of hybrids used as a breeding base (Santos et al. 2012).In subsequent decades, intercrossing between modern hybrids became the main characteristic for developing new varieties.This plant group is characterized as aneuploid and polyploid hybrids, where recombination is the factor responsible for the genetic diversity of modern sugarcane accessions.Considering the genetic narrowing that has been caused in other crops by the use of a few clones as the genetic base (Wouw et al. 2010), there is the possibility of genetic narrowing in modern sugarcane hybrids.
Microsatellite markers are prominent among molecular tools for the genetic analysis of germplasm because they are highly polymorphic, allowing the analysis of genetic similarity between morphologically similar plant materials.This type of marker has been used in crops such as rice (Wu and Tanksley 1993), barley (Saghai-Maroof et al. 1984), and wheat (Roder et al. 1995).In the genus Saccharum, these markers can be used for the preselection of progenies in a breeding program to germplasm studies (Cordeiro et al. 2000, Cordeiro et al. 2001, Cordeiro et al. 2003, Santos et al. 2012, Silva et al. 2012).The objective of the present study was to analyze the basis of the genetic pool of sugarcane over the years using microsatellite molecular markers.

Plant material
A set of 47 parents of sugarcane (Table 1) was obtained from the germplasm bank of "Serra do Ouro."The genotypes were classified according to the decades in which they were developed (year of crossing) and comparative evaluations were performed by means of contrasts restricted to cultivars of the same decade (Table 1).

DNA extraction and amplification of the SSR Loci
DNA was isolated from the tissues of young leaves stored at −80 °C, using the procedure described by Saghai-Maroof et al. (1984).The integrity and concentration of each DNA sample was determined by spectrophotometry (absorbance at 260 and 280 nm) and gel electrophoresis in 1% agarose, respectively.Working stocks containing 50 µL were prepared at a concentration of 25 ng µL −1 .We used four SSR markers (SCC03, SCC05, SCC06 and SCC93) developed by the laboratory of genetic resources at the Arapiraca Campus of the Federal University of Alagoas.The names of the primers are listed in Table 2 and the sequences of the primers are found in Duarte Filho et al. (2010) and Silva et al. (2012).
For PCR amplification, a final volume of 50 µL was used containing: 25 ng of genomic DNA, 10× buffer, 2.0 mM MgCl2, 0.2 mM dNTP, 2.5 U of Taq-DNA polymerase, 30 pmol of each primer (forward and reverse) and sterile distilled water.The PCR amplifications were carried out in a thermal cycler (Applied Biosys-tems), with the following conditions: one cycle at 94 0 C over 3 min for pre-denaturation, followed by 35 cycles at 94 0 C for 1 min, 62 0 C for 1 min and 72 0 C for 1 min and a final extension at 72 0 C for 12 min.Forward primers for each pair were labeled with different fluorescent dyes (6FAM or HEX) and arranged in duplex for analysis in an automated analyzer for DNA fragment analysis (ABI-3100, Applied Biosystems).The ROX 500 (Applied Biosystems, Foster City, CA) was used to accurately determine the

DATA ANALYSIS
Despite being considered co-dominant SSR markers, in this study they were considered as dominant markers, because in highly polyploid genomes such as that of sugarcane, the SSR markers have difficulty distinguishing the alleles of homologous chromosomes, making it difficult to determine heterozygosity or homozygosity at any particular locus (Cordeiro et al. 2003, Oliveira et al. 2009).From this assumption, all possible alleles detected in the varieties have been converted to a binary system.For each clear and distinct peaks were classified as absent (0) or present (1) to form the matrix that was used to estimate the following variables: Number of alleles with absent or present among genotypes (Np).

Number of alleles without absent among genotypes (Nnp).
Polymorphism information content (PIC) obtained through the expression: PIC = 2fi(1 -fi), where fi is the frequency of the amplified fragments and 1-fi is the frequency of the non-amplified fragments (Roldan-Ruiz et al. 2000) and the PIC for each primer was obtained using the average from all fragments.
Genetic similarity (GS) between pairs of genotypes of the same decade (Table 1) was calculated using the Jaccard coefficient, obtained by the expression: , where GS is the measure of genetic similarity between genotype i and j.For each pair of accessions, "a" represents the number of coincidences of the type 1−1, "b" the number of coincidences 1-0 and "c" the type of 0-1 (Reif et al. 2005).
Linear regression between genetic similarity and decade of the genotypes was applied for best-fit model.

RESULTS AND DISCUSSION
Sugarcane hybrids are polyploid and aneuploid and may have variable numbers of chromosomes; therefore, they exhibit high levels of allelic variation among individuals in the same locus (Santos et al. 2012).In the present study, the number of alleles per locus among genotypes analyzed varied between 3 and 16, with a mean of 7.5 (Table 2).The four loci amplified a total of 90 alleles, ranging in size from 81 to 236 bp.The number of alleles for all individuals at each locus ranged between 17 and 27 with a total average of 22.5 alleles (Table 2).The four SSR markers used had a number of alleles for discriminating all the subjects analyzed.This ability is the result of sugarcane's aneuploid and polyploid nature, with many alleles in the same individual, which increases the discriminatory capacity of the microsatellite markers (Cordeiro et al. 2001, Pinto et al. 2004, Silva et al. 2012).Previous studies using the Serra do Ouro germplasm bank have shown that microsatellite markers are highly efficient in distinguishing different genotypes (Santos et al. 2012, Silva et al. 2012).The polymorphic information content (PIC) using all the markers was 0.925, while individually, the microsatellites SCC03, SCC05, SCC06, and SCC93 showed values of 0.942, 0.938, 0.897, and 0.913, respectively (Table 2).The genetic similarity between varieties using all the markers ranged from 0.166 to 0.823 (Table 2).The genetic similarity among genotypes within decades showed that the 1980s had the highest similarity coefficient, corresponding to 0.823, while the lowest genetic similarity, with a value of 0.241, was found in the 1990s.
To analyze the genetic similarity as a function of time, a regression model was fitted to the Jaccard similarity coefficient.The results showed quadratic regression as the most suitable statistical model for describing the data, in view of the significance of the model parameters.The results indicated an increase in genetic similarity during the period 1970-1980, while the varieties developed in the 1990s have levels of genetic similarity close to those from the 1960s (Figure 1).
The use of just a few genotypes in the sugarcane breeding programs during the 1970s and 1980s, which mainly aimed at increasing sucrose levels, was one of the factors that raised the levels of genetic similarity.In the present study, the average similarity was 0.448, indicating values near those found by Lima et al. (2002), Pinto et al. (2006) who reported moderate genetic similarity coefficients (0.48 and 0.62, respectively) among accessions used in three sugarcane genetic breeding programs.From the 1990s onwards, there was a change in sugarcane breeding strategy because of the emergence of new pests and diseases and the need for cultivars that were more adapted than the existing ones.Given the new scenario, the breeding programs included new parent varieties, contributing to a lower genetic similarity index among the varieties developed after 1990.
Analyses of levels of genetic diversity in various crops showed a narrowing trend of the genetic base during the 1960s, 1970s, and 1980s (Wouw et al. 2010.This behavior was observed in the present study, with the highest similarity levels in the 1970s and 1980s (significant quadratic parameter); however, there was not a linear increase or decrease of the genetic bases.The increase in similarity levels during the 1970s and 1980s resulted from interbreeding with few parents; after the 1990s, with the introduction of new parents, there was a decrease in genetic similarity levels.It should be noted that the levels of genetic diversity in sugarcane breeding have remained the same over the past five decades, and that  D Crystian et al.
the varieties developed resulted from the genetic diversity of the first hybrids.These were obtained by interbreeding S. officinarum and S. spontaneum, resulting in hybrids with different genomic constitution ratios (Piperidis et al. 2010) because of different chromosomal proportions (10-20% of S. spontaneum and 80-85% of S. officinarum).This condition allows a large number of combination ratios of the two genomes during chromosomal segregation in meiosis, which can be exploited in breeding.However, given the new requirements for sugarcane, such as the use of fiber for second generation ethanol, coproduction of electricity, and use of fiber as an organic fuel, new parents with different ratios of S. spontaneum should be incorporated into crossings.

Table 1 .
Sugarcane cultivars differentiated according to the decade of development and used to assess the genetic pool of the fragments detected by capillary electrophoresis.These were then visualized in the form of peaks with the respective sizes and intensities and analyzed by Peak Scanner Software (Applied Biosystems). size

Table 2 .
Polymorphic Information Content (PIC), Jaccard coefficient, Number of alleles with absent or present among genotypes (N p ), Number of alleles only with present among genotypes (N np ), allelic interval (bp) and number of alleles per individual

Table 3 .
Genetic Similarity among genotypes by decade