Characterization of wild genotypes of Aroeira: Subsidy for plant breeding

The selection of genetically superior materials of different forest species with economic potential involves studies of germplasm characterization. Facing the exploitation of aroeira (Schinus terebinthifolius Raddi.) in areas of natural occurrence the genetic and chemical characterization of individuals is extremely important to suit the selection. The work was carried out to identify genetic and chemical differences in individuals of aroeira located in different forest fragments. The genetic polymorphism was accessed using 19 inter-simple sequence repeat (ISSR) primers and the fragments generated were used to evaluate the Nei's gene diversity, percentage of polymorphic loci, Shannon index, number of alleles observed and effective number of alleles. The fruits of all individuals were submitted to hydro distillation of essential oils. The oil chemistry composition was analyzed by Gas chromatography/mass spectrometry (GC/ME) and the averages of the compounds were compared by Scott-Knott (p<0.05). The correlations using genetic, geographic distances and chemical composition were analyzed by the Mantel test. There are considerable levels of genetic and chemistry variability among the individuals studied, which allows markers assisted selection aiming to identify potential genitors for use in plant breeding.


INTRODUCTION
The potential of native species with agro-industrial use is endless.Brazil is considered to have the greatest number of higher plant species identified, and majority of these native plants contain essential oils (Mittermeier and Werner, 1990) with potential industrial use.However, Brazil's biodiversity is being destroyed at an accelerated pace due to city expansions and human developments.Reduction of genetic diversity implies reduction of opportunity to identify natural products with economic use, as many native species are important sources of biologically active natural compounds.For example, along the Basin of Lower São Francisco river in Sergipe, grows aroeira (Schinus terebinthifolius Raddi.), a native plant species extensively explored due to the chemical properties of its essential oils that are used in medicine, cosmetics and food.Studies on the essential oil of aroeira collected in different regions of the globe, has shown different chemotypes, with prevalence of different chemical compounds (Silva et al., 2010a).Plants in India, for example, higher content of α-pinene has been found *E-mail: renatamann@hotmail.comAuthor(s) agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0 International License (51%) (Singh et al., 1998;Chowdhury and Tripani, 2001), whereas in plants of Egypt the highlight was αphellandrene (24 %) (Ibrahim et al., 2004).
Genetic diversity within a species like S. terbinthifolius, risks being depleted due to extensive development of cities without appropriate management practices, which may lead to the loss of the genotypes with the most favorable characteristics, and possibly only plants of lower agro-industrial value.
Modern breeding relies on associating phenotypic differences with genetic characterization using molecular markers based on the analysis of nucleotide sequence differences and the identification of specific sequence differences between populations (Telles et al., 2010).Therefore, the identification of phenotypic differences of the chemical makeup of essential oils allows the selection of characters and individuals of economic importance that can then be analyzed genetically in breeding programs.The basic premise that this study proposes for aroeira breeding of is to estimate the genetic variation present in the species that could be estimated using ISSR molecular markers (Julio et al., 2008;Brandão et al., 2011).Thus, the aim of the present work was to phenotypically and genotypically characterize individuals of aroeira located in riparian forest fragments along the region of Lower São Francisco river in Sergipe, Brazil.

Plant sampling
The study was performed in riparian forest fragments of different phytogeographic regions of the State of Sergipe, Brazil, located in the municipalities of Brejo Grande, Neópolis and Propriá.

Collecting plant material
For molecular analysis, tender leaves of individual aroeira plants were collected and stored at -20°C until DNA isolation.Mature fruits were also collected from the same individual plants for chemical analysis of essential oil composition.

DNA isolation
The isolation of genomic DNA was performed from approximately 2 g of tissue according to the protocol described by Nienhuis et al. (1995).The quality and concentration of DNA from each sample was determined by spectrophotometry (CARY 50 PROBE UV-VISIBLE), based on absorbance at 260 and 280 nm.
The PCR products were separated by horizontal electrophoresis, using 1% agarose gel in 0.5X TBE buffer, at 100V for 90 min.Subsequently, the gels were stained with ethidium bromide (0.5 μg ml -1 ) for 30 min, visualized and photographed under exposure to ultraviolet light.The size of the amplified fragments was estimated by comparison with molecular markers within a 100 bp DNA ladder.

Extraction of essential oils
The essential oils from 200 g of aroeira fruits were hydro distilled in a Clevenger apparatus during 180 min.The recovered oils were stored at -20°C in amber bottles.The essential oils content, expressed as a percentage of total volume, was obtained based on 100 ml.g -1 weight of matter.

Gas chromatography-mass spectrometry (GC/ME)
The chemical composition was determined by gas chromatography coupled to a mass spectrometer (GC/ME, Shimadzu, modelo QP 5050A), equipped with a fused silica capillary column DB-5 of 30 m x 0.25 mm i.d., 0.25 μm coating, using He as the carrier gas with a flow rate of 1.0 ml min -1 .The temperature was maintained at 50°C for 1.5 min, followed by an increase of 4°C min -1 until 200°C, and then 10°C min -1 to 250°C, keeping this temperature constant for 5 min.The injector temperature was 250°C and detector temperature (or interface) was 280°C.Forty milligram of the oil was dissolved in 1.5 ml of ethyl acetate and injected into a volume of 0.5 µl of the solution, at a partition volume ratio of 1:100 and injected on the column at a pressure of 64.20 kPa.
The conditions of ME were ion detector of the quadrupole, operating by electron impact and impact energy of 70 eV, scan speed 1.000; scan interval of 0.50 s -1 fragments and fragments detected in the range of 40-500 Da.The identification of oil components was based in MS data comparison followed by retention index and compared to the literature retention indices (Adams, 2007).For the retention rate, the van den Dool and Kratz (1963) equation was used in relation to a homologous series of nalcanos (nC9-nC18).Were also used three libraries of the equipment WILEY8, NIST107 e NIST21, which allow the comparison of the spectral data with those contained in the libraries, using a similarity index of 80%.

Molecular data
The ISSR fragments were used to obtain a binary array, considering the presence (1) or absence calculating the percentage of polymorphism obtained using each primer.The fragments with weak staining and low resolution were not considered.The genetic similarity index (Sij) between each pair of individual plants was estimated using the Jaccard coefficient, using the program NTSYS pc2.1 (Rohlf, 2000).
In order to verify the correlation between genetic and geographic distances of individuals analyzed, we used the Pearson correlation coefficient (r) (Pearson, 1892), by Mantel (Manly, 1997).Simplified representation of the similarity data was exposed in a dendrogram obtained by the clustering method Unweighted Pair-Group Method Arithmetic Average (UPGMA) (Sneath and Sokal, 1973), using the computer program NTSYS pc2.1, and the consistency of each group performed by bootstrap (Coelho, 2000).
Given the dominant nature of the data, the program required the calculation of allele frequencies, to verify the loci Hardy-Weinberg equilibrium.The percentage of polymorphic loci (P%), number of alleles observed, number of effective alleles, the genetic diversity of Nei (1973) (Ĥ e ) and the Shannon Index (I) for individuals were estimated.To characterize the genetic diversity, the program POPGENE (Population Genetic Analysis) versão 1.32 (Yeh et al., 1999) was used, employing parameters for diploid dominant data.

Chemistry data
The performance data, content and major compounds of the essential oils were analyzed by ANOVA and means compared by Scott-Knott (p<0.05)(Ferreira, 2000).The presence (1) or absence (0) of the chemical components was used to obtain a binary matrix.
The chemical similarity index ( ij Sg ) between each pair of individuals was calculated using the Jaccard coefficient, and the simplified representation of the data was expressed in a dendrogram obtained by the UPGMA method, using the NTSYS pc2.1 (Rohlf, 2000).The chemical composition data were analyzed by the UPGMA method, considering the Euclidean distance as similarity coefficient (Cruz et al., 2004), and the principal components analysis (ACP), both using the program STATISTICA versão 6.0 (Statsoft, 2001).The correlation between the chemical profile and geographical distance, and genetic similarities were performed by Mantel (Manly, 1997) with 1,000 permutations.

RESULTS AND DISCUSSION
Five individuals of each fragment were selected, taking into account phenotypic characteristics such as height, trunk and branches with minimum levels of pests and pathogens and evident by abundant fruiting (Table 1).

Molecular characterization
Analyses of gel electrophoresis images identified 104 ISSR fragments showing polymorphism over an averaging of 76.92% of the total fragments obtained (Table 2).ISSR molecular markers have been successfully used to estimate of genetic variability in wild and cultivated species, including medicinal and aromatic species (Santana et al., 2011).The number of polymorphic fragments used in these analyzes is quite variable (Estopa et al., 2006).Using ISSRs primers, Brandão et al. (2011)  (80.50%).Genetic polymorphism observed among individuals of aroeira can be considered high, implying a low level of homozygosity among them, which can partly explain the phenotypic variation observed in the field.The existence of heterozygosity among individuals can be beneficial to a plant breeding program, and can be used as matrix trees for seed collection and would be the genotypes of choice for use in restoration of degraded areas as these plants would potentially add the maximum genetic diversity to an area.With 80 polymorphic fragments, a dendrogram of genetic similarity was generated (Figure 1).The mean genetic similarity among individuals was 43.31% and the range of similarities ranged 15-72%, which confirms the data of genetic polymorphism.P1 and P2 are the most similar individuals, and the most divergent were BG3 and N2.
By dendrogram analysis of genetic similarities there is no clustering in accord with geographical distances among individuals.This lack of genetic differentiation among fragments can be explained because the species present cross-fertilization and pollen flow (Loveless and Hamrick, 1984) without a pattern of genetic structure.There is a fairly close genetic relationship (63.2%) among the majority of individuals situated in the Propriá fragment, which may be caused by isolation, as this fragment is sited in an island in the São Francisco River, and therefore has a barrier to gene flow.
On-the-other-hand, the explanation for imperfect grouping for each fragment cannot be explained by the species pollination, which is coved by bees moving a relatively short distance, but enough to improve gene flow.Loveless and Hamrick (1984) attest that the species pollinated by insects present low variability, due to the limited movement of the insects.Thus, an explanation for the reduced difference among the fragments could be supported by evolutionary occurrence of such events involving individuals.Populations from different regions, although geographically isolated by habitat fragmentation, contains similar allele frequencies and can be a result of ancient genetic exchanges, that is, until the recent past the populations constituted a metapopulation (Raposo et al., 2007).
The results obtained with the Mantel test prove the lack of association among the individuals when considering the genetic distances and geographical distances.The coefficient of correlation between these distances was negative (r=-0.26426) and significant (0.0028) (Figure 2).This indicates that the fragments are not isolated by distance, so the genetic diversity among them could not be explained by geographic distance (Loss et al., 2006).To evaluate the relationship between genetic diversity and location of individuals Enterelobium contortisiliquum, at Sergipean region of Lower São Francisco, Santana   2008) obtained similar results to this work, verifying that, depending on the species traits (phenology, plant dispersal mechanisms etc.)., the recommended minimum distance between matrices for seed collection does not need to be strictly followed, since not always the genetic diversity between individuals can be explained by the geographical distance between them.The use of molecular markers enables the knowledge of spatial genetic structure of forest species aiming at the samples that are more significant with expressive variability in natural populations.The information is important for breeding purposes as well as for conservation and management, contributing thereby to the sustainability of genetic resources (Lacerda and Kageyama, 2003).The genetic parameters among the studied individuals are presented in Table 3.
The estimative of percentage of polymorphic loci was 76.92%.In contrast, the work of Yu et al. (2011)  individuals.The number of alleles observed was 1.77 and the effective number of alleles was 1.48, which represents the number of alleles that effectively participate of gene flow.The index diversity by Nei, also called heterozygosity was used to assess the polymorphic content of each locus with a value of 0.27.Taking into account the stadium of fragmentation of the studied regions, the heterozygosity could be explained by a low or absent activity of genetic erosion processes in those fragments.Furthermore, the high levels of genetic variability among individuals of aroeira should be supported by self-incompatibility alleles and outcrossing.Considerable levels of genetic variation and the absence of endogamy, due to incompatibility mechanisms, are indicative of long-term maintenance (Raposo et al., 2007).However, compared to the fast pace of exploration, only these types of conditions will not be sufficient to sustain populations of aroeira.Prioritization for pre-breeding actions aiming to establishment clones for commercial production and conservation of genetic variability is necessary with a focus on ex situ banks.Freitas et al. (2006) states that even populations with a high rate of genetic variability, which are not categorized as endangered, is important to conserve genetic resources for plant breeding and use of specific genes of interest.
Most indices of genetic diversity depends on the estimative of heterozygosity in the population, and, considering the dominant character of the ISSR, the estimation of these indices following this method is an approximation (Freire et al., 2007).The Shannon Index is not based on heterozygosity and has been used with dominant markers, such as RAPD and ISSR.The value obtained for this parameter was 0.41 indicating a relatively higher diversity (Botrel et al., 2006).Initially, the most divergent genetic materials are the best choice for starting a breeding program or establishing commercial plantations.However, recommendations should only be made after a careful analysis of their performance for commercial traits (Cruz et al., 2004).The fruits of this species have been extracted without optimal management, and therefore, they present great opportunity for enhanced profit for food industries.Therefore correlations with the organoleptic characteristics of fruit obtained from trees selected using chemical and molecular markers and identification of ideal genotypes should be used in breeding programs.

Chemical characterization
The content and yield of essential oils varied among individuals from 0.66 to 2.43 ml, and levels from 1.32 to 4.86% for individuals BG4 and N3, respectively (Table 4).The time of hydrodistillation (180 min) for fruits allowed obtaining optimal performance.According to Barbosa et al. (2007) the initial 20 min is possible to obtain a ratio of 78% of the total oil present in the fruit, and in the final minutes of 180 min only a small amount (2%) of the total oil is extracted.Barbosa et al. (2007) using individuals naturally from Viçosa-MG, found a similar value (4.65%).
In comparison with our study, the maximum value obtained from fruits was 4.87.For the same species higher values were obtained by Roveda et al. (2010) (8.5%) using individuals in the Garden of Medicinal Plants of UFGD-MT.
Chemical compounds, obtained by extraction of essential oils from aroeira, were used for clustering analysis (Figure 3).The average of chemical similarity among individuals was 42.53%, and the similarities ranged from 13 to 90%.The most similar individuals were P4 and P5, and they both had high values.Low values were obtained for BG3 and P4.
The chemical similarity values of 79.45% was estimated for individuals located in Propriá.Given the observed genetic divergence between these individuals, as well as the relative distinction of this group, the values may be related to unique environmental conditions of Propriá, which, unlike the other fragments, is more characteristic of semiarid regions, and the synthesis of active medicinal, aromatic and spice compounds have been shown to be highly affected by temperature and soil type, as well as additional climatic conditions (Rosal et al., 2011).This direct influence of environmental factors on the composition and yield of essential oils is a challenge for farmers to establish stable and productive genotypes and maintain uniformity required by the chemical and food industries (Yamamoto et al., 2008).
The results obtained with the Mantel test presented a negative correlation (r=-0.43)and high significance (0.00) using chemical and geographical distances (Figure 4).The chemical diversity of individuals could not be explained by geographical/ environmental conditions.Essential oils extraction from aroeira identified 51 chemical compounds.Major compounds of essential oils of aroeira were α-pinene, β-pinene, α-phellandrene, δ-  carene, o-cymene, limonene, β-phellandrene and myrtenol (Table 5).There is a significant difference between the levels of the chemical compounds present in the oils.This chemical variation, with prevalence of different compounds in the essential oils, obtained from the same species collected at different locations (ecotypes), is often explained by the high chemical complexity of essential oils.However, the genetic variability of plants should be considered, in addition to geographical factors (location) and ecological (habitat), which are closely related to the quality of essential oils and is expressed through chemotypes (Roveda et al., 2010).As observed by Singh et al. (1998) using leaves and flowers of the species, and by Barbosa et al. (2007) using oils from leaves and fruits, that both analyses were characterized by the presence of mostly monoterpenes.Silva et al. (2010b) verifying chemical compounds present in the essential oils from fruit of aroeira, determined that α-pinene was the main constituent 0.29 0.47 0.65 0.82 1.00 Coeficiente de Jaccard  (29.39%) of the oils, followed by δ-carene (19.69 %), limonene (18.15%) and α-phellandrene (9.39%), in addition to a large concentration of monoterpenes.Barbosa et al. (2007) chemically analyzed a sample of commercial essential oils of aroeira (FLAVÉX) and obtained the same major compounds found here: αpinene (18.82%), α-phellandrene (23.55%), δ-carene (6.32%), β-phellandrene (16.88%).They also found germacreno-D (11.89%), which in this study was detected at a much lower concentration (0.36%).This chemical similarity between the major compounds is an indication that the oil composition is consistent enough to be feasible for larger scale commercialization, and that breeders should select genetic materials that will reliably express these major compounds of interest.Many samples of essential oils of aroeira present Analyzing the different concentrations of the major compounds in the oils samples analyzed, there was a cluster with six groups, two with individuals with higher concentration of α-pinene, α-phellandrene, o-cymene, limonene and β-phellandrene (BG1 and BG2, N2 and N4).The third group is characterized by a higher tenor of α-pinene, α-phellandrene, δ-carene and limonene (BG3 and BG5).N1 presented the highest concentrations of αpinene, β-pinene, α-phellandrene, δ-carene and limonene.α-pinene and β-pinene clustered the individuals BG4, N5 and P4; and α-pinene, β-pinene, δcarene and myrtenol clustered the individuals N3, P1, P2, P3 and P5 (Figure 5).

Conclusions
There are genetic and chemical differences among individuals of aroeira, even those that are close geographically.The existence of variations in genetic and chemical profiles is useful for selection of potential individuals for commercial use, and therefore for plant breeding.The evidences for genotypes selection could include oil composition, and phenotype traits of fruits  production, and also fruits quality.

Figure 6 .
Figure 6.Principal component analysis of chemical compounds of the essential oils of aroeira (S. terebinthifolius Raddi.).
, detected 70 polymorphic fragments (91.10%) for genetic analysis of Myrcia splendens, a value very close to the values used in this study.At the same locus, 55 polymorphic fragments (82.08%) were used by Manica-Cattani et al. (2009) for analysis of Lippia alba Mill, while for Pogostemon cabli, Wu et al. (2010) obtained 194 polymorphic fragments
R = Purine (A or G) and Y = pirimidine (C or T).

Table 3 .
attested for individuals of Magnolia officinalis is 83.2%.The value is considered high for genetic variability among tree Genetic diversity among individuals of aroeira (S. terebinthifolius Raddi.) using ISSR markers.
cMeans followed by the same letter in columns do not differ by test Scott-Knott (p<0.05).