Allelic diversity in populations of Solanum lycocarpum A . St .-Hil ( Solanaceae ) in a protected area and a disturbed environment

Allelic diversity in populations of Solanum lycocarpum A. St.-Hil (Solanaceae) in a protected area and a disturbed environment). Th is study aimed to compare the genetic diversity of populations of Solanum lycocarpum A.St.-Hil between natural and human disturbed environments, with the assumption that protected areas have greater genetic diversity than disturbed areas. For this study, two populations were sampled in Goiás State, Brazil. One was located in a conservation unit, Serra de Caldas Novas State Park, in the Caldas Novas municipality. Th e other was located in a pasture area in the municipality of Morrinhos. Th e two populations are 41 km apart. We sampled 60 individuals from each population, which were genotyped with fi ve microsatellite loci (SSR). Th e highest number of alleles was recorded in the population of the conservation unit, where we found 11 exclusive and fi ve rare alleles. In the disturbed area, we recorded only three exclusive alleles and one rare allele. Although we did not observe signifi cant inbreeding in these populations, genetic divergence between them was high (GST (Hedrick)=0.147 =0.147) for a species with long distance seed dispersal. Th e results corroborate the hypothesis that the population in the less disturbed area harbors greater allelic diversity. Th ey also confi rm the eff ectiveness of using protected areas to preserve the genetic diversity of the species.


Introduction
lycocarpum A. St.-Hil. is a typical and widely distributed shrub from Brazilian Cerrado.It is andromonoecious, fl owers and fruits throughout the year (Oliveira Filho & Oliveira 1988) and the fruits have many seeds that easily and rapidly germinate (Vidal et al. 1999).Pollination is mostly by Xylocopa bees (Oliveira Filho & Oliveira 1988) and the dominant seed disperser is the maned wolf (Chrysocyon brachyurus Illiger), which is capable of carrying the seeds over long distances (Courtenay 1994).
Solanum lycocarpum easily occupies disturbed areas, such as roadsides and pastures.Although its seeds are dispersed over long distances, its pollinators are capable only of short distance fl ights.Th us, the species is ideal for studying the genetic structure of populations, which was previously undertaken by Martins et al. (2006) and Moura et al. (2009).However, studies contrasting the population genetic structure in natural and disturbed areas have not been published.Still, the natural habitats of this species have been devastated in recent years because areas have been transformed into extensive monocultures.In order to develop effi cient strategies of genetic conservation of this species, knowing the distribution of the genetic diversity of the remaining populations is crucial, because it permits the selection of populations for both in situ and ex situ conservation.
Although S. lycocarpum occurs in greater density in disturbed areas, this study hypothesized that protected and undisturbed areas harbor greater genetic diversity than disturbed ones.Th us, our goal was to describe the genetic structure of two populations of S. lycocarpum, comparing a disturbed environment with an undisturbed one to verify whether there is an eff ect of human disturbance on the genetic diversity of this species.

Material and methods
We studied two populations of S. lycocarpum in the southern region of Goiás State, in central Brazil.Th e disturbed population was located in an area of pasture (17° 55' S, 49° 00' W) in Morrinhos.Th e other consists of a natural (undisturbed) population, located in Serra de Caldas Novas State Park (Parque Estadual da Serra de Caldas Novas) in Caldas Novas (17° 46' S, 48° 40' W).Th ese populations are 41 km from each other.
Young leaves were collected from 60 individuals of each population.Genomic DNA extraction followed the CTAB procedure described by Ferreira & Grattapaglia (1998).Th e SSR primers were developed for Capsicum spp.(Buso et al. 2000) and transferred to S. lycocarpum (Martins et al., 2006).A detailed description of the analysis of microsatellites can be found in Martins et al. (2006).Five loci were used in this study (Table 1).Genetic diversity was estimated using the soft ware GDA (Lewis & Zaykin 2000).Qui square tests were performed to verify the signifi cance of the diff erence in number of alleles found between the populations.Estimates of genetic divergence between populations were: θ p (GDA soft ware; Lewis & Zaykin, 2000) and R ST (R ST Calc version 2.2; Goodman, 1997); both with 95% confi dence interval obtained by 1000 bootstrap resampling over loci.By comparison purposes, genetic diff erentiation between populations was also estimated using the statistic G ST (Hedrick) according to Hedrick (2005) and using the soft ware GDA_NT (Degen, 2006).Th e historical gene fl ow Nm was estimated indirectly, according to the 'island model' proposed by Crow & Aoki (1984).For results of other estimates of genetic diversity and spatial genetic structure see Moura et al. (2009).

Results and discussion
A higher number of alleles was detected in the undisturbed population of Serra de Caldas Novas State Park than in the disturbed population located in a pasture in Morrinhos.Th e Caldas Novas population had 26 alleles, 11 being exclusive and some having very low frequency (minimum of 0.009).Th e disturbed population (Morrinhos) had 18 alleles, 3 of which were exclusive and rare (frequencies varied from 0.008 to 0.025( Table 1).Th e number of alleles was signifi cantly diff erent between the populations (χ 2 = 6.03; p < 0.05; d.f.= 1).Th is showed that even a species commonly found in disturbed areas, such as pastures and roadsides, have a higher genetic diversity in undisturbed areas than in disturbed ones.Moura et al. (2009), studying the same populations, detected less spatial genetic structure in Serra de Caldas Novas State Park than in the pasture.Th ese authors suggested that this result might have been due to the larger distance of seed dispersal in the park, leading to small levels of spatial genetic structure in this population.
Estimates of genetic divergence among populations were signifi cantly higher than zero (Table 2), confi rming that genetic drift has aff ected the genetic diff erence between the populations.Th e highest divergence was obtained by the unbiased G ST (Hedrick) estimator (G ST (Hedrick) =0.147), showing that 85.3% of the genetic diversity is located within the populations.Th e estimate of historical gene fl ow among the populations was very low (Nm < 1, Table 2).
As previously mentioned, Solanum lycocarpum occupies disturbed environments and is commonly found along roadsides, which possibly act as gene fl ow corridors, connecting populations (Martins et al. 2011).Th us, the signifi cant genetic divergence between populations can demonstrate the eff ect of fragmentation on these populations, counteracting the potential eff ect of the corridors in facilitating gene fl ow.Th ese populations are at a relatively small distance (41 km) from each other, which would not be expected to cause too much divergence, considering the dispersal biology of this species (the maned wolf can walk about 40 km per night).The difference in the estimations of genetic divergence (Table 2) is due to differences in the calculation methods.G ST (Hedrick) , estimated according to Hedrick (2005), was slightly higher than θ p as estimated by Weir & Cockerham's method (1984). Hedrick's method (2005) considers the type of alleles found in the populations and the allele frequencies, while Weir & Cockerham's method (1984) considers only the variation in allele frequencies.
No signifi cant inbreeding was detected in these populations (Moura et al. 2009), indicating that, although genetic drift may have caused genetic diff erentiation between the populations, probably a random mating system has occurred within populations.Despites the absence of inbreeding, the record of less than one migrant per generation between populations and loss of alleles in the disturbed environment can jeopardize the conservation of this species in the long run.Our results show the eff ectiveness of protected areas for in situ conservation of genetic diversity.Further studies are necessary regarding the genetic structure of plants of the Cerrado for implementation of eff ective conservation management plans.
In general, our results suggest that both populations must be conserved in situ.In addition, the pasture population must also be conserved ex situ, because of the occurrence of some exclusive alleles and its location in a very anthropogenic landscape.

Table 1 .
Allele frequencies in fi ve nuclear microsatellite loci, in two populations of Solanum lycocarpum in Goiás state.n is the sample size and the numbers in the second column are alleles, named by their length in bp (base pairs).