Landscape-Level and Fine-Scale Genetic Structure of the Neotropical Tree Protium spruceanum ( Burseraceae )

1 Unidade Acadêmica Especializada em Ciências Agrárias, Universidade Federal do Rio Grande do Norte, 59280-000 Macaı́ba, RN, Brazil 2 Departamento de Oceanografia e Limnologia, Universidade Federal do Rio Grande do Norte, 59014-100 Natal, RN, Brazil 3 Departamento de Engenharia Florestal, Universidade de Braśılia, 70919-900 Braśılia, DF, Brazil 4 Departamento de Ciências Florestais, Universidade Federal de Lavras, 37200-000 Lavras, MG, Brazil


Introduction
Forest loss and spatial isolation of natural populations affect negatively the reproductive success of many tropical plants [1,2].In contrast, some studies have shown that the habitat fragmentation had resulted in facilitated pollen movement [3,4], whilst gene flow measured among populations of several insect-pollinated woody plant species [5,6] and in a wind-pollinated tree species [7] was not reduced, facilitating too the long distance dispersal and high genetic diversity of remnant populations [8].However, when comparing continuous and naturally fragmented populations in a tropical tree species, fragmented populations exhibited slightly lower genetic diversity [9].Landscape structures, such as ecological corridors, are then an important alternative in the demographic and genetic connection of isolated forest fragments, thus minimizing the negative effects of habitat fragmentation [10].Nonetheless, most studies deal only with the use of corridor's importance for fauna conservation [11]; little is known about corridor importance for plant species mainly on genetic processes [12].
The genetic differentiation in fragment/corridor may be quantified with indirect methods which are based on the structure of neutral genetic diversity amongst populations, contributing to the knowledge of how landscape characteristics structure populations [13].Moreover, studies of spatial genetic structure (SGS) provide insight into microevolutionary patterns by elucidating the movement of genes at a range of spatial scales [14].At landscape-level scales between natural populations, genetic structure has been attributed to historical factors and isolation by distance [15].At fine-scale SGS, seed dispersal and demographic processes acting at specific temporal and spatial scales probably have the greatest influence on the presence or absence of significant fine-scale genetic structuring [16].
The forest remnants in southern Minas Gerais state are characterized by a hilly topography covered by Atlantic seasonal forests.These plant formations were seriously plundered since European occupation goes back to colonial times ca.two centuries ago, resulting in the fragmentation and isolation of these populations at a particularly rapid pace.At the same time, ditches to divide rural properties (∼6 m-wide) were constructed by slaves, resulting in the ecological corridors (colonization by native tree species) that connect small fragments [17].Protium spruceanum is a large canopy tree (up to 25 m tall) found in those fragment and ecological corridors, selected for this research as representative of a mass-flowering insect-pollinated and birddispersed tropical tree.The small pale yellowish flowers are functionally unisexual and organized in dense inflorescences, the individuals are dioecious.The effective pollinators are Apis mellifera and Trigona sp.[17].
We examined the neutral genetic differentiation between populations that can help define important strategies for conservation genetics of remaining populations and ecological corridors.We therefore asked the following questions: (1) what are the levels and spatial scales of relatedness in P. spruceanum across landscape-level and within population?(2) is there any correlation between the fragment-corridor connection and genetic differentiation?And if so, are genotypes randomly or distributed clusters in corridors?(3) what are the potential ecological mechanisms likely responsible for the observed fine-scale SGS?

Materials and Methods
2.1.Sampled Sites.The fragment-corridor system studied is located in the city of Lavras, South of Minas Gerais state, in Brazil (Figure 1).The populations studied have rapidly declined because of habitat fragmentation caused by anthropogenic disturbance over the last 200 years, corroborated by detection of recent bottlenecks [17].Five interlinking fragments and an ecological corridor were analyzed.Protium spruceanum presence in fragments F2, F3, and F4 coincides with the presence of water courses.In fragments F1 and F5 the species occur in a large part of the fragment, which may be determined by the almost permanently flooded soil(Figure 4).All samples along the fragments came from trees with diameter at breast height (d.b.h.) >20 cm and ∼ 16 m height, that is, representing a larger diameter class.The sampling in each corridor axis was along the length of each corridor axis.Individuals were mapped by their compass direction and distance from the last individual was sampled.Live tissue was transported on ice and stored at −80 • C until enzyme extraction.

Enzyme Extraction and Electrophoresis.
Enzymes were extracted from 200 mg of frozen leaf tissue in 1 mL of phosphate extraction buffer: 0.2 mM sucrose, 2.5% polyvinylpyrrolidone (PVP), 1 mM ethylenediaminetetraacetic acid (EDTA), 5.7 mM ascorbic acid, 5.8 mM sodium diethyl carbamate (DIECA), 2.6 mM sodium bisulphite, 2.5 mM Borax, and 0.2% β-mercaptoethanol [18].Discontinuous vertical electrophoresis in a polyacrylamide gel was performed using 10% gels and carried out at 4  [19].Allozymes had banding patterns consistent with expected quaternary structures and putative loci, and alleles were designated sequentially.The one with the most anodally migrating allozyme or alleles was designated as 1 and the next as 2 (Figure 5).The Mdh and Per enzyme patterns showed two polymorphic loci; so we used ten polymorphic loci to genotype the individuals.

Genetic Differentiation Analysis at Landscape Level.
A Gtest, based on 3 600 permutations of genotypes among samples, was performed to test for population differentiation F ST at allozymes loci [20].We adjusted the significance level for multiple pairwise comparisons using a sequential Bonferroni correction.The overall F ST was calculated based on allele frequencies of the population of each at allozyme loci [21], using the program FSTAT 2.9.3.2 [22].The degree of mean F ST values of each fragment in comparison with those of the remaining fragments was estimated (Figure 2).To test for isolation by distance (rM Mantel tests), pairwise F ST /1 − F ST matrices [23] were related to geographical distances between fragments, using 1 000 permutations.These calculations were performed using the program FSTAT.Additionally, the Pearson correlation coefficient (r) was used to test if gene diversity (H E ) varies according to fragment size.

Fine-Scale SGS Analysis.
Population structure was further analyzed using Nason's kinship coefficients (F i j ) between pairs of individuals i and j [24].This coefficient can estimate, between pairs of mapped individuals x and y, a ratio of differences of probabilities of identity-in-state between homologous genes [25].Distance class intervals between individuals were determined by testing to get rule of thumb for each distance interval as suggested by Hardy and Vekemans [26].To test for significant deviations from random SGS, observed values for each distance class were compared to the 95% confidence interval derived from 1 000 permutations.The extent of SGS was estimated using the Sp statistic following Vekemans and Hardy [14].The Sp statistic here is used as a simple measure to allow for comparisons among fragments, not as an estimate of the variance in gene dispersal distances.Sp was quantified by , where b log is the regression slope and F (20,m) is the mean kinship coefficient between individuals belonging to the first distance interval (0-20 m).The b log standard errors were obtained by jack-knifing over loci.These calculations were performed using the program SPAGeDi 1.2 g [26].

Genetic Differentiation and Spatial Genetic Structure.
Wright's F ST revealed little evidence of significant genetic structure between fragments.The average genetic differentiation (F ST ) in each forest fragment pair was generally low (F ST = 0.027).Only the genetic differentiation between fragment F3 and F5 was significant (F ST = 0.111, P < .01).Indeed, the mean F ST results showed the greater genetically distinct fragments, F3 and F5 (Figure 2).At landscape-level scales, the Mantel test suggests that the populations'structure appears not to be due to differences in gene flow produced by an isolation by distance model (rM = −0.051,P = .539).The degree of mean F ST values was also calculated considering both fragments and corridors, indicating the decrease of the level of divergence (Figure 2).All fragments have generally lower genetic differentiation with corridors to which they are connected (Figure 1).The highest and most significant genetic differentiation occurred between fragment F5 and F2-F3 (F ST = 10.5%,P < .05)and F3-F4 (F ST = 14.8%,P < .01)axis.Negative correlation was observed between gene diversity and fragment size (Pearson r = −0.626,R 2 = 0.392, d.f.= 4).

Fine Scale SGS.
At fine-scale, in all fragments, coancestry values were within the range of 95% confidence limits in all of the distance classes (Figure 3).Random genotype distribution was observed also when estimating the coancestry for individuals within ecological corridors, in all distance classes.Coancestry values were close to zero or negative, and nonsignificant at 95% probability.Sp values revealed weak patterns SGS (average Sp = 0.008) and the slope of the regression of F i j was not significant.fragments and in corridors (Table 1).In fact, forest trees are generally highly genetically diverse and characterized by high levels of gene diversity within populations and comparatively low gene diversity between them [27].In addition, the highlights importance of small-sized fragments was demonstrated by negative correlation between fragment size and gene diversity.Some studies have shown a tendency of lower genetic diversity within small fragments [28], but little or no genetic diversity difference has been found between populations of large and small fragments [29].However, our findings show that small-sized fragments may be important units within fragmented landscape and may play a fundamental role to connect larger fragments and contribute to gene flow between populations.

4.2.
Landscape-Level and Fine-Scale Genetic Structure.At both spatial scales, the average relatedness estimates indicated essentially random spatial distributions of alleles.At the landscape-scale (fragments), local populations of P. spruceanum exhibit little variation in their internal genetic structures.This suggests that the causal mechanisms generating genetic structure (e.g., seed dispersal, recruitment processes, etc.) are relatively homogeneous at this scale.
Considering that for an insect-pollinated, bird dispersed  tree, connectivity by ecological corridors, low genetic differentiation among fragments was actually expected and observed (F ST = 0.027).Results obtained for genetic differentiation between fragments and ecological corridors corroborate this hypothesis, because all fragments have generally lower genetic differentiation with corridors which they are connected to.The level of divergence (mean F ST ) for each fragment against the rest of the fragments showed the greater genetically distinct fragments (F3 and F5) and could be included in a high-priority conservation programme.At fine scale, many factors interact to determine genetic structure within populations including mating system and  colonization history [14,30].The failure to detect genetic structure within populations may be due to the adult density (average = 232 individuals • ha −1 ), intraspecific competition or random mortality among samplings, resulting in extensive thinning within maternal half-sib groups.To determine whether this form of thinning occurred during recruitment in the study populations, investigations on contemporary patterns of genetic structure within populations and longterm demographic data are necessary to provide a clear picture of the genetic structure across demographic stages.In fact, our findings show that the magnitude of fine-scale SGS in P. spruceanum was observed to be decreasing from smaller-to larger-diameter classes (F.A. Vieira, unpublished data, 2010), within a small fragment (1.0 ha, fragment F1).Thus, we found SGS in the young plants, but this structure almost completely disappeared in the old plants, due to random thinning of individuals (F.A. Vieira, unpublished data, 2010).In general, other studies have also found withinpopulation genetic structure to be greater in juveniles than adult trees [31,32].
The Sp statistic is robust to different sampling schemes, being useful here to compare among species and sites [14].The fine-scale SGS in P. spruceanum (average Sp = 0.008) was similar to that observed for trees (Sp = 0.010 ± 0.0096), outcrossing species (Sp = 0.012 ± 0.0101) and animaldispersed (Sp = 0.008±0.0050)species [14].Spatial patterns of genetic variation within populations depend primarily on the seed dispersal and adult densities [16].This way, the high density of seed sources and random thinning of individuals should favor the random divergence of the allelic frequency of Protium spruceanum in the fragments and corridors.

Implications for Conservation Genetics.
Here, we studied a mass-flowering insect-pollinated and bird-dispersed tropical tree, as representative of a tree species with similar life history characteristics that occur in the fragments and corridor systems.Our findings show higher gene diversity and a low differentiation between the studied forest fragments.Moreover, those ecological corridors presented high gene diversity, absence of inbreeding, and little genetic differentiation with the fragments.Thus, the creation and the protection of ecological corridors should be an effective conservation strategy for tree species.That would be an important alternative in the demographic and genetic connection of isolated forest fragments, thus minimizing the negative effects of habitat fragmentation.Some studies have provided convincing evidence that, in some cases, corridors can enhance migration rates among fragments [10,33].In addition, the functional aspect of such an approach relates to the biological response of the species and landscape structure, thus new approaches of other species with different life history characteristics would be very important aspects to be considered.

Figure 1 :
Figure1: Location of the study system with forest fragments and ecological corridors in Minas Gerais state, Brazil.F1 to F5 (fragments) and Axis F1-F2 to F4-F5 (ecological corridors).Distances between fragments varied from 0.8 to 2 km.The circles represent the contribution to total genetic differentiation of each fragment with the corridors, using allozymes.

Figure 2 :
Figure 2: Degree of mean F ST values of each fragment in comparison with those of the remaining fragments.Solid and open bars represent degree of mean F ST values for fragments group without ecological corridors and considering both fragments and corridors, respectively.

4. 1 .Figure 3 :
Figure 3: Correlograms of coefficient of coancestry (F ij ) per distance classes in fragments (F1 to F5) and ecological corridors.Confidence intervals around each F ij -value were obtained through a jackknife procedure over loci.

Figure 4 :
Figure 4: Distribution of the trees sampled in five fragments for the analysis of the fine-scale SGS in Protium spruceanum.

Figure 5 :
Figure 5: Zymograms of Protium spruceanum for the enzyme systems peroxidase (Per) and glutamate dehydrogenase (Gtdh) with schematic illustrations and genotype designation.

Table 1 :
[17]ments and corridors code sampled in this study.: sample size; H E , Nei's gene diversity; SD: standard deviations; f : mean fixation index.Detailed descriptions of genetic diversity parameters (H E and f ) may be found in de Almeida Vieira and de Carvalho[17]. N