**4. Genetic diversity of tree species**

#### **4.1 Fundamental concepts**

The genetic variability contained in plant species may occur in distinct levels: 1) species within ecosystems; 2) populations within species; and 3) individuals within populations of a species. The genetic structure of a species can be defined as the distribution of the genetic variability within and between its populations as a direct result of the combination among mutation, migration, selection and genetic drift. Also, many tree species harbor effective mechanisms that allow the dispersion of alleles, enabling them to maintain high levels of genetic variability within their populations (Hamrick et al., 1979; Hamrick, 1983; Loveless and Hamrick, 1987). Studying several natural tropical tree populations, Hamrick (1983), concluded that the genetic variability within natural populations is directly linked to their mating system, pollen dispersion syndrome, seed propagation and also by their effective

water reservoirs, industrial plants, mining and urban areas, especially on the margins of water bodies and mountain areas with highly tilted ground. Even so, one should not underestimate the importance of these fragments. For example, in terms of species richness, occur, even today, in the Atlantic Forest biome, about 20.000 species of vascular plants, of

If on one hand the historical processes of economic development led to a predatory pattern of destruction with the formation of relictual fragments of sizes and levels of isolation and different anthropic impact, on the other, the Brazilian environmental legislation, considered one of the most advanced of the planet, through his last Forest Code, established in 1965 and improved in recent decades, defined areas of permanent preservation and legal reserve (Medeiros et al., 2004). The first has the function of conservation of ecosystem services, encompassing riparian forests, river headwaters, hills tops and mountains, hillsides with slopes greater than 45 °, "restingas" and mangroves. The second has the goal of biodiversity conservation and must have, at least, 20% of the property area in the southern and southeastern Brazil, reaching 80% of the area in the Amazon region. Thus, the historical process that led to the intense fragmentation of the Atlantic Forest biome has suffered, in recent decades, the influence of an environmental legislation more effective for conservation. The result is that most of the remaining fragments can be found in permanent preservation areas and legal reserves, especially in riparian areas and steep slopes with

At this time, Brazilian society is burned with a heated debate, in the Brazilian National Congress, on the proposed changes to the Forest Code. In one side, the sectors of society and policymakers linked to the economically strong Brazilian agricultural sector require the flexibility of code, aiming to reduce the need for recovery and conservation of areas of permanent preservation and legal reserve, in order to result in the release of more land for agriculture. In the other side, the Brazilian Society for Science Progress, urban sectors of society and policymakers linked to the environmental movement strongly oppose the proposed changes. There is a concern that, if these changes are approved, the fragmentation process of natural areas and the destruction of relictual fragments will be intensified, which would result in ecological and evolutionary consequences for populations of many species

The genetic variability contained in plant species may occur in distinct levels: 1) species within ecosystems; 2) populations within species; and 3) individuals within populations of a species. The genetic structure of a species can be defined as the distribution of the genetic variability within and between its populations as a direct result of the combination among mutation, migration, selection and genetic drift. Also, many tree species harbor effective mechanisms that allow the dispersion of alleles, enabling them to maintain high levels of genetic variability within their populations (Hamrick et al., 1979; Hamrick, 1983; Loveless and Hamrick, 1987). Studying several natural tropical tree populations, Hamrick (1983), concluded that the genetic variability within natural populations is directly linked to their mating system, pollen dispersion syndrome, seed propagation and also by their effective

which 6.000 are endemic (Fundação SOS Mata Atlântica and INPE, 2011).

unstable soils.

present in these natural areas.

**4.1 Fundamental concepts** 

**4. Genetic diversity of tree species** 

population size. These factors are also related to the geographic distribution and the type of community in which such species naturally occur.

As stated above, understanding the genetic structure of populations that occur in forest remnants is fundamentally necessary to establish adequate criteria in to which these populations will play a role in the re-composition of degraded ecosystems (Kageyama, 1987). For such, it is necessary not only to understand the genetic diversity of a species, but also, how this variability is distributed within and between their populations present in the disturbed areas. To such intention many statistical tools have since been developed to measure, qualify and partition this genetic variability.

In 1951, Sewal Wright established one of the main components of the distribution of genetic diversity in natural populations: the partition of the endogamy coefficient within and among populations. In his method Wright was able to determine how the endogamy coefficient determine not only the level of crossing between close related individuals within a population but also how it can be related to the differentiation of multiple populations and the overall adaptability of a metapopulation. Also he was able to demonstrate how this partition could be directly linked to the matting system present within each species. His method partitioned the components of the endogamy coefficient f into three distinct coefficients: FIS, which is mainly used to measure the degree of crossing between closely related individuals within a population; FST, which can be considered an estimative of the endogamy level among populations. Although such concept might be strange, Wright was able to determine that this endogamy between populations was equivalent to determining the genetic relatedness among these populations; and FIT, which represent the endogamy level present in the whole metapopulation and correspond to the overall adaptability measured for the sum of populations. We can also consider that FIS represents the endogamy level related to the reproductive system present within a species, FST to be the endogamy level due to the partition of the population into subpopulations, and FIT the endogamy level related to the reproductive system and the subdivision of all subpopulations.

This concept of partitioning the f statistic developed by Wright is highly important when we consider the genetic study of the fragmented population as it gives an overview of how the genetic variability is distributed within and among the subpopulations of a species in a determined area and permits us not only to infer the level of fragmentation within a metapopulation but also main type of reproduction present in a species.

When considering the genetic pattern of natural populations, we need also to know specific patterns of genetic richness within each population to answer a wide range of questions like: degree of conservation, percentage of variation within each specific population, differences in genetic diversity and degree of heterozygosity. This being said, we have to consider other important measures of population genetics.

The percentage of polymorphic loci is one of the simplest measurements of the genetic variation that can be used to evaluate the genetic variability present in a population. As the name says, it shows the number of polymorphic loci present in a population in relation to all the amplified loci obtained with the DNA-based Markers.

Another form to evaluate the genetic variability present within a population is to obtain information about its genetic diversity. Initially geneticist borrowed the method of calculating such diversity from our friends Ecologists. The Shannon and Weaver (1949) index (*H*) was employed in ecology to measure the diversity of species within a given area and later was adapted to measure the genetic diversity within the studied populations. Later on, a more specific index was developed to measure this genetic diversity eliminating some of the bias that was generated when adapting the Shannon index for the evaluation of the genetic data. The Nei´s genetic diversity (1973) was developed as a specific way to measure the population genetic diversity using data obtained specifically with the DNA-based markers. This index was able to measure more accurately the degree of genetic variation within each population and presented different considerations when analyzing the data obtained by dominant (there is no way to differentiate the recessive alleles) codominant (all alleles are differentiated, and each pair of primers is considered to amplify only one molecular locus) markers. For dominant (*H*) data Nei´s genetic diversity is analyzed in terms of within population gene diversity (*Hs*) and the total gene diversity present in the pool of populations studied (*Ht*). But as said above, for codominant marker more detailed information can be given by this statistical index, dividing this statistic into observed (*Ho*) and expected (*He*) heterozygosities, making it possible to calculate the excess or deficit of heterozygotes within each population, gene flow and inferences of genetic bottlenecks and genetic drift.

Considering that in the traditional method of calculation the genetic variability was based on the assumption that the populations were in Hardy-Weinberg equilibrium, some of the inferences obtained for this population presented a significant bias. As a good example we can consider the inference obtained with the Nei´s statistic for genetic diversity, the observed heterozygosity (*Ho*) was compared to the expected heterozygosity (*He*) which was obtained as the pattern of distribution of all the alleles amplified if the population was in HW equilibrium. To eliminate this bias the Baeysian statistic method was adapted to the analysis of molecular data, promoting a revolution in the parameters that can be calculated using data obtained by the DNA-based markers. The Bayesian method is characterized by the use of the posterior probability to infer the likelihood of occurrence of a particular event. In this method all the assumptions, like HW equilibrium, are discarded and the obtained results with this method are compared with a chain (algorithms - Markov Chain of Monte Carlo MCMC and Metropolis Hastings) of results that come from the analyses of the same data. This method of chained analyses repeats itself in tandem until the obtained "chain" stabilizes and yields a result that come close to the real pattern of genetic variation present in natural populations. With this method some parameters like population bottlenecks, attribution of genotypes to specific populations, pattern of gene flow and migration, and attribution of individual paternity, allowing inferences on the population genetics based on DNA-based markers more complete and trustworthy.

#### **4.2 Consequences of habitat fragmentation to the genetic diversity of tree species**

Habitat fragmentation is one of the most important and well diffused consequences to the anthropic soil use dynamics (Brooks et al. 2002). It is characterized by the rupture of landscape unity that initially presented continuity, generating smaller parcels with different dynamics from the original habitat. Such parcels become disconnected from the original biological processes that occurred throughout the area (Dias et al., 2000), behaving like isolated "islands of biodiversity" surrounded by non forest areas (Debinsk and Holt, 2000).

calculating such diversity from our friends Ecologists. The Shannon and Weaver (1949) index (*H*) was employed in ecology to measure the diversity of species within a given area and later was adapted to measure the genetic diversity within the studied populations. Later on, a more specific index was developed to measure this genetic diversity eliminating some of the bias that was generated when adapting the Shannon index for the evaluation of the genetic data. The Nei´s genetic diversity (1973) was developed as a specific way to measure the population genetic diversity using data obtained specifically with the DNA-based markers. This index was able to measure more accurately the degree of genetic variation within each population and presented different considerations when analyzing the data obtained by dominant (there is no way to differentiate the recessive alleles) codominant (all alleles are differentiated, and each pair of primers is considered to amplify only one molecular locus) markers. For dominant (*H*) data Nei´s genetic diversity is analyzed in terms of within population gene diversity (*Hs*) and the total gene diversity present in the pool of populations studied (*Ht*). But as said above, for codominant marker more detailed information can be given by this statistical index, dividing this statistic into observed (*Ho*) and expected (*He*) heterozygosities, making it possible to calculate the excess or deficit of heterozygotes within each population, gene flow and inferences of genetic bottlenecks and

Considering that in the traditional method of calculation the genetic variability was based on the assumption that the populations were in Hardy-Weinberg equilibrium, some of the inferences obtained for this population presented a significant bias. As a good example we can consider the inference obtained with the Nei´s statistic for genetic diversity, the observed heterozygosity (*Ho*) was compared to the expected heterozygosity (*He*) which was obtained as the pattern of distribution of all the alleles amplified if the population was in HW equilibrium. To eliminate this bias the Baeysian statistic method was adapted to the analysis of molecular data, promoting a revolution in the parameters that can be calculated using data obtained by the DNA-based markers. The Bayesian method is characterized by the use of the posterior probability to infer the likelihood of occurrence of a particular event. In this method all the assumptions, like HW equilibrium, are discarded and the obtained results with this method are compared with a chain (algorithms - Markov Chain of Monte Carlo MCMC and Metropolis Hastings) of results that come from the analyses of the same data. This method of chained analyses repeats itself in tandem until the obtained "chain" stabilizes and yields a result that come close to the real pattern of genetic variation present in natural populations. With this method some parameters like population bottlenecks, attribution of genotypes to specific populations, pattern of gene flow and migration, and attribution of individual paternity, allowing inferences on the population genetics based on

**4.2 Consequences of habitat fragmentation to the genetic diversity of tree species** 

Habitat fragmentation is one of the most important and well diffused consequences to the anthropic soil use dynamics (Brooks et al. 2002). It is characterized by the rupture of landscape unity that initially presented continuity, generating smaller parcels with different dynamics from the original habitat. Such parcels become disconnected from the original biological processes that occurred throughout the area (Dias et al., 2000), behaving like isolated "islands of biodiversity" surrounded by non forest areas (Debinsk and Holt, 2000).

DNA-based markers more complete and trustworthy.

genetic drift.

For trees, degradation of primary habitat results from two main processes, fragmentation of forest into patches following clearance, and disturbance of habitat following extraction processes, such as selective logging. Tropical trees are thought to be particularly vulnerable to the effects of habitat degradation due to their demographic and reproductive characteristics (Lowe et al. 2005). Estimations made more than 10 years ago predicted that within fifty years, approximately, twenty-five percent of the vascular plant species would be extinct (Kala, 2000). This loss is still an ongoing process, which is not only linked to the loss of the number of individual plants of a species, but also to the loss of condition of the habitat in which they initially inhabited, as some species cannot persist in small fragments due to alteration in microclimatic conditions and to the intensification of the border effect in small fragments (Lovejoy, 1983).

Tropical trees are predominantly outcrossed, present extensive genetic flow and keep high levels of genetic variability. They frequently experience low density as a consequence of habitat fragmentation, are highly dependent upon animal pollination and present mixed mating systems, so they generally exhibit more genetic diversity among populations than temperate species (Dick et al. 2008).

For neotropical forest species, a reduction in habitat patch size or population density is usually equivalent to a reduction in population size (Lowe et al. 2005), or a genetic bottleneck. The genetic bottleneck leads to a very well known phenomenon in the evolutionary context of habitat disturbance, the genetic drift. Genetic drift can be defined as the sum of random changes in the frequencies of alleles within a population (Futuyma, 2005). This process has as principal outcome the decrease in genetic variability, which can be detected as a drop in the proportion of polymorphic loci and in the number of alleles per locus.

The reduced effective size of populations is also accompanied by the increase in endogamy levels. This can be a result of both the decline in pollen vectors and consequent raise of selfing in species with mixed mating systems, and the increased probability of crossing among relatives, given their close distribution and reduced potential mates. This can lead to a declined heterozigosity, augmented homozigosity and consequent fixation of alleles, independently of their effects over fitness.

The consequences of genetic drift and endogamy are maximized by the isolation of the remaining fragments, resulting in an increased genetic diversity among them. The loss of genetic variability can affect population viability and limit evolutionary opportunities to the populations: they are expected to suffer increased disease and pest susceptibility, loss of incompatibility alleles, fixation of deleterious alleles and decline in fitness (Young et al, 1996).

Lowe and coworkers, based in a computer simulation (Lowe et al. 2005), found that with even relatively low levels of gene flow between remnant populations, loss of diversity can be significantly mitigated by increasing effective population size. Changes in genetic diversity and differentiation following a decrease in population size take a number of generations to become apparent, which is not the case for inbreeding coefficient that increases immediately in the first generation following the occurrence of selfing. Long-lived, historically outcrossing species, such as neotropical trees, are expected to harbour a high genetic load, as deleterious recessive alleles will be masked at multiple heterozygous loci. Also, although deleterious mutations are expected to be purged by selection over time in these species, mildly deleterious alleles can persist despite strong selection.

Some morpho-physiological and life-history traits could confer differences in plants' vulnerability to the effects of fragmentation. For example, species with long generation times will suffer weaker negative effects of fragmentation than the ones with short generation times. The same for species able to reproduce asexually, that will have an extended time between generations.

Many naturally outcrossed tree species frequently present selfing as a clear result of fragmentation, since low density/fragmented populations tend to be more autogamous than high density populations. In the hypothetical scenario of anthropogenic fragmentation ceasing and landscapes remaining as they are today, the effects on genetic diversity of plants will still be much stronger in the future if mating patterns continue shifting towards selfing (Dick et al. 2008, Aguillar et al. 2008).

Aguillar et al. (2008) showed that outcrossing species, such as neotropical trees, suffer greater losses of alleles and polymorphic loci than non-outcrossing species. For selfincompatible species in particular, this may result in the loss of low frequency selfincompatibility alleles (S) leading to mate limitation and further reduction of effective population size.

Animal pollinated outcrossed species are also strongly negatively affected in terms of effective pollination service and seed production by habitat fragmentation. Considering that more than 98% of tropical trees species are animal pollinated (Bawa 1990), these species are exceptionally vulnerable to fragmentation as a consequence of both, ecological and genetic mechanisms. Not only pollination but also seed dispersion in tropical forests is much more dependent in animals than in wind: more than 70% of all tropical tree species are animal dispersed (Howe and Smallwood 1982). These species share multiple agents and are generally understory (Dick et al. 2008).

For the majority of tropical trees, pollen flow transposes enormously the distances of the seeds flow. Pollen dispersion distances, although dependent on small animals in most cases, can be surprisingly as high as 500m and still higher, a few kilometers, in low density/fragmented populations (reviewed in Dick et al. 2008). Even if sufficient pollen reaches an isolated tree to fertilize all potential ovules, a reduction in diversity of the pollen cloud, due to fewer pollen donors, can reduce population fitness by allowing fertilizations from self, related, or maladapted parents. A reduction in seed set predicted from fragmentation may be due to one or a combination of a lack of pollination (e.g. from loss of pollinators), lack of compatible pollination (from increased self-pollination through restricted pollinator movement or a loss of incompatibility alleles owing to reduced population size), or inbreeding (Lowe et al. 2005).

Removal of primary habitat also usually decreases the probability that migrant seeds will find suitable sites for establishment. Under such circumstances, pioneers or invasive weeds will be favoured and will increase in occurrence. If a species' life history profile is characterized by frequent extinction and colonization events, the metapopulation is under threat of extinction if the two forces are not balanced. Even if site colonization does occur, founder bottlenecks can drastically reduce diversity (Lowe et al. 2005 and citations therein).

Also, although deleterious mutations are expected to be purged by selection over time in

Some morpho-physiological and life-history traits could confer differences in plants' vulnerability to the effects of fragmentation. For example, species with long generation times will suffer weaker negative effects of fragmentation than the ones with short generation times. The same for species able to reproduce asexually, that will have an

Many naturally outcrossed tree species frequently present selfing as a clear result of fragmentation, since low density/fragmented populations tend to be more autogamous than high density populations. In the hypothetical scenario of anthropogenic fragmentation ceasing and landscapes remaining as they are today, the effects on genetic diversity of plants will still be much stronger in the future if mating patterns continue shifting towards selfing

Aguillar et al. (2008) showed that outcrossing species, such as neotropical trees, suffer greater losses of alleles and polymorphic loci than non-outcrossing species. For selfincompatible species in particular, this may result in the loss of low frequency selfincompatibility alleles (S) leading to mate limitation and further reduction of effective

Animal pollinated outcrossed species are also strongly negatively affected in terms of effective pollination service and seed production by habitat fragmentation. Considering that more than 98% of tropical trees species are animal pollinated (Bawa 1990), these species are exceptionally vulnerable to fragmentation as a consequence of both, ecological and genetic mechanisms. Not only pollination but also seed dispersion in tropical forests is much more dependent in animals than in wind: more than 70% of all tropical tree species are animal dispersed (Howe and Smallwood 1982). These species share multiple agents and are

For the majority of tropical trees, pollen flow transposes enormously the distances of the seeds flow. Pollen dispersion distances, although dependent on small animals in most cases, can be surprisingly as high as 500m and still higher, a few kilometers, in low density/fragmented populations (reviewed in Dick et al. 2008). Even if sufficient pollen reaches an isolated tree to fertilize all potential ovules, a reduction in diversity of the pollen cloud, due to fewer pollen donors, can reduce population fitness by allowing fertilizations from self, related, or maladapted parents. A reduction in seed set predicted from fragmentation may be due to one or a combination of a lack of pollination (e.g. from loss of pollinators), lack of compatible pollination (from increased self-pollination through restricted pollinator movement or a loss of incompatibility alleles owing to reduced

Removal of primary habitat also usually decreases the probability that migrant seeds will find suitable sites for establishment. Under such circumstances, pioneers or invasive weeds will be favoured and will increase in occurrence. If a species' life history profile is characterized by frequent extinction and colonization events, the metapopulation is under threat of extinction if the two forces are not balanced. Even if site colonization does occur, founder bottlenecks can drastically reduce diversity (Lowe et al. 2005 and citations therein).

these species, mildly deleterious alleles can persist despite strong selection.

extended time between generations.

(Dick et al. 2008, Aguillar et al. 2008).

generally understory (Dick et al. 2008).

population size), or inbreeding (Lowe et al. 2005).

population size.

Species rarity can also determine its susceptibility to genetic erosion. Because common species have comparatively higher levels of genetic variability than naturally rare species, they are expected to lose more diversity due to recent fragmentation processes (Aguillar et al. 2008). Given the ubiquitous nature of anthropogenic habitat fragmentation in today's landscapes, this is important and of interest to conservation biology as they situate common species in potential risk of genetic erosion, which is counterintuitive to current conservation principles that almost exclusively emphasize efforts on rare or threatened species.

As discussed above, habitat fragmentation has the potential to erode genetic diversity of a species, and the magnitude of its effects is related to the state of several life-history traits. Among the several factors, deserve special attention the compatibility system, mating system, pollination vector, seed dispersal vector, vegetative growth capability, rarity, time elapsed in fragmentation conditions, (reviewed in Aguilar et al. 2008), and sucessional stage, as we a going to discuss below.
