**4. Discussion**

#### **4.1 Genetic diversity**

Many studies of plants confirm the influence of plants' mating mode on the population genetic structure. Populations of clonal plants are expected to be dominated by one or a few locally adapted genets as distinct from plant population with predominant sexual propagation (Jogaitė et al. 2005). According to Murawski & Hamrick (1990), the mean

Number of polymorphics fragments

Polymorphism (%)

Number of fragments

808 (AG)8C 18 17 94 810 (GA)8T 17 15 90 813 (CT)8T 12 10 88 834 (AG)8YT 12 10 88 842 (GA)8YG 10 10 100 845 (CT)8RG 9 9 100 848 (CA)8RG 7 6 88 890 VHV(GT)7 8 7 88 **Total 93 84 90,3** 

The AMOVA analysis, based on 60 ISSR polymorphic loci, showed that 9.1% of the genetic variability was among population and 90.9% within population and the Φst value was, therefore, equal to 0.0903. The Fst (0.040) value detected for the data based on the SSR markers also suggested no differentiation among populations. The estimated gene flow was 2.5 for the analyses with ISSR markers and 5.74 for

Table 2. ISSR primers used and their respective base sequences, number of amplified

Fig. 4. ISSR marker profiles of 10 samples of *A. fulgens* generated by primer UCB 842 in 2.0%

Many studies of plants confirm the influence of plants' mating mode on the population genetic structure. Populations of clonal plants are expected to be dominated by one or a few locally adapted genets as distinct from plant population with predominant sexual propagation (Jogaitė et al. 2005). According to Murawski & Hamrick (1990), the mean

fragments and polymorphism produced in samples of *Aechmea fulgens.*

agarose gel. M - 1 kb Plus DNA ladder.

**506 bp**

**1, 363 bp**

**2, 036 bp**

**M**

**4. Discussion** 

**4.1 Genetic diversity**

Primers UBC

SSR loci.

Sequence (5' 3')

proportion of polymorphic loci for monocotyledons is 40.3%. These authors examined 18 isoenzyme loci in *Aechmea fulgens* populations and detected 21.1% mean polymorphism. The ISSR markers used in this study therefore generated much higher values than those reported by Murawski & Hamrick (1990) for monocotyledons. The same primers used in this study generated similar results in *Grevillea* (Proteaceae) (Pharmawati *et al*., 2004) and *Cycas guizhouensis* (Cycadaceae) (Xiao et al. 2004). In these studies the number of fragments ranged from 8-23 (UCB 808), 7-10 (UCB 842) and 8-24 (UCB 890). Pharmawati et al. (2004) obtained fragments for the oligonucletides UCB 808 (359-1558 bp) and UCB 890 (369-1385 bp), with sizes close to those detected in this study, that ranged from 350-2000 bp. The percentage of polymorphism (90.3%) was close to that reported by Cavallari et al. (2006), 89.8%, who studied populations of *Ecolirium biflorum* (Bromeliaceae) with RAPD markers.

Microsatellite molecular markers, widely used in plants, are considered to be very reliable because they are highly specific, usually producing a single fragment, frequently polymorphic at the infrageneric level (Oliveira et al. 2006). However, the number of fragments detected depends on the size of the genome, the frequency of the sequences repeated in the genome and the detection method (Barth et al. 2002). In the present study, the primers used to amplify the microsatellite loci (Table 1) generated one to five fragments per loci, depending on the primers. The allele pattern produced by the five pairs of primers used in this study permitted the characterization of the *Aechmea fulgens* species with microsatellites developed for use in studies of other species belonging to different genera.

According to Zane et al. (2003), microsatellites can be transferred among close species, or even among close genera. The transferability of these markers among related species is a consequence of the homology of the regions flanked by the microsatellites and the size of the region between the pair of primers responsible for amplification by PCR. The possibility of using microsatellites developed for one species in the genetic assessment of another is very useful because it reduces costs considerably. Boneh et al. (2003) suggested that the microsatellites they developed for the *Tillandsia* and *Guzmania* genera could be applied to other species within the genus and even to close genera. According to the phylogenetic genetic analysis of the bromeliad family reported by Terry et al. (1997), the *Aechmea fulgens* genus is a very different from those referred to by Boneh et al. (2003) but two primers used by the these authors were useful in the characterization of *Aechmea fulgens*.

The percentage of polymorphic loci (*P*) detected in this study was high compared with tropical tree species, for the three populations studied. Auler et al. (2002) reported a mean of 43% of polymorphic loci and 1.8 alleles per loci in *Araucaria angustifolia* populations with isoenzymatic markers, and considered the values to be high because these populations had a good conservation `status´ compared to other more degraded populations (20% and 1.4).

The total molecular genetic diversity for each population can be observed both by the percentage of polymorphic loci (*P*) or by diversity indexes of Shannon-Wiener (*I*) and Nei's genetic diversity (*h*).

The values detected for the Shannon-Wiener diversity index in this study were, respectively, for the SSR and ISSR markers: Igarassu: *I* = 0.55 and 0.53; São L. da Mata: *I* = 0.51 and 0.51; São V. Ferrer: *I* = 0.51 and 0.56. The Nei genetic diversity ranged from 0.35 to 0.39 in the three populations and for both markers used. Measures of allelic richness are suitable for assessing short-term diversity loss, but for longer-term loss, involving commoner alleles, measures of gene diversity are more appropriate (Lowe et al. 2004).

#### **4.2 Genetic distances**

The Nei genetic distances, calculated for the pairs between populations, ranged from *approximately* 2% to 7% for the SSR and 4% to 7% for the ISSR markers (Tables 3 e 4). The dendrograms among populations distances for both the molecular markers formed two distinct groups: populations 1 and 3 and population 2 (Figure 5 A and B), where the geographic closer populations were in different groups, suggesting absence of association between geographic distance and genetic distance. This result was confirmed by the Mantel test that showed absence of correlation between the Nei genetic distance matrix and geographic distance for the two markers used. It is important point out that the two types of molecular markers gave the same grouping in spite of their different natures.

As the Nei distance value was proportional to the time since divergence and to the genetic substitution rate by loci and generation (Dias 1998) the values of the genetic distances detected in this study and the grouping formed indicated that the populations did not diverge a long time ago suggesting an explanation for the high within population values and lower among population values.

#### **4.3 Genetic structure**

The total genetic diversity revealed that the greatest part of the variation was present within the populations. The results of high within population diversity were in line with several studies with natural plant populations (Auler et al., 2002; Zucchi et al., 2003; Souza et al., 2004; Xiao et al., 2004; Galeuchet et al., 2005; Cavallari et al., 2006). Galeuchet et al. (2005) used seven microsatellite loci and detected Fst = 0.02 in *Lychnis flos-cuculi* L. (Caryophyllaceae). The authors suggested that the low differentiation among populations studied was primarily due to the significant level of historic gene flow (*Nm*) when the populations were more connected and second by the contemporary gene flow and accidental anthropogenic seed dispersion.

The *Nm* value determines whether the gene flow alone can produce substantial genetic variability among locations. If *Nm* is greater than 1.0, the gene flow will be high enough to prevent differentiation due to drift (Moraes & Derbyshire, 2002). In the present study, the estimated gene flow among the populations was high (2.5 and 5.7 for the ISSR and SSR markers, respectively) that together with the high within population genetic variation suggested low impact from the habitat fragmentation on the genetic diversity of *Aechmea fulgens* in the three environments studied. The *Nm* values, according to Galeuchet et al. (2005), should reflect the historic gene flow, because although bromeliads are probably pollinated by birds and/or bats that have a great flight range, most present clonal propagation. Furthermore, the genetic distance dendrograms (Figure 5 A and B) indicated no divergence among the populations studied.

Foré et al. (1992) studied the genetic structure after fragmentation in 15 *Acer saccharum* (Aceraceae) populations and reported the isolation of remnants with levels of genetic diversity. The authors observed no divergence among the remnants (Fst = 0.03) and that the gene flow was greater after fragmentation, in this case, due to the greater incidence of wind

assessing short-term diversity loss, but for longer-term loss, involving commoner alleles,

The Nei genetic distances, calculated for the pairs between populations, ranged from *approximately* 2% to 7% for the SSR and 4% to 7% for the ISSR markers (Tables 3 e 4). The dendrograms among populations distances for both the molecular markers formed two distinct groups: populations 1 and 3 and population 2 (Figure 5 A and B), where the geographic closer populations were in different groups, suggesting absence of association between geographic distance and genetic distance. This result was confirmed by the Mantel test that showed absence of correlation between the Nei genetic distance matrix and geographic distance for the two markers used. It is important point out that the two types

As the Nei distance value was proportional to the time since divergence and to the genetic substitution rate by loci and generation (Dias 1998) the values of the genetic distances detected in this study and the grouping formed indicated that the populations did not diverge a long time ago suggesting an explanation for the high within population values

The total genetic diversity revealed that the greatest part of the variation was present within the populations. The results of high within population diversity were in line with several studies with natural plant populations (Auler et al., 2002; Zucchi et al., 2003; Souza et al., 2004; Xiao et al., 2004; Galeuchet et al., 2005; Cavallari et al., 2006). Galeuchet et al. (2005) used seven microsatellite loci and detected Fst = 0.02 in *Lychnis flos-cuculi* L. (Caryophyllaceae). The authors suggested that the low differentiation among populations studied was primarily due to the significant level of historic gene flow (*Nm*) when the populations were more connected and second by the contemporary gene flow and

The *Nm* value determines whether the gene flow alone can produce substantial genetic variability among locations. If *Nm* is greater than 1.0, the gene flow will be high enough to prevent differentiation due to drift (Moraes & Derbyshire, 2002). In the present study, the estimated gene flow among the populations was high (2.5 and 5.7 for the ISSR and SSR markers, respectively) that together with the high within population genetic variation suggested low impact from the habitat fragmentation on the genetic diversity of *Aechmea fulgens* in the three environments studied. The *Nm* values, according to Galeuchet et al. (2005), should reflect the historic gene flow, because although bromeliads are probably pollinated by birds and/or bats that have a great flight range, most present clonal propagation. Furthermore, the genetic distance dendrograms (Figure 5 A and B) indicated

Foré et al. (1992) studied the genetic structure after fragmentation in 15 *Acer saccharum* (Aceraceae) populations and reported the isolation of remnants with levels of genetic diversity. The authors observed no divergence among the remnants (Fst = 0.03) and that the gene flow was greater after fragmentation, in this case, due to the greater incidence of wind

of molecular markers gave the same grouping in spite of their different natures.

measures of gene diversity are more appropriate (Lowe et al. 2004).

**4.2 Genetic distances** 

and lower among population values.

accidental anthropogenic seed dispersion.

no divergence among the populations studied.

**4.3 Genetic structure** 

and that fragmentation did not result in the isolation of the populations. Souza et al. (2004), however, disagreed with Foré et al. (1992) because they observed high genetic divergence within fragmented populations of *Chorisia speciosa* St. Hil. (Bombacaceae) that indicated, according to the authors, that fragmentation led to genetic drift and to the increase in within population divergence. On the other hand, Nei (1978) stated that when there is reduction in the population size, even though the number of alleles is reduced, the degree of heterozygosity and the genetic diversity can remain as high as in the original population. This is due in most cases to the fact that the alleles lost by drift are rare and contribute little to the level of heterozygosity.


Table 3. Nei's genetic distance matrix ( higher diagonal) and geographical distance (lower diagonal) between pairs of populations of *A. fulgens*. SSR marker data.


Table 4. Nei's genetic distance matrix (higher diagonal) and geographical distance (lower diagonal) between pairs of populations of *A. fulgens*. ISSR marker data.

The genetic structure of *A. fulgens* populations indicates that in situ conservation strategies for this species should be based on the establishment of several small-scale protected areas chosen so as to represent the species' genetic variability. A similar *in situ* conservation schedule was previously suggested by Laguna et al. (1998) for endemic, rare or threatened plant species. Moreover, considering that the samples collected in this study provide a snapshot of the species' distribution area as a whole in Northeastern Brazil and that the genetic variability detected is low, *ex situ* conservation should also be considered in conservation strategies d if *A. fulgens* germplasm represents its overall genetic variability.
