Study Area

The study area includes only the north region of the current distribution of Dicksonia navarrensis in America, the cloud forest of the Sierra Madre Oriental (SMOR), Mexico. This type of vegetation grows in conditions of high atmospheric humidity, frequent fog, and low light. Most plant species retain their foliage throughout the year, with some exceptions during the dry season (Gual-Díaz & Rendón-Correa, 2014; Rzedowski, 2006). The elevation distribution of D. navarrensis ranges between 900 and 2200 m (Figure 1), in sites with slopes of 15°–80°. In the sampling sites Agua Blanca, Medio Monte, Tizapan, Tlanchinol (Hidalgo), and Zilacatipan (Veracruz), the Cwa climate predominates (temperate, with dry winter and hot summer); in Naupan (Puebla) the Cwb climate (temperate, with dry winter and hot summer) and in Xocoyolo (Puebla) the Cfb climate (temperate, without a dry season and with a hot summer; Peel et al., 2007). The average annual temperature ranges between 12 and 23°C, and the total annual rainfall between 1000 and 3000 mm. During the year there is a season of low rainfall that lasts from 0 to 4 months.

Figure 1.

Geographic location of the analyzed wild populations of Dicksonia navarrensis in the Sierra Madre Oriental, Mexico

Sampling

Seven wild populations of D. navarrensis distributed in the Mexican Sierra Madre Oriental were sampled. These populations were relatively close to human settlements, in an elevation interval of 1300 to 2200 m, on steep slopes of up to 70° with northeastern and northwestern orientation, and with canopy covers of 50–78% (Reyes-Ortiz et al., 2019). The distance between populations ranged from 12 to 163 km (Figure 1). Fresh foliar tissue was collected from 30 individuals randomly selected from each population (a total of 210 samples). The samples were labeled, cleaned with 70% ethanol, stored individually in plastic bags filled with silica gel, and preserved at −20°C until they were used for subsequent DNA extraction at the Biological Research Center of the Autonomous University of Hidalgo State, Mexico.

DNA Extraction

A modified cetyltrimethyl ammonium bromide (CTAB) protocol (Doyle & Doyle, 1987) was used to isolate genomic DNA from leaf tissue. DNA concentration was measured in an Eppendorf^TM Bio Spectrometer basic. The quality and integrity of the molecule were assessed with 0.8% agarose gel electrophoresis.

Amplification

Eleven specific SSR markers (Dic01–Dic04, Dic06, Dic08, Dic10, Dic11–Dic14), developed for D. sellowiana, were used for the polymerase chain reaction (PCR); of which, eight were polymorphic for the populations of the south and southeast of Brazil ( Supplementary Material 1; Nazareno et al., 2013). PCRs were carried out in a Thermo Scientific™ Arktik™ Thermal Cycler with a final volume of 8 µL containing 50 ng of genomic DNA, buffer Taq 1X, MgCl₂ (2.5 mM), dNTPs Mix (0.6 mM), forward primer (25 p.m.), reverse primer (25p.m.), Taq DNA (Promega^TM) polymerase (0.8 U) and sterile H₂O. PCR conditions were as follows: 94°C for 3 min of initial denaturation, 38 cycles of 94°C for 30 s, 45 s of annealing at 49–63°C for 45 s ( Supplement Material 1), 1 min of elongation at 72°C, and a final extension step of 72°C for 10 min. Amplified fragments were separated into 15% polyacrylamide gels (140 V, 90 min), stained with ethidium bromide, and visualized under UV light on a transilluminator (2 UVTM-UVP^TM) with a molecular ladder of 100 bp (Promega^TM).

Data Analysis

Genetic Diversity

The results revealed the presence of two monomorphic loci (Dic04 and Dic06), which were excluded from the subsequent genetic analyses (mean of polymorphic loci 74.03 ± 3.4%, across the 11 screened markers). A total of 33 alleles were identified at the remaining nine loci, ranging from 2 (Dic13) to 6 (Dic01 and Dic08), with an average of 2.55 alleles per locus ( Supplementary Material 1). The mean heterozygosity observed (H_O) and expected (H_E) per population were 0.133 and 0.367, respectively. Thus, D. navarrensis populations showed an inbreeding tendency, probably due to heterozygosity loss (H_O = 0.133 ± 0.023, f = 0.65 ± 0.045). Low values of effective alleles by locus (mean A_e = 1.766 ± 0.079) were estimated. Genetic diversity differed between the analyzed populations (Table 1). Three exclusive alleles were found in POP1 (Dic03: 195 bp, Dic04: 185 and 219 bp), three in POP4 (Dic02, Dic03, and Dic08: 253, 249, and 283 bp, respectively), and only one in POP6 (Dic01: 295 bp).

Table 1.

Genetic Diversity Among Populations of Dicksonia navarrensis in Mexico.

Genetic Structure

The Nm and F_ST values were heterogeneous between the populations analyzed. The values of the index of differentiation ranged between 0.045 and 0.328 (mean F_{ST(6, 419)} = 0.174, p < 0.001). The highest differentiation values were those of population 1, with respect to populations 4–7. The AMOVA indicated greater variation among individuals (58%) than within individuals (24%) or populations (18%). Inter-population flow levels ranged from 0.511 to 5.285, and the average number of migrant individuals per generation (Nm) between populations was 1.19 (Table 2).

Table 2.

Genetic Structure Calculated by AMOVA. The F_ST and N_m Values Appear Below and Above the Dashes, Respectively.

The structure of the population inferred with the assignation methods (Evanno et al., 2005; Puechmaille, 2016) revealed that K = 3 and K = 4 were the numbers of most likely clusters for an admixture model within the predefined populations. The highest homogeneity among individuals was recognized in POP1 and POP6, and major heterogeneity was found in POP2 and POP3. POP1, located in the north, was a divergent population in both K models (Figure 2). The first assignment group (K = 3) includes the three northern populations, the second cluster comprises POP4 and POP7, and POP5 and POP6 form the remaining group. In the K = 4 model, POP1 is isolated from the two remaining north cluster populations. Both assignments are similar to the genetic associations identified by neighbor-joining clustering (Figure 3).

Figure 2.

Assignation analysis based on Bayesian model of STRUCTURE software: (a) K = 3, (b) K = 4. Colors represent the groups identified; each population is represented by one chart, showing the proportion of membership of the different clusters.

Figure 3.

Neighbor-joining clustering in wild populations of Dicksonia navarrensis in Mexico.

The DACP results indicate that there are significant differences between all populations, consistent with the AMOVA results, but only 4 loci were needed to explain 65.3% of the variance (F_{(17, 187)} = 8.59 ± 4.23; p < 0.0002) correlated with two factors. In the ordination diagram, it is observed that the individuals of POP1 and POP4 are more differentiated, compared to the rest of the populations (Figure 4). The correlation between the geographic distances (km) and the Nei genetic distances of the D. navarrensis populations was statistically significant (r = 0.58, p = 0.03) ( Supplementary Material 2). Geographic distance between the populations ranged from 12.7 to 163.5 km.

Figure 4.

DAPC grouping of wild populations of Dicksonia navarrensis in Mexico based on linear distances orthogonalized by correlation matrix.

Genetic Diversity

This study provides the first results on the structure and genetic diversity of Dicksonia navarrensis (sensu Noben et al., 2017) populations in Mexico; 100% of the specific loci included in the work were successfully amplified, and 81.81% were polymorphic, although it is possible that other alleles are masked (Izuno et al., 2012). Here, the polymorphism percentage obtained through the use of microsatellites, is similar to that estimated in D. sellowiana in Brazil, the values range between 72.73% (Nazareno et al., 2013) and 100% (Fagundes et al., 2020), with a total number of alleles similar to that of Mexican populations of D. navarrensis, 38 and 32, respectively (versus 33 in the present work). In Alsophila spinulosa populations of China, polymorphism estimated with SSR markers was 21.43% (Zhou et al., 2008), and in Sphaeropteris brunoniana populations of China, polymorphism was 80.6%, estimated using AFLP (Wang & Guan, 2011).

Dicksonia navarrensis presented low genetic diversity (mean H_E = 0.36) compared to other species of the same genera, such as D. sellowiana (mean H_E = 0.50, Fagundes et al., 2020), or with respect to Alsophila spinulosa (mean H_E = 0.45, Zhou et al., 2008), probably as a result of the high degree of fragmentation and loss of its habitat, the montane cloud forest of Mexico (INEGI, 2017), which affects the connectivity between its populations (Graignic et al., 2018). However, it must be considered that habitat loss and fragmentation in other fern species (Pleopeltis crassinervata and Polypodium rhodopleuron) did not have a strong influence on genetic diversity because it remained high; mean H_E = 0.257 and 0.223, respectively (Winkler et al., 2011). In addition to habitat alteration, the density of D. navarrensis populations in the SMO is rapidly declining and geographic isolation is increasing due to illegal harvesting of individuals (Reyes-Ortiz et al., 2019). However, it is necessary to analyze other populations of D. navarrensis to establish the relationship between population size and forest fragment size in the different mountainous regions where it is distributed in Mexico, to complement the results obtained here, since the consequences of habitat reduction on the genetic diversity of species cannot be generalized (Carvalho et al., 2019; Yu et al., 2020).

In this study, the low values of heterozygosity suggest inbreeding (Frankham et al., 2002), which can result from genetic drift events, founder effect, bottlenecks (Luikart, Allendorf, & Cornuet, 1998), and a reduction in population size (Lönn & Prentice, 2003; Pullin, 2002). It has been suggested that the plant mating system increases inbreeding (Wang et al., 2014). In ferns, in general, the probability of inbreeding is high, because the sporophytes are produced from gametophytes that can self-fertilize (Sessa et al., 2016). However, crossing over can be favored, both by the asynchronous production of ovules and sperm in the gametophyte phase, and by the high vagility of the spores, which allows them to colonize new habitats (Haufler et al., 2016). This happens according to the type of mating system they use: outcrossing versus selfing (Haufler et al., 2016; Nazareno et al., 2013). Alsophila firma, for example, has developed a post-zygotic mechanism that enables it to promote outcrossing in its populations (Ramírez-Barahona & Eguiarte, 2015); therefore, it is necessary to analyze the mating system of D. navarrensis, to find whether the specific cause of the deficiency in heterozygotes is attributable to natural selection or self-crossing (Wang et al., 2014).

Fagundes et al. (2020) note that the spores of D. sellowiana are very small, so they are easily dispersed by wind and rain, which facilitates the movement of individuals and interbreeding. This is plausible for this species, but the environmental conditions in which D. navarrensis grows are somewhat different. In Mexico, the primary MCF, which is the ecosystem that meets the ideal conditions for D. navarrensis, occupies between 0.4 and 0.9% of Mexican territory and is considered an ecosystem at risk of extinction (Gual-Díaz & Rendón-Correa, 2014; INEGI, 2017). Even though the dispersal capacity of the spores of this species is very high, the probability that they will find suitable conditions for germination, gametophyte production and sporophyte formation is low, but studies are required to confirm that argument. In contrast, populations of D. sellowiana are distributed over a larger geographic area in the Atlantic Forest of South America (Mallman et al., 2019), have a high density of individuals, and have a continuous influx of gametophytes into populations (Schmitz et al., 2006; Fagundes et al., 2020); these three characteristics could be related to the corresponding values of H_E, higher than those estimated for D. navarrensis in the present study.

Genetic Structure

The F_ST values suggest low differentiation among populations. At the population level, POP2 and POP3 had the lowest differentiation rate and a high Nm per generation (5285). The two populations are relatively close (16.1 km apart), they have similar latitude and elevation (1878 and 1732 m, respectively), and their orientation is similar as well. This suggests that both populations were at some time in the past a single population that now is in a process of fragmentation; recent studies indicate that gene flow is one of the most important factors affecting the population genetic structure, in some cases increasing genetic diversity by sexual reproduction among populations (Ranker, 1994; Yu et al., 2018). In this sense, Pelosi and Sessa (2021) point out that in ferns the value of Nm can be up to 24 migrants per generation; however, values above four indicate random mating between populations.

In ferns species, it has been determined that its genetic structure is affected by the mating system (Pelosi & Sessa, 2021), their evolutionary history (Soltis & Soltis, 1990), and the physical characteristics of their habitat: elevation, precipitation, and latitude, among others (Fagundes et al., 2020). The results indicate that the populations of D. navarrensis that are distributed in the northern part of the study area (POP1–POP3) maintain gene flow among them, which could explain their low level of differentiation. The probable lack of genetic exchange between these three populations and the rest could be an indication of isolation by distance, favored by the high degree of habitat fragmentation (Yu et al., 2020). This suggests a synergy effect between natural selection on disturbed populations and isolation by distance in the north that increases the similarity within populations but increases divergence between latitudinal groups. However, tree ferns produce billions of spores (Large & Braggins, 2004); D. sellowiana, for example, produces approximately 4 billion spores per kilogram (Gomes et al., 2006), which facilitates the formation of spore banks (Fiilippini et al., 1999). Moreover, it has been suggested that in general they are able to disperse even over distances greater than 3000 km (Tryon, 1970; Tryon & Tryon, 1982). In the case of D. navarrensis, a probable answer to found divergence is that, within the SMO, there is a wide range of variation in environmental factors such as elevation, climate, topography, edaphology, and degree of anthropic disturbance (Salinas-Rodríguez, 2018), which alone or in synergy, constitute barriers to spores survival or dispersal along different distances, thus inducing genetic isolation (Tryon, 1970; Tryon & Tryon, 1982). This environmental heterogeneity facilitates the existence of populations with a few isolated individuals, which can lead to the loss of genetic variation (Soltis & Soltis, 1990). Despite the large number of spores produced and their wide dispersal distance, well documented in ferns, the effects of multiple colonization contribute to loss of diversity and low genetic structuring in their populations (Vogel et al., 1999), which could explain the divergence between the three geographic groups in D. navarrensis.

Implications for Population Conservation in the Sierra Madre Oriental

Currently, the populations of D. navarrensis are being decimated through the illegal extraction of individuals, which are sold in local markets as ornamental plants (Reyes-Ortiz et al., 2019). In addition, changes in land use in their distribution area have made sites that meet the ideal high humidity characteristics necessary for the establishment of individuals increasingly scarce, which contributes to reducing their survival rate. The genetic structure and divergence found in this species could be related to the high environmental heterogeneity caused by drastic modifications in its habitat (Carvalho et al., 2019; Yu et al., 2020).

Genetic information on D. navarrensis and a recent ecological analysis of its population structure (Reyes-Ortiz et al., 2019) can be very useful in the design of a management and conservation plan for this species. The limited distribution and small population size of species in the SMO should be considered. On the other hand, the heterogeneity of the SMO (Salinas-Rodríguez, 2018) hardly represents an obstacle for illegal extraction of species. Therefore, we should look for alternatives to avoid the loss of D. navarrensis genetic resources, principally in the populations at the edge of local extinction, each population must be considered individually due their genetic differences (Shrestha et al., 2002), the close relationship between population size and genetic variation (Izuno et al., 2012), and also because information about the proportion of genetic variation that confers potential adaptability to this species is unknown (Kim et al., 2005). However, it is evident that we are still far from knowing the total genetic variation of the species (Shah et al., 2008).

The values of genetic diversity estimated for this species probably are the product of evolutionary processes, habitat fragmentation, and excessive extraction of individuals, whether for purposes of exploitation or not ( Supplement Material 3; Shah et al., 2008). The degree of habitat disturbance differs between populations, and is most severe in POP6 and POP1; the latter is an important genetic reservoir, containing particular adaptations. In situ and ex situ conservation are methods that complement each other (Vovides et al., 2013). The priority should be to categorize all populations in situ, and subsequently carry out ex situ work through the collection of spores or youngest individuals, initially in the populations of highest genetic diversity (Shah et al., 2008) such as POP1 in this study. The ecological characteristics of D. navarrensis populations (Reyes-Ortiz et al., 2019), and genetic information provided here, suggest develop strategies to conservation and management of MCF. If its habitat is conserved, D. navarrensis will be protected properly.