Biological material

Six populations representing the distribution of P. pinceana in northern Mexico (Table 1), specifically those located in the states of Zacatecas and Coahuila and the population recently registered for Nuevo León (Figure 1) (^{Favela, 2009}). 30 individuals separated from one another by a distance of 10 m were selected at random; 10 g of ripe leaves were collected manually from each and stored in plastic bags with 10 g of silica gel.

Figure 1 Geographic distribution of the sampled Pinus pinceana Gordon populations of the Eastern Sierra Madre. 

DNA extraction

The DNA was extracted from 1 g of dry leaves of the 30 individuals collected in each population, following the method of ^{Doyle and Doyle (1990}), and ammonium acetate was added to remove the excess carbohydrates (^{Hollingsworth et al., 1999}).

RAPD analysis

In order to select the oligos to be used in all the samples 15 random sequence decanucleotides of the OPA, OPB, OPC, OPG and OPP operon primers for RAPD (Operon RAPD kits technologies, Alameda). For this purpose, 16 of the 160 existing individuals were selected, and finally six with clear, reproductible shift patterns were chosen for the corresponding analyses.

The PCR reactions were prepared in a final volume of 25 µL of a mixture of the following reagents: 20 ng of genomic DNA in a 1X buffer (Invitrogen 10X) with 2.5 mM MgCl₂, 0.2 nM dNTP´s, 0.50 M primer, 2 % formamide, and 1 Unit of Taq polymerase (Invitrogen, Carlsbad, CA, USA). A control tube without genomic DNA having all the components of the reaction was included in order to rule out contamination.

PCR conditions

The DNA was amplified using a programmable thermal cycler (Thermo Electron Px2 Thermal Cycler PCR PCYL220 Issue 1, HBPX2110). The reaction conditions for RAPD utilized were: 94 °C (2 min) for the initial separation of the DNA chains; 2 touchdown cycles were subsequently added, in which the reaction alignment temperature decreased by 2 °C in each cycle. These were followed by 41 cycles of 93 °C (30 sec), 35 °C (1 min), 72 °C (2 min) and 72 °C (5 min) for the final stretch.

Detection and processing of PCR products with RAPD

The amplifications were replicated for each primer and for each population according to the standardized protocol. The DNA samples of the population were inserted in the agarose gels. The amplified products were separated in an agarose gel electrophoresis at 1.8 % in a TBE 1x buffer, with a constant difference of potential of 135 volts during 2 hours and 30 minutes. The gels were dyed during 15 minutes in a SB 1X buffer with ethidium bromide, and finally, the banding patterns were observed under UV light and photographed on a transilluminator (MultiDoc-It™ Imaging System, UVP^®, USA).

Data analysis

With the banding pattern obtained through the separation during the electrophoresis of the various amplified fragments, a presence-absence matrix was built and usually recorded as present bands (1), consisting of all the dominant homozygotes (AA) and heterozygotes (Aa), and as absent bands (0), consisting of recessive homozygotes (aa). Binary matrices were developed for each primer and population; these were then used to estimate the genetic variability and structure of P. pinceana populations.

The binary matrix was analyzed using the POPGENE v32 software (^{Yeh et al., 1997}) to measure the genetic variability. The genetic variation values for each population were: the Shannon diversity index (I) (^{Lewontin, 1972}) and the polymorphic loci percentage (P). The statistical analysis of ^{Nei's genetic diversity (1973}) was used to calculate the genetic differentiation between populations (Gst).

The genetic distance (Jaccard) was calculated based on the binary matrix and was subsequently subjected to a Principal Coordinates (PCO) analysis using the PAST software, version 3 (^{Hammer et al. 2001}), which produces a visual representation of the genetic relationship within and between the populations. Furthermore, a dendrogram was built using the UPGMA (Unweighted Pair Group Method with Arithmetic averages) method based on the genetic distance between pairs of populations. The bootstrap analysis was carried out with 1 000 repetitions. The FAMD (Fingerprint Analysis with Missing Data) software, version 1.25, was used for both analyses (^{Schlüter and Harris, 2006}).

The analysis of molecular variance (AMOVA) was carried out for the structure of the population in order to evaluate the variation between and within the populations using the Harlequin software (^{Schneider et al., 2000}). The levels of significance for AMOVA are computed by means of non-parametric permutations of the set of data with 100 permutations. The Arlequin software generates a statistic-ɸ (^{Excoffier et al., 1992}) analogous to Wright's FST (Wright, 1951). This method has been used with RAPD data (^{Favela, 2004}, ²⁰¹⁰).

Amplification using RAPD markers

Six of the first 15 assayed markers yielded reproductible band patterns that are easily resolved; these were used with 180 sampled individuals. The six primers generated a total of 76 bands whose sizes ranged between 1 500 and 250 pb. A range of 10 to 16 bands per primer was obtained.

Data analysis

Genetic diversity. The analysis of the genetic diversity of the populations carried out with POPGENE revealed that the population of Cuatrocienégas, Coahuila, exhibits a larger genetic diversity, as shown in Table 2, with high I values (I= 0.44 and I = 0.43). The lowest value corresponded to the population of Las Lajas, Zacatecas (I= 0.37).

The average polymorphic loci (P) of the populations ranged between 77.32 and 84.21 %. The population of La Noria Cuatrocienégas, Coahuila, exhibited the greatest polymorphism, and the one with the least polymorphism was that of Las Lajas, Zacatecas. (Table 2).

It should be noted that the populations with the highest diversity values are those La Noria and La Palmosa, in Cuatrociénegas, Coahuila; both are located on the northernmost strip of the distribution interval of P. pinceana.

Genetic differentiation

The total coefficient of genetic differentiation (Gst) between the populations was 0.15. This value describes how the variation is distributed between the studied populations and ranges between 0 to 1; the value closest to 0 means that the populations share more genes, and the closer the value is to 1, the further the populations are from each other and the less genetic material they share.

Genetic structure of P. pinceana populations

The AMOVA indicates that (85.18 % of the total variation is registered within the populations (P< 0.001). Significantly, 14. 82 % of the remaining variation is found between the six P. pinceana populations (P< 0.001) (Table 3). A FST value of 0.1482, equivalent to the GST value cited above (0.15), was observed.

Genetic distance

Jaccard's distance was utilized to carry out an analysis of principal coordinates, which is a graphic representation of the genetic relationship between the six analyzed P. pinceana populations. The first two principal components of the distance for the RAPD describe 4.64 and 4.24 % of the variation, respectively. The samples of the individuals of each population were represented as a continuous dispersion, with the populations of La Noria, La Palmosa, La Casita and El Jaralito in the state of Coahuila. The populations of Cañón del Moroso, Nuevo León and Las Lajas de Zacatecas appear clearly separate from the rest of the populations in a distinct group (Figure 2).

Figure 2 Graph of principal coordinates of the genetic distance for six Pinus pinceana Gordon populations (Jaccard Similarity Coefficient). 

The dendrogram shown in Figure 3 was obtained from the genetic distance matrix of pairs using the unweighted pair clustering method with arithmetic average; it revealed that the populations of Coahuila are closer and constitute two clades: one consisting of the populations of Cuatro Cienégas (La Noria and La Palmosa), and another, of the populations of Parras and General Cepeda (La Casita and El Jaralito). Equally, the populations of Nuevo León (El Moroso Canyon) and those of Zacatecas (Las Lajas) are a separate clade, as they are genetically close. The bootstrap values for each population were highly significant and range between 96 and 100.

Figure 3 Dendrogram of the groups of the six Pinus pinceana Gordon populations, according to the unweighted pair clustering analysis with arithmetic average based on the Jaccard distance estimated with RAPD markers. 

Genetic diversity between Pinus pinceana populations

Estimated using the Shannon index (I=0.48), the genetic diversity follows a pattern in which the northernmost populations have the highest values: La Noria and La Palmosa are first (0.44 and 0.43), followed by El Jaralito and Sta. Catarina (0.41), La Casita and Las Lajas (0.39 and 0.37); we may infer that the framentation may have triggered the divergence between the populations, some of which generated new, unique alleles as a result of the isolation.

However, in average, the Shannon indices obtained for all the studied populations (0.48) do not differ from that of other pine species (0.45) and of other conifers and brodleaves (0.53) (Araucaria, Fitzorya and Cedrela) (^{Gillies et al.;1997}; ^{Allunt et al.,1999}; ^{Bekessy et al., 2002}).

The polymorphic loci percentage is another important indicator for determining the level of genetic variation of an area, as a species with a high value exhibits a high adaptability to the environment and, conversely, a species with a poor adaptability to the environment may be eliminated through natural selection (^{Zhang et al., 2013}); the value for P. pinceana was relatively high (79.80 %) in contrast with other pinyon pine species like P. culminicola Andresen & Beaman (57.3 %), (^{Favela, 2010}), P. cembroides var. bicolor Little and P. johannis M.-F. Robert, with 69.9 and 78.1 %, respectively (^{Favela, 2004}). This shows that each population contains a considerable fraction of the genetic diversity of the species, with values above those documented for endemic species (^{Hamrick et al., 1992}; ^{Hamrick and Godt, 1996}).

In general, the genetic variation in terms of the various diversity indices exhibits a geographic pattern as higher values are observed in the populations of Coahuila, the northernomost of which -La Noria and La Palmosa in Cuatro Ciénegas- have the highest value, followed by those of La Casita and El Jaralito. These populations are separated from those of Nuevo León and Zacatecas, probably due to the existence of geographic barriers that have favored their isolation and, as the information flow between populations is determined by the geographic distance. Nevertheless, they still share alleles, perhaps because they were originally all one, and with the passage of time this original population became fragmented -a fact that favored the onset of the divergence between the populations, which in turn generated new, unique alleles in some of them as it accomplishes to finish the genetic structure of the species as fragmented in three patches, given that in its study includes six populations, each of which marks the fragmentation, allotting two for each patch to the north, center and south of the geographic distribution.

Genetic differentiation between Pinus pinceana populations

The statistic (FST) is utilized to estimate the proportion of the genetic variation within and between the populations; it is particularly useful for dominant data such as RAPD (^{Excoffier, 2001}), whose variation ranges between 0 and 1. Values closer to 0 indicate that the allele frequencies are the same in all the populations and that there has been no differentiation; the maximum value possible is 1, when each population is fixed at different alleles (^{Piñero et al., 2008}). The value in the studied populations (0.15) was comparable to the FST (0.152) registered by ^{Ledig et al. (2001}), but it was lower than those estimated by ^{Molina-Freaner (2001}) (0.24) and ^{Ramírez-Herrera (2007}) (0.16) for Pinus pinceana using isoenzyimes, and for other Pinus taxa located in scattered, isolated populations (^{Hamrick, 2004}).

This value indicates that 15 % of the genetic diversity of the species corresponds to the one existing between populations and is very close to the estimate for other pinyon pine, P. rzedowskii Madrigal & M. Caball. (17 %) (^{Delgado et al., 1999}), an endemic taxon in Mexico also distributed in fragmented, isolated populations. It is also worth mentioning that a FST above 0.10 is regarded as indicating a strong interpopulational genetic differentiation, which is unusual in allogamous species like those of the Pinus genus (^{Hamrick et al., 1992}; ^{Ledig, 1998}). Only in certain conifers with a restricted natural distribution or with a strong fragmentation of their populations have similar genetic differentiation values been documented between populations (^{Ledig, 1998}).

Most of the variation in Pinus pinceana populations was registered within these (85.18 %). This result is similar to those registered for other pine species (^{Ledig, 1998}; ^{Favela, 2010}) and woody plants (^{Hamrick et al., 1992}). Furthermore, they are comparable to data obtained from RAPD, in which most of the examined species exibit high levels of interpopulational genetic variation (^{Xue et al., 2006}; ^{Favela, 2010}), while the degree of genetic variation between populations was 14.82 %. High levels of diversity and differentiation between populations seem to be usual among Mexican conifer taxa (^{Ledig et al., 2001}; ^{Fazekas and Yeh, 2006}; ^{Favela, 2010}).

Genetic and geographic distances between Pinus pinceana populations

Figure 2, which depicts the Jaccard similarity coefficient and is subjected to the PC analysis, evidences a separation between the populations of Nuevo León and Zacatecas and those of Coahuila, indicating that they are genetically the most distant. From this it may be inferred that there is a greater genetic flow between the populations of Coahuila (La Noria, La Palmosa, La Casita in Parras and El Jaralito) than between those of Nuevo León (Cañón del Moroso) and Zacatecas (Las Lajas). Previous studies (^{Ramírez-Herrera, 2007}) also show a separation with respect to the populations of Sierra de Parras and Las Norias, respectively.

According to ^{Hollingsworth and Ennos (2004}), the clustering of individuals of different populations in the same clade in a tree is interpreted as evidence of the genetic flow between populations. This can be equally observed in the tree obtained using the UPGMA method because two clades are formed; the first groups the populations of Coahuila and comprises two subclades, one encompassing the populations of Cuatro Ciénegas -La Noria and La Palmosa-, and another, those of Parras and Gral. Cepeda. The populations of Cañón del Moroso, Sta. Catarina, Nuevo León and Las Lajas, Zacatecas, form a single group. This makes it possible to infer that there is a greater genetic flow between the populations of Nuevo León and Zacatecas, compared to the populations of Coahuila. This is an important datum, for in previous studies (^{Ledig, 2001}; ^{Ramírez-Herrera, 2007}) the populations of Zacatecas were separated from among the northern population groups, together with those of Coahuila.

The variation between the populations may be due to the ecological differences in each of the two groups. The first is constituted by regions with less precipitation (350 a 400 mm) than the regions belonging to the second group -where the precipitation is 400 a 500 mm- and, in general, by the climatic and orographic regions defined as the Eastern Sierra Madre.

Various studies on P. pinceana record similar data, separating the northern populations (Coahuila) from those of central and southern Mexico; now, with the results obtained from the recently described population of Nuevo León (Favela et al., 2009), the manner in which they are separated from those of Zacatecas is evident.

In researches by ^{Ledig et al. (2001}) it is possible to distinguish the formation of two clades clearly separating the northern from the southern populations, as well as the populations of Las Lajas and Parras, Coahuila, also included in this study, in subclades. ^{Ramírez-Herrera (2007}) has submitted a graph in which two clades are formed: one for the northern populations only, and another with those of the center and the south. In the northern populations we may observe two subclades, one of which separates the population of Las Norias, considered in the present study, while the other subclade consists of the populations of Zacatecas. Finally, ^{Villarreal et al. (2009}) conclude that there are two groups of P. pinceana pinyon pine forests, based on their floristic and distribution differences: those of the northern region (Coahuila, Zacatecas and San Luis Potosí) and those of the southern region (Querétaro and Hidalgo).