Introduction
Jatropha curcas L. is a new bioenergy crop in the world, belonging to the family Euphorbiaceae, the centre of origin, diversification and domestication of which is probably the Mesoamerican region (Dehgan and Webster, Reference Dehgan and Webster1979; Heller, Reference Heller1996; Carels, Reference Carels, Kader and Delseny2009; Ovando-Medina et al., Reference Ovando-Medina, Adriano-Anaya, Vázquez-Ovando, Ruiz-González, Rincón-Rabanales and Salvador-Figueroa2013). Although this species has been used for millennia as a hedge or ‘living fence’ and some genotypes are edible, many authors consider it to be still in the process of domestication (Carels, Reference Carels, Kader and Delseny2009; Granados-Galván, Reference Granados-Galván2009; Achten et al., Reference Achten, Nielsen, Aerts, Lengkeek, Kjaer, Trabucco, Hansen, Maes, Graudal, Akinnifesi and Muys2010). It has other important uses; however, its current relevance is the possibility of converting the seed oil into biodiesel (Gubitz et al., Reference Gubitz, Mittelbach and Trabi1999).
Due to its importance as a newly extensive crop, it is estimated that in a few years J. curcas will have been planted in nearly 15 million hectares in Mexico as well as in Asia, Africa and the Americas (Renner and Zelt, Reference Renner and Zelt2008). However, there are still many technological challenges for its commercial establishment, including the lack of genetically improved varieties of high agronomic performance (Mishra, Reference Mishra2009; Ovando-Medina et al., Reference Ovando-Medina, Adriano-Anaya, Vázquez-Ovando, Ruiz-González, Rincón-Rabanales and Salvador-Figueroa2013).
Knowledge of the diversity of J. curcas on distinct levels (molecular, morphological, physiological and productive) is of primary importance, as from germplasm characterization genotypes could be selected for extensive planting (Basha and Sujatha, Reference Basha and Sujatha2007) or genetic improvement could be accelerated (Singh et al., Reference Singh, Singh, Mishra and Bhatia2010). Molecular studies are particularly important as DNA-based markers are not influenced by environmental conditions (Ganesh et al., Reference Ganesh, Parthiban, Senthil-Kumar, Thiruvengadam and Paramathma2008) and therefore reflect genetic variation more faithfully. In this respect, the molecular diversity of J. curcas has been studied, although mainly germplasm collected in Asia (Abdulla et al., Reference Abdulla, Janagoudar, Biradar, Ravikumar, Koti and Patil2009; Yu et al., Reference Yu, Sun, Wud and Peng2010), Africa (Basha et al., Reference Basha, Francis, Makkar, Becker and Sujatha2009; Ambrosi et al., Reference Ambrosi, Galla, Purelli, Barbi, Fabbri, Lucretti, Sharbel and Barcaccia2010; Ricci et al., Reference Ricci, Chekhovskiy, Azhaguvel, Albertini, Falcinelli and Saha2012), South America (Rosado et al., Reference Rosado, Laviola, Faria, Pappas, Bhering, Quirino and Grattapaglia2010) and, to a lesser extent, Mexico and Central America (Van-Loo et al., Reference Van-Loo, Jongschaap, Montes-Osorio and Arzudia2008; Ovando-Medina et al., Reference Ovando-Medina, Sanchez-Gutierrez, Adriano-Anaya, Espinosa-Garcia, Núñez-Farfán and Salvador-Figueroa2011a; Pecina-Quintero et al., Reference Pecina-Quintero, Anaya, Zamarripa, Montes, Núñez, Solís, Aguilar, Gill, Langarica and Mejía2011) has been characterized. Additionally, the marker systems used have been mainly dominant, such as random amplified polymorphic DNA (RAPD), amplified fragment length polymorphism (AFLP) and inter-simple sequence repeats (ISSRs), among others (Ovando-Medina et al., Reference Ovando-Medina, Espinosa-García, Núñez-Farfan and Salvador-Figueroa2011b).
On the other hand, microsatellite markers (simple sequence repeats (SSRs)) have the advantage of detecting heterozygous individuals, so they are considered to be co-dominant, as well as of having a greater number of polymorphisms than dominant markers (Kalia et al., Reference Kalia, Rai, Kalia, Singh and Dhawan2011). However, in the case of J. curcas, SSR-based studies have generally detected less diversity in Asian and South American germplasm (Basha et al., Reference Basha, Francis, Makkar, Becker and Sujatha2009; Rosado et al., Reference Rosado, Laviola, Faria, Pappas, Bhering, Quirino and Grattapaglia2010; Ricci et al., Reference Ricci, Chekhovskiy, Azhaguvel, Albertini, Falcinelli and Saha2012) compared with those found with dominant markers. AFLP-based studies conducted with germplasm from the state of Chiapas, Mexico, and from Guatemala have detected the highest genetic diversity of this species, although questions remain about the structure and variation that can be detected with co-dominant markers. For that reason, this study was carried out with the main aim of analysing, using SSR markers, the genetic structure and diversity of populations of J. curcas located in the centre of diversification of this species.
Materials and methods
Plant material
Plants from the Biosciences Centre (CenBio, initials in Spanish) Germplasm Bank of the Autonomous University of Chiapas (Mexico) were used and samples were obtained from cuttings. We selected 93 representative samples from seven sites of two regions in the state of Chiapas (Table 1) grouped by geographical proximity.
Table 1 Samples of Jatropha curcas (Euphorbiaceae) from the CenBio Germplasm Bank of the Autonomous University of Chiapas collected in Southern Mexico
DNA extraction
Samples used were fresh foliar tissues, which were transferred on ice to the laboratory, where they were washed three times with sterile distilled water and 70% ethyl alcohol and maintained at − 30°C until processing. The genomic DNA was isolated and purified by the method described by Doyle and Doyle (Reference Doyle and Doyle1990) modified for processing samples of 200 mg (Ovando-Medina et al., Reference Ovando-Medina, Sanchez-Gutierrez, Adriano-Anaya, Espinosa-Garcia, Núñez-Farfán and Salvador-Figueroa2011a). In brief, 200 mg of healthy leaves and 60 μg of polyvinylpyrrolidone were ground with liquid nitrogen and then extracted with 1 ml of CTAB buffer (0.1% w/v hexadecyltrimethylammonium bromide, 5 mM EDTA, 1.5 M NaCl, 50 mM Trizma base and 0.1% v/v β-mercaptoethanol, pH 8.0). Extractions were carried out with chloroform–isoamyl alcohol and precipitation with isopropanol. The extracted DNA was purified by re-extraction with a mixture of phenol–chloroform–isoamyl alcohol (25:24:1). The integrity of DNA, dissolved in 60 μl of Milli-Q® water (Sigma-Aldrich®, St. Louis, MO, USA), was verified by electrophoresis on 1% agarose gel and quantified spectrophotometrically at 260 nm (Genova Nano®; Jenway®, Staffordshire, UK).
Microsatellite DNA study
Ten pairs of specific oligonucleotides were used for regions of microsatellite DNA of the J. curcas genome reported by Pamidimarri et al. (Reference Pamidimarri, Sinha, Kothari and Reddy2009a). The descriptive data of the oligonucleotides are given in Table S1 (available online).
Individual amplifications were carried out by polymerase chain reaction (PCR) following the protocol reported by Pamidimarri et al. (Reference Pamidimarri, Sinha, Kothari and Reddy2009a), as modified with the following reaction mixture: 25 ng of DNA, 50 pmol of each oligonucleotide, 4 μl of Taq® 10 × buffer, 2 μl of MgCl2 (25 mM), 0.5 μl of dNTPs (10 mM) and 1 U of Taq polymerase (GoTaq® Flexi DNA Polymerase; Promega®, Madison, WI, USA); the final volume was 25 μl. The thermal profile consisted of an initial denaturation step at 94°C for 4 min and 40 amplification cycles of three steps each: 30 s of denaturation at 94°C, 30 s of alignment at the temperature specific for each pair of oligonucleotides (Table S1, available online) and 1 min of polymerization at 72°C. It involved a final extension step at 72°C for 7 min, followed by cooling at 15°C. The PCR products were stored at − 30°C until electrophoresis. Negative controls without DNA were used to rule out contamination in the samples.
The amplification products were separated by electrophoresis on 12% acrylamide gels in TAE buffer (48.4 g Trizma base, 10.9 g acetic acid, 2.92 g EDTA and 1 litre H2O) at 80 V for 250 min and revealed with ethidium bromide (Sambrook et al., Reference Sambrook, Fritsch and Maniatis1989). The size of the bands was estimated using the molecular marker 25 bp DNA Step Ladder (Promega®, Madison, WI, USA). Image analysis was carried out using the Universal Hood (ChemiDoc®; Model 170-8126 Bio-Rad®, USA) equipment coupled to the Quantity One© software (Bio-Rad®, Hercules, CA, USA).
Population analysis
The locus informative descriptors were calculated using the program PicCal© 1.0 (Vekemans et al., Reference Vekemans, Beauwens, Lemaire and Roldán-Ruiz2002) and those of diversity of populations using the program GenAlEx© version 6.5 (Peakall and Smouse, Reference Peakall and Smouse2012), specifically the percentage of polymorphism (P%), the effective number of alleles (N e), Shannon's informative index (I), observed heterozygosity (H o) and expected heterozygosity (H e), and fixation index (F). To test the hypothesis of isolation by distance, a Mantel correlation test was conducted between Nei's genetic distance and geographical distance, which was performed with 10,000 permutations using the GenAlEx© program, version 6.5. The degree of differentiation within and among the populations and regions was determined by analysis of molecular variance (AMOVA), while estimating F statistics, including the F ST. The AMOVA was conducted with 100,000 permutations using the Arlequin© software version 3.5.1.3 (Excoffier and Lischer, Reference Excoffier and Lischer2010). To show the existence of evolutionary forces acting on individuals and populations, a series of tests of Hardy–Weinberg equilibrium (HWE) (for each locus) was conducted with the same software. The genetic population structure was studied with Bayesian statistics using the Structure© program, version 2.3.2. (Pritchard et al., Reference Pritchard, Stephens and Donnelly2000), with 100,000 burn-in iterations and 1,000,000 iterations after burn-in or Markov Chain Monte Carlo simulation. We used a model of admixtured ancestry, with 20 repetitions of each set of genetic populations (K1–K20). The value of K was estimated following the procedure of Structure Harvester© program version 0.6.93 (Earl and Vonholdt, Reference Earl and Vonholdt2012), which uses the procedure of Evanno et al. (Reference Evanno, Regnaut and Goudet2005) for estimating the actual number of genetic groups. For comparison purposes, a database of AFLP markers was used for the same populations and individuals, created by Sánchez-Gutiérrez (Reference Sánchez-Gutiérrez2010) and Ovando-Medina et al. (Reference Ovando-Medina, Sanchez-Gutierrez, Adriano-Anaya, Espinosa-Garcia, Núñez-Farfán and Salvador-Figueroa2011a), from which diversity descriptors for dominant markers were determined.
Results
Marker informativeness
The ten pairs of oligonucleotides allowed for the detection of 27 amplified fragments, of which 77.78% were polymorphic, and it was possible to distinguish homozygous individuals from heterozygous ones (Fig. 1). The microsatellite regions Jcds58 and Jcps6 were found to be monomorphic, as they only exhibited alleles of 108 and 110 bp, respectively, in all the individuals studied. The average number of alleles (N a) per locus was 2.286, and the most informative microsatellite was Jcps20, which exhibited nine alleles (158, 196, 200, 208, 214, 224, 230, 234 and 244 bp), with a polymorphic index content (PIC) value of 0.784. Only two alleles were private: 158 bp from the locus Jcps20 for the population Soconusco and 114 bp from the locus Jcds10 for the population Riviera. Table 2 lists the alleles and PIC data of all the studied microsatellites.
Fig. 1 Acrylamide gel showing alleles of 150 and 170 bp of the microsatellite Jcds24 in 13 accessions of Jatropha curcas of Chiapas, Mexico. Accessions CHIC1 and OCZ1 are heterozygous for this locus. M: Molecular marker 25 bp DNA Step Ladder (Promega®, Madison, WI, USA).
Table 2 Informativeness of microsatellite loci for the populations of Jatropha curcas from Chiapas, Mexicoa
PIC, polymorphic information content.
a The data were obtained by amplification of 93 samples from seven populations.
b Data from Pamidimarri et al. (Reference Pamidimarri, Sinha, Kothari and Reddy2009a, Reference Pamidimarri, Singh, Mastan, Patel and Reddyb, Reference Pamidimarri, Mastan, Rahman, Ravi Prakash, Singh and Reddy2010), Ricci et al. (Reference Ricci, Chekhovskiy, Azhaguvel, Albertini, Falcinelli and Saha2012); and Vischi et al. (Reference Vischi, Raranciuc and Baldini2013).
Genetic diversity
The polymorphism per population ranged from 37.5% (Mapastepec) to 87.5% (Centre), with an average of 58.9%. There was a marked difference between the populations from the Chiapas Coast, which had 50% or less polymorphism, and those from Central Chiapas, with values >62.5%. Table 3 summarizes the diversity descriptors as found, indicating that in all the populations the observed N a was greater than the N e.
Table 3 Diversity descriptors from seven populations of Jatropha curcas from Chiapas, Mexico
n, Number of individuals; N a, number of alleles; N e, effective number of alleles; I, Shannon's information index; H o, observed heterozygosity; H e, expected heterozygosity; F, fixation index.
The I value ranged from 0.324 in the least diverse population (Mapastepec) to 0.538 in the most diverse one (Centre), with an overall average of 0.427. Populations from the coastal region of Chiapas exhibited heterozygote deficiency, as their values of H e were higher than those of H o and therefore their F were positive. On the other hand, populations from the central region of Chiapas had negative values of F, indicating excess of heterozygotes (Table 3).
Reproductive pattern
The HWE tests revealed that the majority of loci were in equilibrium; that is, the genotype frequencies observed for each locus in each population were consistent with expectations. However, the loci Jcps20, Jcds66 and Jcms21 departed significantly from HWE, and in particular, the locus Jcps20 was in disequilibrium in all the populations (Table S2, available online), indicating a reproductive pattern of non-random mating.
Genetic structure of populations
The AMOVA (Table 4) revealed that the populations were structured and moderately differentiated. As expected, the largest proportion of variation was found within the populations; however, the F ST differentiation index had a value of 0.087 (P< 0.05), indicating that there are important genetic differences between the regions (F RT= 0.056; P< 0.000) and, to a lesser extent, among the populations within each region (F SR= 0.033; P< 0.000). The differentiation found among the populations was not due to isolation by distance, as the Mantel test was not significant (r(x, y) = 0.513, P= 0.137, r 2= 0.263). Evidence was found that differentiation might be due to allopatry (i.e. the speciation process that occurs when populations become isolated from each other preventing the genetic interchange), with the mountain range Sierra Madre de Chiapas being the main geographical barrier between the coastal and central regions. Bayesian analysis revealed that accessions belonged to only four genetic clusters (ΔK= 4) and confirmed the differentiation between the regions. In Fig. 2, the proportion of alleles in each genetic group (four colours) for each individual (upper vertical bar) for each geographical population is shown. Pie charts for each geographical population are also shown in the Chiapas satellite image in the figure. In each population, there are a relatively large fraction of alleles (colours) from other populations, showing mixed ancestry.
Table 4 Analysis of molecular variance of 93 accessions of Jatropha curcas, belonging to seven populations and two regions of Chiapas, Mexico
a Obtained using data from eight microsatellite loci.
Fig. 2 Genetic structure of seven populations of Jatropha curcas in Chiapas, Mexico. Upper part scale shows individuals (vertical bars) and their ancestry proportion. This was obtained using the Structure© 2.3.2. software (Pritchard et al., Reference Pritchard, Stephens and Donnelly2000) with 100,000 burn-in, 1,000,000 repetitions after burn-in and 20 runs of each K (1–20). Ancestry model: admixtured. The K value was 4 (Evanno et al., Reference Evanno, Regnaut and Goudet2005).
Discussion
The microsatellite loci used in this study were selected for being among the most informative from previous studies (for a review, see Achten et al., Reference Achten, Nielsen, Aerts, Lengkeek, Kjaer, Trabucco, Hansen, Maes, Graudal, Akinnifesi and Muys2010); however, two of them were monomorphic and a further five had PIC values < 0.100 (Table 2), showing that such markers are scarcely informative for at least the populations studied. Nevertheless, the eight polymorphic loci did allow for estimating population diversity, because they all had the capacity for detecting heterozygotes as three of them had high PIC values, especially Jcps20 (PIC = 0.784). The last locus exhibited nine alleles, while only four had been reported (Pamidimarri et al., Reference Pamidimarri, Sinha, Kothari and Reddy2009a). In general, the alleles detected were within the range of those that had been detected in the Asian accessions of J. curcas (Pamidimarri et al., Reference Pamidimarri, Singh, Mastan, Patel and Reddy2009b; Ricci et al., Reference Ricci, Chekhovskiy, Azhaguvel, Albertini, Falcinelli and Saha2012), although there were 20 alleles that had not been found in this species. Also, the degree of polymorphism (eight of ten microsatellites) was similar to that reported by Pamidimarri et al. (Reference Pamidimarri, Singh, Mastan, Patel and Reddy2009b), who found that seven of the 12 microsatellites studied were polymorphic. These findings let us infer that further research is required for more hypervariable co-dominant markers, as those available might not detect variation in populations of J. curcas with low diversity, such as Asian (Sun et al., Reference Sun, Li, Li, Wu and Ge2008; Wen et al., Reference Wen, Wang, Xia, Zou, Lu and Wang2010; Sato et al., Reference Sato, Hirakawa, Isobe, Fukai, Watanabe, Kato, Kawashima, Minami, Muraki, Nakazaki, Takahashi, Nakayama, Kishida, Kohara, Yamada, Tsuruoka, Sasamoto, Tabata, Aizu, Toyoda, Shin-i, Minakuchi, Kohara, Fujiyama, Tsuchimoto, Kajiyama, Makigano, Ohmido, Shibagaki, Cartagena, Wada, Kohinata, Atefeh, Yuasa, Matsunaga and Fukui2011; Yadav et al., Reference Yadav, Ranjan, Asif, Mantri, Sawant and Tuli2011) and South American (Rosado et al., Reference Rosado, Laviola, Faria, Pappas, Bhering, Quirino and Grattapaglia2010). On the other hand, and in case of confirmation of absence in other populations in a larger study, the two specific or private bands could be isolated, sequenced and converted to robust sequence characterized amplified region (SCAR)-type markers. The availability of SCAR markers would facilitate the detection of migrant genotypes in other populations, whether by natural causes or human dispersal.
Recent years have seen an exponential increase in the number of studies carried out on the genetic diversity of J. curcas; however, most of the research has been based on dominant molecular markers (for a review, see Ovando-Medina et al., Reference Ovando-Medina, Espinosa-García, Núñez-Farfan and Salvador-Figueroa2011b). The dominant marker systems, such as AFLP, RAPD, and ISSR, among others, have the advantage of generating a large number of DNA bands, but have the disadvantage of not revealing the proportion of homozygous and heterozygous individuals. Among the several co-dominant markers, the SSRs or microsatellite DNAs have several advantages due to their co-dominant nature, are abundant in the genome, are highly reproducible, possess hyperpolymorphism, are rapidly evolving, and can be applied both on an infraspecific level and between related species (Yadav et al., Reference Yadav, Ranjan, Asif, Mantri, Sawant and Tuli2011). The type of marker used can skew the values of diversity indices; for example, the use of SSR markers entails the risk of overestimating the diversity indices, especially when using a less number of loci, owing to the high allelic variability in SSR sequences (Xiang et al., Reference Xiang, Song, Wang, Chen, Yang and Long2007). Nevertheless, studies using SSR markers have found low-to-moderate genetic diversity in the germplasm of J. curcas in Asia and South America (Sun et al., Reference Sun, Li, Li, Wu and Ge2008; Rosado et al., Reference Rosado, Laviola, Faria, Pappas, Bhering, Quirino and Grattapaglia2010; Wen et al., Reference Wen, Wang, Xia, Zou, Lu and Wang2010; Sato et al., Reference Sato, Hirakawa, Isobe, Fukai, Watanabe, Kato, Kawashima, Minami, Muraki, Nakazaki, Takahashi, Nakayama, Kishida, Kohara, Yamada, Tsuruoka, Sasamoto, Tabata, Aizu, Toyoda, Shin-i, Minakuchi, Kohara, Fujiyama, Tsuchimoto, Kajiyama, Makigano, Ohmido, Shibagaki, Cartagena, Wada, Kohinata, Atefeh, Yuasa, Matsunaga and Fukui2011; Yadav et al., Reference Yadav, Ranjan, Asif, Mantri, Sawant and Tuli2011). The present study has demonstrated a low informative level of microsatellites reported for J. curcas (see Achten et al., Reference Achten, Nielsen, Aerts, Lengkeek, Kjaer, Trabucco, Hansen, Maes, Graudal, Akinnifesi and Muys2010, for a review of the SSR loci discovered in this species); however, this is not the cause for the low polymorphism detected in the accessions of Old World and South America, but rather its low diversity, as high diversity was found in the accessions of Chiapas at the same loci. These results confirm the results reported by Ambrosi et al. (Reference Ambrosi, Galla, Purelli, Barbi, Fabbri, Lucretti, Sharbel and Barcaccia2010), who studied microsatellite allelic variation in J. curcas from around the world and included nine individuals of Jalisco, Mexico, finding that the greatest genetic diversity was in the Mexican germplasm.
In general, the genetic diversity of J. curcas in Mexico estimated with SSR markers is similar to that estimated using dominant markers such as AFLP (Ovando-Medina et al., Reference Ovando-Medina, Sanchez-Gutierrez, Adriano-Anaya, Espinosa-Garcia, Núñez-Farfán and Salvador-Figueroa2011a). In the present study, a comparison of the diversity detected with AFLP and SSR markers (see the Materials and methods section) was made, from which it may be concluded that although global values of H e (0.230) and I (0.427) obtained using SSR markers are higher than those obtained using AFLP markers (H e= 0.190, I= 0.294), the diversity pattern is the same; that is, populations from Central Chiapas were more diverse (I= 0.340) than those from the Chiapas Coast (I= 0232). In another study of Jatropha populations of Chiapas using fatty acids as chemical markers, a genetic barrier was detected, isolating populations from Central Chiapas with respect to those from Guatemala and the Chiapas Coast (Ovando-Medina et al., Reference Ovando-Medina, Espinosa-García, Núñez-Farfan and Salvador-Figueroa2011c). This confirms the differentiation between J. curcas populations at the regional level.
The results of this study support the hypothesis of Mesoamerica being the centre of diversification of the genus Jatropha, and J. curcas, in particular, although other studies have obtained results that do not support this idea. For example, Vischi et al. (Reference Vischi, Raranciuc and Baldini2013) analysed 26 loci, but only seven exhibited polymorphism (16 alleles) in 29 J. curcas accessions from six countries, including Mexico. Most of the genotypes that they studied were completely or highly homozygous and the genetic variability was very low, even in accessions from Mexico. These authors attribute their results to the sampling method, because they collected isolated plants instead of populations. We hypothesize that the other possibility is that sampling was biased towards non-toxic genotypes, whereas in Mexico most of the J. curcas genotypes are toxic.
Regarding the genetic structure, the AMOVA revealed that the largest proportion of variation was within the populations; however, the differentiation index F ST (0.087) indicated that the populations were moderately differentiated, which was mainly due to the contrast between the regions (F RT= 0.056). In the same vein, it was found that the values of fixation index (F), which indicates the rate of inbreeding in a population, were negative in populations from Central Chiapas, indicating excess of heterozygotes, while populations from the Chiapas Coast had heterozygote deficiency (F positives). In the same vein, Bayesian analysis revealed that the seven populations declared a priori had alleles of only four genetic groups, indicating genetic migration (mixed ancestry) among the populations within regions. However, there exists a clear differentiation between the central and coastal regions of Chiapas, especially between Soconusco and Border (Fig. 2), demonstrating that Sierra Madre de Chiapas is an effective geographical barrier that has prevented gene flow between them. The mountain range called Sierra Madre de Chiapas arose about the middle Miocene–early Pliocene (Burkart, Reference Burkart1978; Aguayo and Trápaga, Reference Aguayo and Trápaga1996), probably before the Jatropha colonization of Chiapas (after the closure of the Isthmus of Panama, about three million years ago; Carels, Reference Carels, Kader and Delseny2009; Ovando-Medina et al., Reference Ovando-Medina, Adriano-Anaya, Vázquez-Ovando, Ruiz-González, Rincón-Rabanales and Salvador-Figueroa2013), possibly indicating two routes of Jatropha dispersion: throughout the coastal zone and to the North of the mountain. Furthermore, differentiation was not due to isolation by distance (Mantel P= 0.137), but probably due to allopatry. It is known that genetic differentiation among populations depends on gene flow through pollen and seed dispersal (Loveless and Hamrick, Reference Loveless and Hamrick1984) and in the case of cultivated plants, such as J. curcas, by human dispersion. The latter, together with the fact that the plant is mainly propagated by cloning, may explain the differentiation pattern found, as farmers tend to exchange germplasm on a regional level and less likely among the regions. The HWE tests revealed that the loci Jcps20, Jcds66 and Jcms21 departed from the equilibrium and especially the locus Jcps20 was in disequilibrium in all the populations, indicating a reproductive pattern of non-random mating, and possibly that evolutionary forces that act upon some individuals of the populations exist (Hedrick, Reference Hedrick2011). Such forces include migration (caused by the germplasm interchange by farmers) and asexual reproduction, due to the propagation technique used by Mesoamerican Jatropha farmers. Other researchers raised similar ideas to explain the Hardy–Weinberg disequilibrium exhibited by J. curcas populations. For example, five microsatellite loci studied in the Asian accessions of J. curcas exhibited equilibrium deviations, due to the presence of null alleles or as a result of natural dispersal disturbances due to human activity (Basha and Sujatha, Reference Basha and Sujatha2007; Sudheer et al., Reference Sudheer, Pandya, Reddy and Radhakrishnan2008; Pamidimarri et al., Reference Pamidimarri, Sinha, Kothari and Reddy2009a).
The results of this study reveal that J. curcas in Chiapas has genetic diversity that is greater than that reported in other parts of the world, which represents a potential germplasm pool for the improvement of the plant or for the selection of genotypes for field planting.
Supplementary material
To view supplementary material for this article, please visit http://dx.doi.org/10.1017/S1479262114000550
