Introduction
The genus Vanilla Plum. ex Mill comprises 106 species (Soto Arenas and Cribb, Reference Soto Arenas and Cribb2010) distributed in the tropical regions of America, Asia, Africa and islands of the Indian and Pacific Oceans (Portéres, Reference Portéres and Lechevalier1954). However, only three species have economic value: V. planifolia, V. pompona and V. tahitensis. Of these, commercial vanilla (V. planifolia) occupies 95% of the market (Bory et al., Reference Bory, Grisoni, Duval and Besse2008a) and is used in food, perfumes and pharmaceuticals (Krushnamurthy et al., Reference Krushnamurthy, Shyamala and Madhava Naidu2013).
V. planifolia is native to southern Mexico and Central America, and southern Mexico has been identified as a possible domestication centre (Soto Arenas, Reference Soto Arenas1999; Lubinsky et al., Reference Lubinsky, Bory, Hernández, Seung-Chul and Gómez-Pompa2008a). In Mexico, this wild material is a species of low abundance, with only one individual per 2–10 km2 in well-conserved regions (Soto Arenas, Reference Soto Arenas1999). This material is extremely important as a source of genetic variability to broaden the genetic base within the commercial cultivars of V. planifolia. As revealed by RAPD (random amplified polymorphic DNA) and AFLP (amplified fragment length polymorphism) studies, the majority of the cultivated plantations in the world are derived from one to three genotypes of the Mexican morphotype Mansa (Schluter et al., Reference Schluter, Soto Arenas and Harris2007; Minoo et al., Reference Minoo, Jayakumar, Veena, Vimala, Basha, Saji, Nirmal Babu and Peter2008; Bory et al., Reference Bory, Lubinsky, Risterucci, Noyer, Grisoni, Duval and Besse2008b).
This lack of variability is also thought to be a primary contributor to the low tolerance to biotic and abiotic stresses, which have caused fruit drop and significant yield losses worldwide. The main problems are high sensitivity to drought and temperature changes and susceptibility to diseases caused by Fusarium, Phytophthora and Glomerella and by cymbidium mosaic virus (CymMV) (Duval et al., Reference Duval, Bory, Andrzejewski, Grisoni, Besse, Causse, Charon, Dron, Odoux and Wong2006; Minoo et al., Reference Minoo, Jayakumar, Veena, Vimala, Basha, Saji, Nirmal Babu and Peter2008; Bory et al., Reference Bory, Grisoni, Duval and Besse2008a; Hernández, Reference Hernández2011). To improve the tolerance of commercial cultivars and reduce losses, the wild species and their nearest relatives must be genetically characterized in depth to make full use of the germplasm.
Although the genetic variability of V. planifolia has been evaluated with different markers, inter-simple sequence repeats (ISSRs) have not been sufficiently explored. ISSR markers combine most of the advantages of microsatellites (SSRs) with those of AFLPs and the universality of RAPDs. The dominant ISSR markers are highly polymorphic, can track sequences found abundantly in the eukaryotic genome, and evolve quickly, yielding robust results (Zietkiewicz et al., Reference Zietkiewicz, Rafalski and Labuda1994; Fang and Roose, Reference Fang and Roose1997; Reddy et al., Reference Reddy, Sarla and Siddiq2002).
In a comparison between ISSR and RAPD markers to detect the genetic diversity between nine species of Vanilla (V. planifolia, V. tahitensis and seven wild Asian species), it has been found that the ISSR detected more polymorphisms (Verma et al., Reference Verma, Chakrabarty, Jena, Mishra, Singh, Sawant and Tuli2009). They have also been used to characterize the germplasm of various species such as Zingiber officinale (ginger) (Kizhakkayil and Sasikumar, Reference Kizhakkayil and Sasikumar2010) and Nelumbo spp. (lotus) (Zuo et al., Reference Zuo, Xiuqun, Wahiti, Juntawong, Zhou and Chen2010).
The Instituto de Biotecnología y Ecología Aplicada (INBIOTECA-Universidad Veracruzana) has collected 20 accessions of wild and cultivated germplasm of V. planifolia in southeastern Mexico. The objective of this study was to characterize this germplasm molecularly and morphologically to identify the sources of intra- and interspecific variation that can be used in breeding programmes and in the conservation of this valuable resource.
Materials and methods
Plant material
The collection sites were selected according to the previous information that indicated the presence of genetic variation in V. planifolia (Soto Arenas, Reference Soto Arenas1999). Also, we considered the information provided by producers about the presence of wild individual or different morphotypes of V. planifolia. Wild species of Vanilla were collected from the commercial cultivars of V. planifolia or at sites nearby.
In each site, we collected healthy cuttings of length 80 cm. The number of individuals collected in each site depended on the abundance and quality of plants. The cuttings were taken from different plants. Each plant was considered as an individual. Each individual or group of individuals of one species or morphotype from one site was labelled as an accession. Their passport data were registered and the cuttings were established under the greenhouse condition to integrate the genebank of INBIOTECA.
V. planifolia had both wild and cultivated accessions. The wild accessions were from Agua Blanca, Quintana Roo (QS) and an accession was from Armadillo Chico Oaxaca known as ‘Rajada’ because its fruits had a dehiscence more pronounced than other morphotypes of V. planifolia. The majority of the accessions of V. planifolia had the Mansa morphotype (Fig. 1). This is a commercial morphotype from which commercial plantations have been derived worldwide. Also, we collected the vanilla known as ‘Colibri’ from Oaxaca. Finally, we collected an uncommon commercial morphotype known as ‘Rayada’ or ‘Acamaya’, which had stems and leaves variegated with green and yellow stripes. It is similar in appearance to the genotype known as variegated in Reunion Island (Fig. 2).
Fig. 1. Collection sites. 
, Cultivated accessions; 
, wild accessions. The letters represent the names of the locations of the colleted sites: A, Armadillo Chico; B, La Bandera; E, Emiliano Zapata; F, Francisco Sarabia; J, Jalcomulco; JP, Joloapan; L, La Gran Lucha; M, Mesa de Guadalupe; O, Ojital Chico; P, Potingo; QC, Calderitas; QS, Agua Blanca; R, Rivera de Puntilla de Aldama; U, La Unión.
Fig. 2. Plant material. (a) V. pompona, (b) V. insignis, (c) V. affin. odorata, (d) V. planifolia (Mansa morphotype), (e) V. planifolia (Rayada morphotype), (f) V. planifolia (wild accessions from Quintana Roo) and (g) V. planifolia (wild accessions commonly known as Rajada).
We analysed 20 accessions that comprised 108 individuals of V. planifolia, one individual of V. pompona, three of V. insignis and one of V. affin. odorata from the INBIOTECA Genebank. These accessions were collected from 14 sites in the states of Veracruz, Oaxaca and Quintana Roo in Mexico (Table 1).
Table 1. Location and accession information and number of gene pools by site for individuals used in the molecular analysis
W, wild; VP, V. pompona; Ver., Veracruz; VIJ, V. insignis Jalcomulco; VIE, V. insignis Emiliano Zapata; VIA, V. insignis Armadillo Chico; Oax., Oaxaca; VOC, V. affin. odorata, Calderitas; QR, Quintana Roo; RJ, Rajada; QS, Agua Blanca; C, crop; RY, Rayada morphotype; RYP, Rayada Potingo; RYU, Rayada La Unión; Co, Colibrí; A, Armadillo Chico; L, La Gran Lucha; (M), Mansa morphotype; B, La Bandera; QC, Calderitas; E, Emiliano Zapata; F, Francisco Sarabia; J, Jalcomulco; JP, Joloapan; M, Mesa de Guadalupe; O, Ojital Chico; P, Potingo; R, Rivera de Puntilla de Aldama.
DNA extraction and generation of gene pools
DNA was extracted from young leaves following the method of Doyle and Doyle (Reference Doyle and Doyle1987). The DNA obtained was treated with 0.5 μl RNase (Invitrogen, Carlsbad, CA, USA) per 50 ml DNA suspension. The concentration and integrity of the DNA were verified by electrophoresis on 1.5% (w/v) agarose gels. As V. planifolia is a species that is established by cuttings, we can observe the presence of the same genotype in a plantation and even in several plantations. Therefore, we performed genetic pools to detect the genetic variation between the collection sites. For the molecular analysis of gene pools, the total number of individuals of a collection site was divided into two to three groups, each containing equal amounts of DNA from three to five individuals. The method of Gilbert et al. (Reference Gilbert, Lewis, Wilkinson and Caligari1999) was used to generate 33 gene pools (Table 1).
DNA amplification
The reaction mixture for PCR was prepared in a final volume of 20 μl, containing 30 ng genomic DNA, 15.2 μl ultrapure water, 2 μl of 1 × Taq buffer, 0.8 μl MgCl2 (50 mM), 0.4 μl dNTPs (10 mM), 0.4 μl primer (10 mM) and 0.2 μl of Taq polymerase (5 U/ml) (Promega, Madison, WI, USA). A negative control, without any genomic DNA, was run to rule out any self-amplification of primer or genomic DNA contamination. Six ISSR primers were used: five were designed by Verma et al. (Reference Verma, Chakrabarty, Jena, Mishra, Singh, Sawant and Tuli2009) and the primer (017) was designed by Martínez-Castillo (Reference Martínez-Castillo2005) (Table 2). Amplifications were performed in a GeneAmp PCR System 9700 thermal cycler (Applied Biosystems, Foster City, CA, USA) using the following programme: an initial denaturation step of 7 min at 94°C, 34 cycles of 30 s of denaturation at 94°C, 45 s alignment with temperatures between 47 and 54°C depending on the primer used, and 63 s extension at 72°C and a final extension of 10 min at 72°C.
Table 2. Results of molecular characterization using ISSR molecular markers, separated on polyacrylamide gels and silver-stained, at two levels [among four Vanilla species (V. planifolia, V. insignis, V. affin. odorata and V. pompona) and among V. planifolia accessions]
Separation and visualization of products
Each amplification product (5 μl) was mixed with 4 ml bromophenol blue and separated electrophoretically on 6% polyacrylamide gels for 13 h at 350 V in a vertical electrophoresis chamber (SQ3 sequencer; Hoefer Scientific Instruments, San Francisco, CA, USA). Bands were visualized using the silver staining technique (Martinez-Castillo et al., Reference Martínez-Castillo, Colunga-García Marín and Zizumbo-Villarreal2008).
Data analysis
For the four species, the clustering patterns and genetic relationships among accessions were represented by dendrograms constructed using the Jaccard similarity coefficient and the UPGMA (unweighted pair group method with arithmetic mean) algorithm. Their reliability was assessed by bootstrap analysis with 1000 replicates using the program Past v. 2.17 (Hammer et al., Reference Hammer, Harper and Ryan2001). Thereafter, an individual assignment test was performed using the program STRUCTURE v. 2.3.3 (Pritchard et al., Reference Pritchard, Stephens and Donnelly2000). STRUCTURE uses a clustering model based on a Bayesian approach to infer the number of K groups (i.e. populations or accessions) by means of an adjustment to the Hardy–Weinberg and linkage disequilibrium. The model used was that of ancestry admixture with correlated allele frequencies. Values of K from 1 to 10, with five independent simulations for each value of K, were evaluated. To obtain reliable data, the analyses were made with replications for a burn-in period of 10,000 and a run length of 100,000. With the results generated, the optimal value of K (number of genetically distinct groups or subgroups) was determined according to the method of Evanno et al. (Reference Evanno, Regnaut and Goudet2005) using the program Structure Harvester v. 0.6.93 (Earl and VonHoldt, Reference Earl and VonHoldt2012). Finally, the graphs of ancestry for the optimum value of K were generated using STRUCTURE.
For V. planifolia, the following indices of genetic diversity were estimated using the Popgene program v. 1.31 (Yeh and Boyle, Reference Yeh and Boyle1999): number of polymorphic loci; percentage of polymorphic loci; dominant nature of the ISSR; Shannon–Weaver diversity index. The analysis was conducted at three levels, within an accession, between groups and among all accessions, considering each accession as an independent population. Results of UPGMA and STRUCTURE were considered for determining the groups.
Results
The six primers generated 154 loci, of which 99.3% were polymorphic. Loci ranged from 201 to 3196 bp. An average of 25.6 bands was obtained per primer. Primers C03 and C09 detected more loci (Table 2). Within the group of V. planifolia, 88 loci were obtained, with 70.45% of these being polymorphic. The size of the sampled loci ranged from 420 to 2706 bp (Table 2).
The UPGMA (Fig. 3) showed three distinct groups for the four species of Vanilla evaluated. Group I represents V. pompona as a distant group separated from the other accessions. Group II comprised four accessions of the species V. insignis and V. affin. odorata. Group III comprised all accessions of V. planifolia, with three subgroups designated as IIIA, IIIB and IIIC. Subgroup IIIA comprised individuals of the Rayada morphotype (variegata) (RYP and RYU) and a wild individual of Oaxaca (RJ). Subgroup IIIB comprised wild individuals of Agua Blanca, Quintana Roo, and subgroup IIIC comprised all cultivated accessions of the Mansa morphotype from Veracruz and Quintana Roo, and accessions of Oaxaca (vanilla colibrí).
Fig. 3. Dendrogram generated from the Jaccard similarity coefficient. The numbers indicate the values generated with 1000 bootstrap repetitions. The asterisk (*) denotes the accessions of wild origin. Group I consisted of VP (V. pompona). Group II consisted of VIJ (V. insignis Jalcomulco), VIE (V. insignis Emiliano Zapata), VIA (V. insignis Armadillo Chico) and VOC (V. affin. odorata, Calderitas). Group III consisted of all the accessions of V. planifolia. Subgroup IIIA consisted of RYP (Rayada Potingo), RYU (Rayada La Unión) and RJ (Rajada Armadillo Chico). Subgroup IIIB consisted of QS (Agua Blanca, Quintana Roo). Subgroup IIIC consisted of A (Armadillo Chico), B (La Bandera), E (El Palmar), F (Francisco Sarabia), J (Jalcomulco), JP (Joloapan), L (La Gran Lucha), M (Mesa de Guadalupe), O (Ojital Chico), P (Potingo), R (Rivera de Puntilla de Aldama) and QC (Calderitas).
The method of Evanno et al. (Reference Evanno, Regnaut and Goudet2005) yielded an optimal value of K= 2, indicating the existence of two genetically distinct groups. Figure 4 shows the grouping pattern generated by STRUCTURE based on K= 2. Individuals of wild species (V. insignis, V. affin. odorata and V. pompona) belonged to one group and all accessions of V. planifolia belonged to the other group. Figure 4 also shows the existence of some accessions with some degree of ancestry from both groups. These are the wild accession from Oaxaca (RJ) (approximately 50% ancestry from the group of wild species and the group of V. planifolia) and individuals of the Rayada morphotype of V. planifolia (RYP and RYU).
Fig. 4. Analysis of assignment of individuals. The red group comprised the accessions of wild origin: VIJ, V. insignis Jalcomulco; VIE, V. insignis Emiliano Zapata; VIA, V. insignis Armadillo Chico; VOC, V. affin. odorata, Calderitas; VP, V. pompona Mesa de Guadalupe. The green group comprised the cultivated accessions of V. planifolia: A, Armadillo Chico; B, La Bandera; E, El Palmar; F, Francisco Sarabia; J, Jalcomulco; JP, Joloapan; L, La Gran Lucha; M, Mesa de Guadalupe; O, Ojital Chico; P, Potingo; QC, Calderitas; R, Rivera de Puntilla de Aldama. Mixed individuals are as follows: RJ, Rajada Armadillo Chico; RYP, Rayada Potingo, Rayada morphotype (variegata); RYU, Rayada La Unión, Rayada morphotype (variegata); QS, Agua Blanca; A, Armadillo Chico; O, Ojital Chico; R, Rivera de Puntilla de Aldama.
Genetic diversity in V. planifolia
The accession of V. planifolia with the greatest genetic diversity was Armadillo Chico (A) of Oaxaca, with a Shannon–Weaver diversity index of 0.0957 and 14.94% polymorphism, followed by Ojital Chico from Papantla Veracruz with a Shannon index of 0.0487 and 8.05% polymorphism at the group level. The greatest genetic diversity according to the Shannon index was found in group IIIA (R= 0.1536). Although group IIIC had a higher percentage of polymorphic loci (31.03%), group IIIB had less genetic diversity with a Shannon index of 0.0139. Considering all accessions, a Shannon index of 0.2143, with 71.26% of polymorphic loci, was obtained (Table 3).
Table 3. Indices of diversity among 16 accessions of V. planifolia collected from three states in Mexico
I, Shannon index; A, Armadillo Chico; B, La Bandera; E, Emiliano Zapata; F, Francisco Sarabia; J, Jalcomulco; JP, Joloapan; M, Mesa de Guadalupe; L, La Gran Lucha; O, Ojital Chico; P, Potingo; QC, Calderitas; QS, Agua Blanca; R, Rivera de Puntilla de Aldama; RYP, Rayada Potingo; RYU, Rayada La Unión; RJ, Rajada.
Discussion
Molecular characterization of all species
The use of the ISSR in conjunction with polyacrylamide gel electrophoresis is a technique that provided a significant increase in the mean number of bands. Verma et al. (Reference Verma, Chakrabarty, Jena, Mishra, Singh, Sawant and Tuli2009) used agarose gels, obtaining a mean of 13.4 bands with primers C03, C07, C09, T05 and T06. In our study using polyacrylamide gels and those five primers, the mean was 25.6. The products with primer T05 were the least polymorphic and increased by only three bands in the polyacylamide gels.
The four interspecific relationships, the four species of Vanilla could be differentiated despite their close relationship. The genetic relationships among the species closely related to V. planifolia have been studied with different molecular markers (AFLP, RAPD, ITS (internal transcribed spacer) and plastid trnH-psbA). Although results varied in previous reports, V. odorata and V. tahitensis appear to be the closest species to V. planifolia. V. tahitensis is an interspecific hybrid between V. planifolia and V. odorata (Soto Arenas, Reference Soto Arenas1999; Schluter et al., Reference Schluter, Soto Arenas and Harris2007; Lubinsky et al., Reference Lubinsky, Bory, Hernández, Seung-Chul and Gómez-Pompa2008a, Reference Lubinsky, Cameron, Molina, Wong, Lepers-Andrzejewski, Gómez-Pompa and Seung-Chulb; Bory et al., Reference Bory, Lubinsky, Risterucci, Noyer, Grisoni, Duval and Besse2008b).
Molecular characterization of V. planifolia
With respect to the genetic diversity of V. planifolia, our results indicate that the wild individual from Armadillo Chico, Oaxaca (RJ) was the most genetically differentiated accession. This accession is locally known as ‘Rajada’ because its fruits have a dehiscence more pronounced than other morphotypes of V. planifolia. Also, the fruit dehiscence starts when the fruit is not yet fully mature, so the producers consider that these fruits have no commercial value. It is important to mention that this individual was the only one present in the field crop where it was collected; the producers were exterminating it because such dehiscence was undesirable for commercialization. From the data generated in this study, we cannot determine whether this individual represents a new species of Vanilla or is a wild individual of V. planifolia.
Furthermore, the accessions of the morphotype Rayada (acamaya or variegata) are extremely important for breeding programmes and conserving diversity. In this study, we found the first clear molecular distinction between the Rayada and Mansa morphotypes. The Rayada accessions (RYP and RYU) had numerous unique loci, which were shared only with the wild accession of Oaxaca (Rajada RJ). Therefore, we can rule out the suggestion that the Rayada morphotype has arisen by somatic mutations from a Mansa morphotype as reported by other authors (Soto Arenas, Reference Soto Arenas1999; Bory et al., Reference Bory, Lubinsky, Risterucci, Noyer, Grisoni, Duval and Besse2008b). In addition, the genetic relationship of the Oaxaca wild individual may be an indication that the Rayada morphotype is the product of sexual reproduction between the Mansa morphotype and some wild material. However, a thorough investigation is required to classify the origin of the Rayada morphotype in Mexico. However, it is noteworthy that this pattern has not been observed in other studies, which could be due to the type of marker (isozyme, AFLP and RAPD) used in other studies or to different Rayada genotypes studied (Soto Arenas, Reference Soto Arenas1999; Schluter et al., Reference Schluter, Soto Arenas and Harris2007; Bory et al., Reference Bory, Lubinsky, Risterucci, Noyer, Grisoni, Duval and Besse2008b). Although these results are very interesting and novel, we should not rule out the possibility that epigenetic variation in Rayada vanilla (variegata) could be responsible for the change in leaf colour in response to light. We observed that a variegated plant can have green shoots when it is exposed to a higher intensity of light. Thus, a plant can present variegated shoots and completely green.
Another important source of genetic variation was observed in the wild material from Agua Blanca, Quintana Roo. This material has several unique loci but is more close to cultivated accessions that are of the Rayada morphotype and the wild individual from Oaxaca. We must take into account the hypothesis of Lubinsky et al. (Reference Lubinsky, Bory, Hernández, Seung-Chul and Gómez-Pompa2008a) that vanilla cultivation may have originated in the Maya region and subsequently was successfully adapted in the Veracruz region under the care of the Totonacos. However, we must also consider that our results may indicate that the material from Veracruz is more closely related to the material from Oaxaca. Other studies of genetic diversity in vanilla have suggested that individuals in the states of Oaxaca, Chiapas and Quintana Roo are genetically very similar (Soto Arenas, Reference Soto Arenas1999; Schluter et al., Reference Schluter, Soto Arenas and Harris2007). However, we found marked differences between the genotypes of Quintana Roo and Oaxaca, so a more detailed analysis of the diversity of vanilla in the Maya region and northern Oaxaca is needed.
The northern region of Oaxaca has been cited as harbouring great diversity in the genus Vanilla (Schluter et al., Reference Schluter, Soto Arenas and Harris2007; Lubinsky et al., Reference Lubinsky, Bory, Hernández, Seung-Chul and Gómez-Pompa2008a; Soto Arenas and Dressler, Reference Soto Arenas and Dressler2010). Oaxaca is the second largest producer of vanilla in Mexico. and therefore it has a greater influence on the forms of production in the Veracruz region. In this region, cuttings of V. planifolia from wild individuals (known as vanilla colibrí) are grown with commercial cuttings from Veracruz. In our results, we found that some accessions from Armadillo Chico are genetically distinct individuals, while others are genetically identical to the Veracruz accessions. Individuals from La Gran Lucha, Oaxaca, are genetically separated from other accessions, but are very close to the Veracruz material. However, this region of Oaxaca remains strategic as a source of diversity for the genus Vanilla, and studies on the population genetics of the material from this area are needed to help determine strategies for the use and conservation of wild genotypes.
For commercial individuals grown in the state of Veracruz, we note that the region of Nautla has been popular as a commercial source of cuttings, which is reflected by accessions from different backgrounds having the same genetic basis. We know that vanilla production at El Palmar (E) was established over 50 years ago in central Veracruz from five cuttings from San Rafael (near Nautla). However, we observed some differences between the gene pools of this accession that can be attributed to somatic mutations. Such mutations also explain the differences between accessions from Rivera de Puntilla de Aldama (R), La Bandera (B) and Francisco Sarabia (F). Variability in V. planifolia generated by somatic mutations has been reported by several authors for accessions from the Asian-Pacific region (Duval et al., Reference Duval, Bory, Andrzejewski, Grisoni, Besse, Causse, Charon, Dron, Odoux and Wong2006; Minoo et al., Reference Minoo, Jayakumar, Veena, Vimala, Basha, Saji, Nirmal Babu and Peter2008; Bory et al., Reference Bory, Lubinsky, Risterucci, Noyer, Grisoni, Duval and Besse2008b). The material grown in Calderitas Quintana Roo (QC) is identical genetically to the material grown in Joloapan Veracruz (JP). This site is very close to the town of Primero de Mayo from Papantla, Veracruz, which is also a popular source of cuttings.
Conclusions
The ISSR analysis used in conjunction with other markers allowed us to differentiate the species of Vanilla and provided a better detail to detect the variation among the accessions of V. planifolia. Thus, we consider this technique to be suitable for characterizing germplasm and to study the population genetics of Vanilla.
Some authors have shown the potential importance of using other species of Vanilla as sources to genetically improve V. planifolia, e.g. for breeding for tolerance against various pathogens. For example, V. andamanica (a distant species from V. planifolia) is tolerant to Fusarium oxysporum and is being used in hybridization work with V. planifolia (Minoo et al., Reference Minoo, Jayakumar, Veena, Vimala, Basha, Saji, Nirmal Babu and Peter2008). Duval et al. (Reference Duval, Bory, Andrzejewski, Grisoni, Besse, Causse, Charon, Dron, Odoux and Wong2006) mentioned V. pompona and V. tahitensis as species tolerant to CymMV. Other close wild relatives of V. planifolia (e.g. V. insignis) should also be explored for any evidence of tolerance to biotic and abiotic stresses. Promising materials should then be characterized and evaluated for use as parental lines for hybridization work.
Among the accessions of V. planifolia are at least four that have greater genetic variability that can be used to improve the genetic diversity in the crop. As discussed already, the most diverse materials proved to be a wild individual from Oaxaca (Rajada), two accessions of the Rayada morphotype (variegata) and a wild accession from Quintana Roo.
The individual with dehiscent fruits from Oaxaca and the accession from Quintana Roo should be analysed primarily for their tolerance to factors that negatively affect cultivated V. planifolia and to broaden the genetic base of the cultivated species.
Vanilla Rayada accessions (variegata) are of great interest for genetic improvement because of their variability and commercial value (similar to the Mansa morphotype). The Rayada morphotype is unfortunately rare in Mexico. Among the Asociación Veracruzana de Vainilleros are only four producers that have any individuals of this morphotype, and Rayada accessions are also present in the genebanks of INBIOTECA and Benemérita Universidad Autónoma de Puebla. In other vanilla-producing areas, such as the Asian-Pacific region, there are more records of variegated accessions. Accessions of the Rayada morphotype and those of wild origin should be given priority in conservation programmes because they are the most locally endangered ones due to the scarcity of individuals, destruction of habitat and replacement by the commercial morphotype. Together with breeding programmes in INBIOTECA, we work on micropropagation (Ramos-Castellá et al., Reference Ramos-Castellá, Iglesias-Andreu, Bello-Bello and Lee-Espinosa2014) and in vitro conservation (Bello-Bello et al., Reference Bello-Bello, García-García and Iglesias Andreu2015). This is intended to promote a comprehensive programme of conservation and breeding of this important resource.
Acknowledgements
The authors thank the Programa de Mejoramiento del Profesorado for the financial support of the project ‘Biotechnological Basis for the Genetic Improvement of Vanilla planifolia’ within the Conservation, Management and Plant Breeding network. They also thank the Laboratorio de Marcadores Moleculares from the Unidad de Recursos Naturales (CICY) where the ISSR data were generated.
