FIGURE 2.3 Factorial analysis from AFLP markers on different American Vanilla species illustrating the increased diversity for V. planifolia selfed progenies.
FIGURE 2.4 Exclusive sexually reproducing species. Left: V. palmarum in the CIRAD collection. Fruits were spontaneously obtained in an insect proof quarantine glasshouse. (Courtesy of M. Grisoni.) Right: V. mexicana in Guadeloupe. (Courtesy of P. Besse and P. Feldmann.)
Interspecific Hybridization
The main factors preventing interspecific hybridization in the Orchidaceae family are pre-pollination mechanisms such as pollinator specificity, flowering phenologies, or mechanical barriers in flowers (Dressler, 1981; Gill, 1989; Grant, 1994; Paulus and Gack, 1990; Van Der Pijl and Dodson, 1966). On the contrary, genetic incompatibility between closely related species is rarely observed (Dressler, 1993; Johansen, 1990; Sanford, 1964, 1967). This is also the case for Vanilla. Indeed, most inter specific artificial crosses attempted to date in Vanilla have been successful showing the lack of genetic incompatibility between the species involved. Interspecific hybrids were successfully obtained between closely related American species (V. planifolia × V. tahitensis J.W. Moore—accession Hy0003 in Figures 2.1 and 2.3, V. planifolia × V. pompona) in breeding programs in Madagascar (Bory et al., 2008c), and even between distantly related species such as the Indian V. aphylla Blume and the American V. planifolia in breeding programs in India (Minoo et al., 2006).
There is a growing evidence for the occurrence of natural interspecific hybridization in Vanilla. A study on three native species of Vanilla, V. claviculata, V. barbellata, and V. dilloniana in the western part of the island of Puerto Rico, showed the possibility of interspecific hybridization between V. claviculata and V. barbellata in sympatric areas (Nielsen, 2000; Nielsen and Siegismund, 1999). This was demonstrated by using isozyme markers, and floral morphological observations confirmed the hybrid status of sympatric populations. On the other hand, V. dilloniana, showing a different phenology, did not hybridize with the other two species. Recent work using AFLP and SSR markers also suggested the possibility of interspecific hybrid formation in tropical America, involving species such as V. bahiana, V. planifolia, or V. pompona (Bory, 2007; Bory et al., 2008d) (Figure 2.5). The species V. tahit-ensis was also recently shown using nuclear ITS and cp DNA sequences to result from intentional or inadvertent hybridization between the species V. planifolia and V. odorata C. Presl that could have happened during the Late Postclassic (1350– 1500) in Mesoamerica (Lubinsky et al., 2008b).
FIGURE 2.5 Flowers, fruits, and leaves from accession CR0068 (a) from Costa Rica, putatively identified as a hybrid species derived from a maternal V. planifolia donor species (b) based on AFLP, SSR, and cp DNA markers data. (Courtesy of M. Grisoni.)
Polyploidization
We recently demonstrated the occurrence of recent polyploidization events (in less than 200 years as V. planifolia was introduced in 1822) in Reunion Island (Bory et al., 2008a). Congruent evidences (AFLP markers, genome size, chromosome counts, and stomatal length) showed the formation of auto-triploid self-sterile types (“Stérile”), as well as auto-tetraploid types (“Grosse Vanille”), with genome sizes of 2C = 7.5 and 10 pg, respectively, as opposed to 5 pg for conventional “Classique” varieties (Bory et al., 2008a) (Table 2.1). The most likely formation of these types was suggested to be through manual self-pollination accompanied by the formation of unreduced 2n gametes (Bretagnolle and Thompson, 1995), seed germination from a forgotten pod, followed by vegetative multiplication of the individuals (Bory et al., 2008a) (Figure 2.6). Polyploidy was also reported for the cultivated species of V. tahitensis in Polynesia with the existence of diploid and tetraploid (i.e., “Haapape”) types (Duval et al., 2006; see Chapter 13) and this species might have resulted from both polyploidy and sexual regeneration following its V. planifolia × V. odorata origin (Lubinsky et al., 2008b).
FIGURE 2.6 Schematic representation of the possible formation of autotriploid and autotetra-ploid V. planifolia types in Reunion Island.
Polyploidization could therefore be a major phenomenon in the evolution of Vanilla. In order to put this hypothesis to test, we conducted a preliminary survey on genome size variation in different Vanilla species. Thirty-eight accessions were analyzed by flow cytometry according to the protocol detailed by Bory et al. (2008a) using wheat as an internal standard: Triticum aestivum L. cv. Chinese Spring, 2C = 30.9 pg, 43.7% GC (Marie and Brown, 1993). These accessions belong to 17 different Vanilla species and also included 3 artificial hybrids (V. planifolia × V. planifolia, V. planifo-lia × V. tahitensis, V. planifolia × V. phaeantha Rchb. f.). The entire leaf samples were collected from vines maintained in the Vanilla genetic resources collection of CIRAD in Reunion Island (see Chapter 3 and Grisoni et al., 2007). Details for each accession (species, putative continent of origin, place of sampling, and genome size) are presented in Table 2.1. For each species, fluorescence histograms revealed five endoreplicated peaks and the marginal replication ratio was still irregular (from 1.5 to 1.8 instead of 2), as encountered in V. planifolia (Bory et al., 2008a).