Genetic Resources and Crop Evolution

, Volume 55, Issue 4, pp 551–571

Biodiversity and preservation of vanilla: present state of knowledge

Authors

  • Séverine Bory
    • Pôle de Protection des PlantesCIRAD, UMR PVBMT Cirad – Université de la Réunion
    • Université de La Réunion, UMR PVBMT Cirad – Université de la Réunion
  • Michel Grisoni
    • Pôle de Protection des PlantesCIRAD, UMR PVBMT Cirad – Université de la Réunion
  • Marie-France Duval
    • CIRAD, UPR Multiplication Végétative
    • Université de La Réunion, UMR PVBMT Cirad – Université de la Réunion
Research Article

DOI: 10.1007/s10722-007-9260-3

Cite this article as:
Bory, S., Grisoni, M., Duval, M. et al. Genet Resour Crop Evol (2008) 55: 551. doi:10.1007/s10722-007-9260-3

Abstract

The genus Vanilla belongs to the Orchidaceae family and Vanilla planifolia, probably endemic from tropical forests in Eastern Mexico, is the main source for commercial vanilla. There has recently been an important number of publications covering Vanilla taxonomy, particularly using molecular genetics, but the taxonomy of the genus is still unclear and numerous synonyms remain. Recent studies showed that inter-specific hybridization and perhaps even polyploidization played an important role in the evolution of the genus. There has also been an important increase in the knowledge of the genetic diversity and reproductive biology of V. planifolia in natural conditions, showing that mating system diversity exists in Vanilla and that this genus could be a good model to study the role of fragrance in orchid evolution. Recent studies on the genetic consequences of V. planifolia domestication are also presented and raise major scientific questions regarding the origin of phenotypic diversity in a vegetatively propagated crop. Finally, all these studies have demonstrated the urgent need for preservation of the genetic resources of V. planifolia (primary and secondary gene pools, and cultivated resources) and current conservation efforts are presented.

Keywords

ConservationGenetic diversityMolecular geneticsOrchid taxonomyPlant reproductionVanilla planifolia

Introduction

The most comprehensive and recent review on vanilla history, taxonomy and ecology, is the book edited by Bouriquet (1954). We present here an updated bibliographical review of the current knowledge (including recent data obtained using molecular genetics) on the taxonomy, phylogeny, and the cultivation and dissemination history of vanilla (genus Vanilla Plumier ex Miller), particularly V. planifolia G. Jackson, syn. V. fragrans (Salisb.) Ames. The information is supplemented with a description of diversity and reproduction in natural conditions and in areas of introduction such as Reunion Island. Present worldwide efforts for the ex situ preservation of vanilla genetic resources are also reviewed. Some of the research perspectives that could be considered in these different areas are presented. This review shows that Vanilla can be a model of choice to study the consequences of crop domestication (through vegetative propagation) but also to unravel the early evolution history of the largest plant family: the Orchidaceae.

Vanilla taxonomy and phylogeny

Vanilla within the Orchidaceae

Distribution

The Orchidaceae family is very ancient and the genus Vanilla, which is monophyletic, supposedly belongs to a primitive lineage (Soto Arenas 1999b; Cameron 2003). Vanilla species are distributed throughout all continents, except Australia, between the 27th north and south parallels. Most (52) of the species are found in tropical America, 31 species are found in south-east Asia and New Guinea, 17 in Africa, 7 in the Indian Ocean islands, and 3 in the Pacific area (Portères 1954). From morphological observations of flowers, Portères drew hypotheses for the origin of vanilla and suggested that the primary diversification centre for the genus could be Indo-Malaysian. The Indo-Malaysian stock diversified and evolved on one hand in Madagascar, Mascarhenas islands and Africa, and on the other hand in oriental Asia and occidental Pacific islands. Asian species have evolved with a migration towards the Pacific and from there, either directly towards America either (more doubtfully) indirectly towards continental Asia and Europe during the Tertiary (65.5–2.5 mya) (Portères 1954). However, other more recent historical bio-geographical studies based on phylogeny data suggest a different scenario (Cameron 1999, 2000). Even if Vanilla is a pantropical genus, species from South America are sister to those from Africa and Asia. Vanillinae lineage evolved prior to the breakup of Gondwana (160 mya) in South America. Then, a migration to the Old World before 100 mya occurred, explained more by series of vicariance events than by long-distance dispersal. Consequently, Orchidaceae may have evolved much earlier than is traditionally believed. We favour an American origin for the Vanilla genus because recent bio-geographical data are based on molecular phylogeny and consider the Orchidaceae family as a whole, and because a trans-pacific migration is unlikely (Portères 1951a).

According to Portères (1954), 18 species are aromatic but Soto Arenas (2003) recognizes 35 known or expected aromatic species which are mostly of American origin. V. planifolia is the main cultivated aromatic species and 95% of the world commercial vanilla originates from it. V. planifolia is probably endemic from eastern Mexico tropical forests and its natural habitat roughly follows a straight line between the Oaxaca state towards Guatemala and Belize (Fig. 1) (Soto Arenas 1999a). Its population density is very low, with one specimen per 2–10 km2 in Chimalapas (Fig. 1) and one specimen per 4 km2 in Oaxaca (Fig. 1) (Soto Arenas 1999b). Another Vanilla species, V. tahitensis J.W. Moore is also cultivated in several Pacific countries. Some other aromatic species are grown locally or harvested in the wild but have no economical importance, for example, V. pompona Schiede in the West Indies, V. chamissonis Klotzsch in Brazil, V. odorata C. Presl in America, V. claviculata (W. Wright) Sw., V. griffithii Rchb. f. and V. abundiflora J.J.Sm. in the West Indies and in Asia (Soto Arenas 2003).
https://static-content.springer.com/image/art%3A10.1007%2Fs10722-007-9260-3/MediaObjects/10722_2007_9260_Fig1_HTML.gif
Fig. 1

Geographical localization of Mexican and Meso-American places cited in this paper

Taxonomy and phylogeny

The Vanilla genus belongs to the Orchidaceae family which contains more than 800 genera distributed in more than 25,000 species (Govaerts et al. 2006). Floral characters, such as anther configuration and pollinarium structure, have been used traditionally to classify the family ((Dressler and Dodson 1960), in: (Cameron 2004)) but have been hypothesized to be subjected to selective pressure (Cameron et al. 1999). Cameron et al. (1999) cite the most recent phenetic treatment of Orchidaceae to be that of Dressler (1993). This classification is the most comprehensive system at the moment and is widely accepted by botanists and growers.

Molecular phylogenetic studies have changed substantially how Vanilla and its allies are now treated by orchidologists. Plastid DNA was used to carry out phylogenetic studies within the Orchidaceae family (Fay and Krauss 2003; Cameron 2004). The Vanilloideae sub-family has been further studied by Cameron et al. (Cameron 1996, 2004, 2005b; Cameron and Chase 1998, 1999; Cameron and Dickison 1998; Cameron et al. 1999). Sequence data for the rbcL plastid gene divided the Orchidaceae family into five major monophyletic clades: Apostasioideae, Cypripedioideae, Vanilloideae, Orchidoideae and Epidendroideae. These clades correspond to the classically recognized sub-families except for the Vanilloideae, which is now considered a new sub-family (Cameron et al. 1999). Sequence data for the psaB gene showed the same division of the family into five monophyletic clades but failed to resolve the positions of the Cypripedioideae and the Vanilloideae sub-families (Cameron 2004). The Vanilloideae sub-family is composed of two sub-tribes: Pogonieae and Vanillineae (Cameron 1996; Cameron and Chase 1999). Other gene sequences were used to complete phylogenetic relationships in orchids: atpB (Cameron 2005b), psbB/psbC particularly for Vanilla (Cameron and Molina 2006), where eight species of Vanilla were analyzed; matK gene and intergenic non-coding spacer trnL-F for Apostasioideae (Kocyan et al. 2004) and Diurideae (Kores et al. 2001), matK and rbcL for Arethuseae (Goldman et al. 2001), where only one accession of V. planifolia was analyzed. All the sequences are available in GenBank. These studies place the Vanilla genus in the Orchidaceae family, Vanilloidae sub-family, Vanilleae tribe and Vanillinae sub-tribe (Cameron 2005b).

The Vanilla genus

The genus organization

The number of species given for the Vanilla genus varies depending on the authors: 90 (Cameron and Chase 1999), 107 (Soto Arenas 2003), 110 (Portères 1954). The International Plant Names Index is a base which lists 188 names of Vanilla species (The International Plant Names Index 2004). In the list of the world’s orchids compiled by Govaerts (2006) from different authors’ reviews, 110 species names are accepted while 71 are not. The main goal of this list was to provide baseline on accepted data in orchid families rather than revise families. Acceptance of a species name or an infra-specific taxon is based not only on literature and common usage but also, when possible, on specialist advice and on herbarium or living collections. Soto Arenas (1999b, 2006b) cites six additional species, which do not appear in the three previous lists. All Vanilla species cited in the literature are compiled in Table 1.
Table 1

List of Vanilla species found in the literature

Genus

Species

Author

Source

Vanilla

abundiflora

J. J. Sm.

a

b1

 

d

Vanilla

acuminata

Rolfe

a

b1

 

d

Vanilla

acuta

Rolfe

a

b1

 

d

Vanilla

acutifolia

Lodd. Cat. ex W. Baxt.

   

d

Vanilla

africana

Lindl.

a

b1

 

d

Vanilla

africana subsp. cucullata

(Kraenzl. ex J. Braun et K. Schum.) Szlach. et Olszewski

 

b2

 

d

Vanilla

africana subsp. ramosa

(Rolfe) Szlach. et Olszewski

 

b2

 

d

Vanilla

albida

Blume

a

b1

 

d

Vanilla

anaromatica

Griseb.

 

b2

 

d

Vanilla

andamanica

Rolfe

a

b1

 

d

Vanilla

angustifolia

Willd.

 

b2

 

d

Vanilla

angustipetala

Schltr.

a

b1

 

d

Vanilla

annamica

Gagnep.

a

b1

 

d

Vanilla

anomala

Ames et L.O.Williams

a

b2

 

d

Vanilla

aphylla

Blume

a

b1

 

d

Vanilla

aphylla

Eggers

   

d

Vanilla

aphylla

Wight

   

d

Vanilla

appendiculata

Rolfe

a

b1

 

d

Vanilla

argentina

Hicken

a

b2

 

d

Vanilla

aromatica

Willd.

 

b2

 

d

Vanilla

aromatica

Sw.

 

b2

 

d

Vanilla

articulata

Northr.

 

b2

 

d

Vanilla

axillaris

Mill.

 

b2

 

d

Vanilla

bahiana

Hoehne

 

b1

 

d

Vanilla

bakeri

Schltr.

a

b1

 

d

Vanilla

bampsiana

Geerinck

 

b1

 

d

Vanilla

barbellata

Rchb. f.

a

b1

 

d

Vanilla

barrereana

Y.Veyret et Szlach.

 

b1

 

d

Vanilla

beauchenei

A.Chevalier

 

b2

 

d

Vanilla

bertoniensis

Bertoni

a

b1

 

d

Vanilla

bicolor

Lindl.

a

b1

 

d

Vanilla

borneensis

Rolfe

a

b1

 

d

Vanilla

bradei

Schltr. ex Mansf.

a

b1

 

d

Vanilla

calopogon

Rchb. f.

a

b1

 

d

Vanilla

calyculata

Schltr.

a

b1

 

d

Vanilla

carinata

Rolfe

a

b1

 

d

Vanilla

chalottii

A. Finet

a

b1

 

d

Vanilla

chamissonis

Klotzsch

a

b1

 

d

Vanilla

chamissonis var. brevifolia

Cogn.

 

b2

 

d

Vanilla

claviculata

Lindl.

 

b2

 

d

Vanilla

claviculata

Sw.

a

b1

 

d

Vanilla

columbiana

Rolfe

a

b1

 

d

Vanilla

correllii

R. P.Sauleda et R. M. Adams

 

b1

 

d

Vanilla

coursii

H. Perrier

a

b1

 

d

Vanilla

crenulata

Rolfe

a

b2

 

d

Vanilla

cribbiana

Soto Arenas

  

c1

 

Vanilla

cristagalli

Hoehne

 

b1

 

d

Vanilla

cristatocallosa

Hoehne

 

b1

 

d

Vanilla

cucullata

Kraenzl. ex J. Braun et K. Schum.

a

b1

 

d

Vanilla

decaryana

H. Perrier

a

b1

 

d

Vanilla

denticulata

G. F. J. Pabst

 

b1

 

d

Vanilla

diabolica

P. O’Byrne

 

b1

 

d

Vanilla

dietschiana

Edwall

a

b2

 

d

Vanilla

dilloniana

Correll

a

b1

 

d

Vanilla

domestica

(L.) Druce

 

b2

 

d

Vanilla

dominiana

Hort. ex Gentil

   

d

Vanilla

dressleri

Byrd

  

c2

 

Vanilla

dubia

Hoehne

 

b1

 

d

Vanilla

duckei

Huber

a

b2

 

d

Vanilla

dungsii

Pabst

 

b1

 

d

Vanilla

edwallii

Hoehne

 

b1

 

d

Vanilla

eggersii

Rolfe

a

b2

 

d

Vanilla

ensifolia

Rolfe

a

b2

 

d

Vanilla

epidendrum

Mirb.

 

b2

 

d

Vanilla

fasciola

(G. Forst.) Spreng.

 

b2

 

d

Vanilla

fimbriata

Rolfe

a

b1

 

d

Vanilla

fragrans

(Salisb.) Ames

a

b2

 

d

Vanilla

françoisii

H. Perrier

a

b1

 

d

Vanilla

gardneri

Rolfe

a

b1

 

d

Vanilla

giulianettii

F.M.Bailey

a

b1

 

d

Vanilla

grandiflora

Lindl.

 

b1

 

d

Vanilla

grandifolia

Lindl.

a

b1

 

d

Vanilla

grandifolia var. lujae

(De Wild.) Geerinck

 

b2

 

d

Vanilla

gratiosa

Griseb.

 

b2

 

d

Vanilla

griffithii

Rchb. f.

a

b1

 

d

Vanilla

griffithii var. formosana

T. Itô

 

b2

  

Vanilla

griffithii var. ronoensis

(Hayata) S. S. Ying

 

b2

 

d

Vanilla

guianensis

Splitg.

 

b2

 

d

Vanilla

hallei

Szlach. et Olszewski

 

b1

 

d

Vanilla

hamata

Klotzsch

a

b1

 

d

Vanilla

hameri

Soto Arenas

  

c1

 

Vanilla

hartii

Rolfe

a

b1

 

d

Vanilla

havilandii

Rolfe

a

b1

 

d

Vanilla

helleri

A.D.Hawkes

 

b1

 

d

Vanilla

heterolopha

Summerh.

a

b1

 

d

Vanilla

hirsuta

M. A. Clem. et D. L. Jones

 

b2

 

d

Vanilla

hostmanni

Rolfe

a

b1

 

d

Vanilla

humblotii

Rchb. f.

a

b1

 

d

Vanilla

imperialis

Kraenzl.

a

b1

 

d

Vanilla

imperialis var. congolensis

De Wild.

 

b2

  

Vanilla

inodora

Schiede

a

b1

 

d

Vanilla

insignis

Ames

a

b1

 

d

Vanilla

kaniensis

Schltr.

a

b1

 

d

Vanilla

kempteriana

Schltr.

a

b1

 

d

Vanilla

kinabaluensis

Carr

a

b1

 

d

Vanilla

klabatensis

Schltr.

a

   

Vanilla

latisegmenta

Ames et C. Schweinf.

a

b1

 

d

Vanilla

laurentiana

De Wild.

 

b2

 

d

Vanilla

laurentiana var. gilletii

De Wild.

 

b2

  

Vanilla

leprieurii

R. Porteres

a

b1

 

d

Vanilla

lindmaniana

Kraenzl.

a

b1

 

d

Vanilla

lujae

De Wild.

a

b2

 

d

Vanilla

lutea

C. Wright ex Griseb.

 

b2

 

d

Vanilla

lutescens

Moq. ex Dupuis

 

b2

 

d

Vanilla

madagascariensis

Rolfe

a

b1

 

d

Vanilla

majaijensis

Blanco

 

b2

 

d

Vanilla

marowynensis

Pulle

a

b1

 

d

Vanilla

martinezii

?

  

c2

 

Vanilla

methonica

Rchb. f. et Warszc.

a

b1

 

d

Vanilla

mexicana

Mill.

 

b1

 

d

Vanilla

microcarpa

H. Karst.

 

b2

 

d

Vanilla

montagnacii

R. Porteres

a

  

d

Vanilla

montana

Ridl.

a

b2

 

d

Vanilla

moonii

Thwaites

a

b1

 

d

Vanilla

nigerica

Rendle

a

b1

 

d

Vanilla

ochyrae

Szlach. et Olszewski

 

b1

 

d

Vanilla

odorata

C. Presl

a

b1

 

d

Vanilla

organensis

Rolfe

a

b1

 

d

Vanilla

oroana

Dodson

 

b2

 

d

Vanilla

ovalifolia

Rolfe

 

b2

 

d

Vanilla

ovalis

Blanco

a

b1

 

d

Vanilla

ovata

Rolfe

a

b1

 

d

Vanilla

palembanica

Teysm. et Binn.

a

b1

 

d

Vanilla

palmarum

(Salzm. ex Lindl.) Lindl.

a

b1

 

d

Vanilla

palmarum var. grandifolia

Cogn.

 

b2

 

d

Vanilla

parishii

Rchb. f.

a

b2

 

d

Vanilla

parvifolia

Barb. Rodr.

a

b1

 

d

Vanilla

pauciflora

Dressler

 

b2

 

d

Vanilla

penicillata

Garay et Dunst.

 

b1

 

d

Vanilla

perexilis

Bertoni

a

b1

 

d

Vanilla

perplexa

Soto Arenas

  

c1

 

Vanilla

perrieri

Schltr.

a

b1

 

d

Vanilla

pfaviana

Rchb. f.

a

b2

 

d

Vanilla

phaeantha

Rchb. f.

a

b1

 

d

Vanilla

phalaenopsis

Reichb. f. ex Van Houtte

a

b1

 

d

Vanilla

philippinensis

Rolfe

 

b2

 

d

Vanilla

pierrei

Gagnep.

a

b1

 

d

Vanilla

pilifera

R. E. Holttum

a

b1

 

d

Vanilla

pittierii

Schltr.

a

b2

 

d

Vanilla

planifolia

Jacks. ex Andrews

 

b1

 

d

Vanilla

planifolia

Griseb.

   

d

Vanilla

planifolia var. gigantea

Hoehne

 

b2

 

d

Vanilla

planifolia var. macrantha

M. Gomez

 

b2

 

d

Vanilla

planifolia var. macrantha

Griseb.

   

d

Vanilla

platinilla

Hort. ex Gentil

   

d

Vanilla

platyphylla

Schltr.

a

b1

 

d

Vanilla

pleei

Porteres

a

b2

 

d

Vanilla

poitaei

Rchb. f.

a

b1

 

d

Vanilla

polylepis

Summerh.

a

b1

 

d

Vanilla

pompona

Schiede

a

b1

 

d

Vanilla

ponapensis

Kanchira et Yamamoto

a

b2

 

d

Vanilla

porteresiana

Szlach. et Y. Veyret

 

b1

 

d

Vanilla

preussii

Kraenzl.

a

b2

 

d

Vanilla

pseudopompona

Soto Arenas

  

c1

 

Vanilla

pterosperma

Lindl. ex Wall.

 

b2

 

d

Vanilla

purusara

Barb. Rodr. ex Hoehne

 

b1

 

d

Vanilla

ramificans

J. J. Sm.

a

   

Vanilla

ramosa

J. J. Sm.

 

b2

 

d

Vanilla

ramosa

Rolfe

a

b1

 

d

Vanilla

ribeiroi

Hoehne

a

b1

 

d

Vanilla

rojasiana

Hoehne

 

b1

 

d

Vanilla

ronoensis

Hayata

a

b2

 

d

Vanilla

roscheri

Rchb. f.

a

b1

 

d

Vanilla

rubiginosa

Griff.

 

b2

 

d

Vanilla

rubra

(Lam.) Urb.

 

b2

 

d

Vanilla

ruiziana

Klotzsch

a

b1

 

d

Vanilla

sativa

Schiede

 

b2

 

d

Vanilla

savannarum

Britton

 

b2

 

d

Vanilla

schliebenii

Mansf. ex Schlieben

 

b2

 

d

Vanilla

schwackeana

Hoehne

 

b1

 

d

Vanilla

seranica

J. J. Sm.

a

b1

 

d

Vanilla

seretii

De Wild.

a

b1

 

d

Vanilla

siamensis

Rolfe ex Downie

a

b1

 

d

Vanilla

somai

Hayata

a

b2

 

d

Vanilla

speciosa

Bovall ex Naves

   

d

Vanilla

sprucei

Rolfe

a

b1

 

d

Vanilla

sumatrana

J. J. Sm.

a

b1

 

d

Vanilla

surinamensis

Rchb. f.

 

b1

 

d

Vanilla

sylvestris

Schiede

 

b2

 

d

Vanilla

tahitensis

J. W. Moore

a

b2

 

d

Vanilla

taiwaniana

S. S. Ying

 

b2

 

d

Vanilla

tiarei

Costantin et Bois

 

b2

 

d

Vanilla

tisserantii

R. Porteres

a

b2

 

d

Vanilla

tolypephora

Ridl.

 

b2

 

d

Vanilla

trigonocarpa

Hoehne

 

b1

 

d

Vanilla

uncinata

Huber ex Hoehne

 

b1

 

d

Vanilla

utteridgei

J. J. Wood

 

b1

 

d

Vanilla

vanilla

(L.) H. Karst.

   

d

Vanilla

vanilla

Huth

 

b2

 

d

Vanilla

vellozii

Rolfe

a

b2

 

d

Vanilla

verrucosa

Hauman

a

b1

 

d

Vanilla

viridiflora

Blume

 

b2

 

d

Vanilla

walkeriae

Wight

a

b1

 

d

Vanilla

wariensis

Schltr.

a

b1

 

d

Vanilla

weberbaueriana

Kraenzl.

a

b1

 

d

Vanilla

wightiana

Lindl.

a

   

Vanilla

wightii

Lindl. ex Wight

 

b1

 

d

Vanilla

wrightii

Rchb. f.

a

b2

 

d

Vanilla

yersiniana

Guillaumin et Sigaldi

 

b2

 

d

Vanilla

zanzibarica

Rolfe

a

b2

 

d

a: species cited by Portères (1954); b1: accepted names of species published in the World Checklist of Orchidaceae of Royal Botanic Gardens, Kew (Govaerts et al. 2006); b2: not accepted names of species published in the World Checklist of Orchidaceae of Royal Botanic Gardens, Kew (Govaerts et al. 2006); c1: species cited by Soto Arenas (1999b); c2: species cited by Soto Arenas (2006b); d: species published on the Internet in The International Plant Names Index (2004)

Rolfe in 1896 first proposed a taxonomic classification for Vanilla species with two sections: Foliosae, where vines have developed leaves and Aphyllae, where plants have bract-like leaves (Portères 1954). Portères (1954) further divided the Foliosae section into three sub-sections: Papilloseae have thick leaves and labellum with more or less fleshy hairs, Lamellosae have also thick leaves but labellum wear scaly lamellae and Membranaceae have thin membraneous to sub-membraneous leaves. Soto Arenas (2003) invalidated this classification by stressing that sub-sections are heterogeneous and incomplete. However, as no revision of the taxonomy has been proposed to date, Portères’ classification is still used.

Phylogenetic relationships among Vanilla species

Vegetative anatomical characters have been used to generate a cladistic analysis of 14 Vanilla species (Stern and Judd 1999). Although results were partially in agreement with Portères’ classification, the sampling was too small to conclude further.

Nuclear sequence data for the Internal Transcribed Spacer (ITS) of the 18S and 26S rRNA genes, were used to unravel relationships within sub-tribes and genera in the Orchidaceae (Cameron 2004). In order to establish phylogenetic relationships between Mexican Vanilla species, Soto Arenas (1999b) studied the same sequences. The ten species of Mexican vanilla included in the study belonged to a monophyletic group, divided into three clades, which is in agreement with Portères classification. The most external clade contained V. inodora Schiede of the Foliosae section and Membranaceae sub-section. The second clade contained V. barbellata Rchb. f. of the Aphyllae section, and the third clade contained the eight remaining species, which are aromatic and belong to the Foliosae section and Lamellosae sub-section. In the third clade, two groups could be distinguished based on morphology. The group containing V. pompona included V. cribbiana Soto Arenas and V. hameri Soto Arenas, and shared characteristic trigonal fruits and large, fragrant yellow flowers without papilla. The group containing V. planifolia included V. insignis Ames, V. odorata and V. phaeantha Rchb.f., which were characterized by less trigonal, smaller fruits and greenish flowers. At the intra-specific level, the ITS data showed little variation, which might reflect methodological (alignment or sequencing) artefacts rather than a genuine variation.

Analysis of the plastid gene rbcL sequences of 21 Vanilla species had showed congruent results with sequences from the other analyzed gene regions (ITS, matK) as well as with morphology (Soto Arenas 2003), with the presence of an external clade that includes species of the Foliosae section and Membranaceae sub-section such as V. inodora, V. mexicana Mill., V. angustipetala Schltr. and V. martinezii, which are confined to the Neotropics. The remaining species are separated in two main clades, one including the Old World and Caribbean species and the other the American fragrant species (Soto Arenas 2003).

The origin of V. tahitensis

Vanilla tahitensis is the second most cultivated aromatic vanilla species worldwide. Historically, it was introduced into Tahiti from the Philippines by Amiral Hamelin in 1848 (Correll 1953; Portères 1954). However, the genetic identity of this species is unclear. Portères (1954) cites various authors giving each a different name: V. ovalis Blanco (1845) = V. planifolia Andrews (by Naves in 1879) = V. majaijensis Blanco (by Rolfe in 1896) = V. aromatica S.W. (by Naves in 1879) and V. ovalis Blanco = V. philippinensis Rolfe (1896).

Some authors even suggested that V. tahitensis results from inter-specific hybridization between V. planifolia and a second parent, either a specimen of a V. pompona–V. odorata complex (Portères 1951b; Soto Arenas 1999b) or V. pompona (Portères 1954). This was based upon morphological traits shared either with V. planifolia (leaf, inflorescence and flower morphologies and the general aspect of the fruit (Portères 1954)), with V. pompona (the reflection of the labellum disc, the presence of heliotrope (piperonal) and a low vanillin content in the pods, and indehiscence of the mature pod), or with V. odorata (the stem anatomy (P. Roux, in: (Portères 1951b)) and the fruit anatomy (R. Simony, in: (Portères 1953)), the capacity to flower less seasonally and in equatorial areas such as New Guinea (Soto Arenas 1999b)). Moreover, the leaf dimensions of V. tahitensis are exactly intermediate between that of V. planifolia and V. odorata (Soto Arenas 2006b).

Molecular markers were used to elucidate V. tahitensis origin. In a first Random Amplified Polymorphic DNA study (RAPD), 93% of all detected markers for V. tahitensis were shared with V. planifolia and none of the V. pompona specific fragments was shared with V. tahitensis. The authors favoured the hypothesis of a close relation between V. tahitensis and V. planifolia (Besse et al. 2004). In another study carried out with RAPD markers, Schlüter did not find good evidence in support of an hybrid origin of V. tahitensis (all combination between three putative parents V. planifolia, V. odorata, and even V. insignis were tested) (Schlüter 2002).

In two other studies, the Amplified Fragment Length Polymorphism technique (AFLP) confirmed that V. planifolia and V.tahitensis are genetically very closely related. The genetic distance calculated between V. tahitensis and V. pompona was approximately three times greater than between V. tahitensis and V. planifolia. In addition, very few bands present in both V. tahitensis and V. pompona were found (Bory 2004; Duval et al. 2006).

So far, the molecular results did not validate the suggested inter-specific hybrid origin of V. tahitensis and favoured the hypothesis of a variation of V. planifolia. Nevertheless the use of co-dominant markers is needed for confirmation, such as nuclear Simple Sequence Repeats (SSR). These were developed in orchids (Gustafsson 2000; Soliva et al. 2000; Fay and Krauss 2003) as well as recently for V. planifolia (Bory et al., unpublished data).

The taxonomy of the Vanilla genus continues to be puzzled. Remaining questions on taxonomy and phylogeny are numerous, such as the positions of the Cypripedioideae and the Vanilloideae sub-families in the Orchidaceae family, the position of Vanilla and the sister group within Vanilloideae, the number of Vanilla species and the numerous synonymy and the phylogeny and taxonomical position of V. tahitensis. Sequence data for the plastid and nuclear gene introns, as performed by Cameron et al. (Cameron et al. 1999; Cameron 2004; Cameron and Molina 2006) and Soto Arenas (1999b), will be very useful to revise the taxonomy of the Vanilla genus.

Reproductive biology

The reproduction of American vanilla in natural conditions

Sexual reproduction is rarely observed in natural conditions. Natural reproduction of V. planifolia in Puerto Rico was observed for less than 1% of the flowers (Childers and Cibes 1948). Weiss (2002) reported similar rates (between 1 and 3%) in Central America. Soto Arenas (1999b) reported an even lower rate in Mexico (1 fruit for 100–1,000 flowers).

The floral structure of vanilla makes natural self-pollination difficult. The rostellum prevents contact between stamen and stigmata except for some rare species like V. palmarum and V. inodora, which are self-pollinating (Soto Arenas 2006b). Nevertheless, spontaneous self-pollination rates of 6.06% are reported for V. chamissonis, in the São Paulo area (Brazil) (Macedo Reis 2000). Likewise, some V. planifolia cultivars in Mexico show natural self-pollination rates from 4–6% to 20% (Soto Arenas 1999b). Similarly, in 1997, Hernandez-Apolinar conducted experiments on reproduction costs associated with cultivated V. planifolia and found that 6% of the wild individuals used to establish crops in Oaxaca by the end of the 1980’s were self-pollinating (Lubinsky 2004).

The floral morphology of vanilla favours out-crossing. In V. chamissonis, 15.16% of natural allogamy is reported (Macedo Reis 2000). Similarly, in Mexico, most V. planifolia varieties, like ‘Mansa’, are unable to naturally self-pollinate (Soto Arenas 1999b). Some self-incompatible V. planifolia varieties are even reported, like ‘Oreja de Burro’, for which 80–100% of fruit aborts three months after self-pollination (Castillo Martinez and Engleman 1993). Dequaire (1976) observed that successive self-fertilization of V. planifolia led to inbreeding depression with plants being less vigorous and more sensitive to diseases. All these observations are in favour of an out-crossing reproductive mode for V. planifolia. Surprisingly, Soto Arenas (1999b) found very low allogamy rates (using isozyme markers) in the Veracruz (0.106) and Oaxaca regions (0.082), and very low observed heterozygosity. He suggested that the dominant sexual reproductive mode in these regions was autogamy, and this was explained by the possible occurrence of a majority of self-compatible individuals in these regions (Soto Arenas 1999b). The higher allogamy rate observed in Veracruz was supposedly due to the presence of the self-incompatible ‘Oreja de Burro’ type (Soto Arenas 1999b). It was concluded that V. planifolia possesses a mixed reproductive system in which the real proportion of self-compatible and self-incompatible individuals is still unknown (Soto Arenas 1999b). It seems as well essential to assess the relative success rates of self-pollination versus allo-pollination in V. planifolia.

In America, pollinators are not well known. According to some authors (Lecomte 1901; Ridley 1912; Bouriquet 1954; Purseglove et al. 1981; Torregrossa 1988) pollinators are rare and only exist in South and Central America. According to these authors, V. planifolia is pollinated by social bees of the Melipona genus and by hummingbirds. Other pollinators have been reported, like bees of the genus Trigona suspected to carry out pollination in Guadeloupe (Stehlé 1952). Soto Arenas (1999b) reported the existence of three pollination systems among Mexican vanilla species. The first system, restricted to V. inodora, involves carpenter bees of the Xylocopa genus. The second system, specific to V. pompona, V. hameri and V. cribbiana, involves bees from the Euglossa genus, in which males collect flower nectar. The third system concerns V. planifolia, V. odorata and V. insignis and is a deceptive pollination system, where flowers do not reward the insects and are visited equally by males and females. The latter is commonly observed in orchids with low population densities (Ackerman 1986). According to Soto Arenas (2006b), Euglossa viridisima and maybe Eulaema spp. are the real pollinators of V. planifolia. Lubinsky et al. (2006) have carried out field observations in Oaxaca of V. planifolia for 2 weeks of the spring of 2004. They have noted occasional visit of flowers by ants, Melipona, Euglossa, Exeretes and hummingbirds without occurrence of pollination event. They have observed on V. grandiflora Lindl., in the Peruvian Amazon in September 2005, visit of flowers by Melipona and Euglossa without pollen removal due to their too small size, but pollination by Eulaema meriana Olivier is suspected following observation of successful pollen removal by this orchid bee species (Lubinsky et al. 2006).

Like for other orchids, it was suggested that V. planifolia seeds could be dispersed by air or water (Arditti and Ghani 2000). Soto Arenas (1999b, 2003) suggested that aromatic Vanilla fruits (a character almost exclusively restricted to American species) are an adaptation to bat dispersal. It was even proposed that V. planifolia seeds could be endo-ornithochors with bird digestive fluids helping germination (Bouriquet 1954; Arditti and Ghani 2000). Recently in V. grandiflora, fragrance collection by orchid bees from fruits was also observed (Lubinsky et al. 2006) which may lead to dispersion of sticky seeds. This raises important questions regarding the early evolution history of the orchid family which might have involved a transition away from animal-mediated dispersal (Lubinsky et al. 2006).

Seeds produced by vanilla rarely germinate. They have an undifferentiated embryo, little reserve matter, very hard and waxy teguments containing germination inhibitors (Kleinert, 1963, in: (Dequaire 1976)). Childers et al. (1959) noted that a small number of seeds could germinate in ideal conditions of humidity, temperature, and nutrition. At the Federal Experiment Station of Mayaguez in Puerto Rico, they observed seeds germinating and growing on moist, rotten wood support. The dependence of orchid seeds on mycorrhizal fungi for germination is commonly admitted (Arditti and Ghani 2000). This might therefore be the case for Vanilla, although, to our knowledge, this has never been demonstrated.

Vegetative propagation ultimately remains the predominant reproduction mode in Vanilla. It naturally occurs from stem cuttings (Bouriquet 1954). In natural conditions, one individual of V. planifolia can cover very large areas, up to 0.2 ha, although not very densely (Soto Arenas 1999b). The predominance of vegetative propagation as strategy to the development of the settlements of V. chamissonis is reported (Macedo Reis 2000). For this reason, vanilla crops can be easily established from stem cuttings of 8–12 nodes, collected from healthy and vigorous vines (Stehlé 1952; Purseglove et al. 1981; Soto Arenas 2003).

Inter-specific hybridization

Vanilla is notable for providing one of the few cases in which natural hybridization in Neotropical orchids has been reported (Lubinsky et al. 2006). Nielsen et al. (Nielsen and Siegismund 1999; Nielsen 2000) evidenced some spontaneous wild hybrids between the Aphyllae species V. claviculata and V. barbellata in the Puerto Rico region using morphological and iso-enzymatic markers. However, inter-specific differentiation was found to be low between both species and was mainly due to differences in allelic frequencies rather than the presence of specific alleles. It was then suggested that these species are closely related with a relatively recent common ancestor (Nielsen and Siegismund 1999; Nielsen 2000).

Inter-specific hybridization between American species of the Foliosae section (Lamellosae sub-section) was used successfully in the breeding programme conducted in Madagascar where V. planifolia X V. tahitensis and V. planifolia X V. pompona hybrids were obtained (Delassus 1960; Dequaire 1976; FOFIFA 1990). It is thus highly probable that spontaneous hybrids between V. planifolia and other sympatric species could occur, as suggested between V. planifolia and V. pompona or V. insignis, sympatric near Papantla (Lubinsky 2004).

Soto Arenas (1999b) suggested that fragrant species such as V. barbellata (American species of the Aphyllae section) and V. inodora (American species of the Foliosae section Membranaceae sub-section) are too divergent from V. planifolia to expect successful hybridization with it. However, recently, breeding programs in India successfully created hybrids between the distantly related V. planifolia and V. aphylla (American species of the Foliosae section and Asiatic species of the Aphyllae section, respectively) (Minoo et al. 2006b), demonstrating the absence of reproductive barriers between distantly related Vanilla species.

The occurrence of inter-specific hybridization is an issue that must be considered and tested in taxonomic and phylogenetic studies of the genus. It could be one of the reasons for the difficulties in clearly delimiting and identifying species and for the present confusion in Vanilla taxonomy.

V. planifolia in its area of origin

The history of V. planifolia in Mexico

The pre-Columbian history of vanilla in Mexico is poorly documented. Stehlé (1952) refers to archive documents dating from 1427, which mention the gathering of vanilla pods by Aztecs. Vanilla belongs to a group of native plants in the Maya Lowlands that could have been subjected to human selection before 3400 B.C. However, its domestication origin is still unresolved (Colunga-GarciaMarin and Zizumbo-Villareal 2004). The pods were used for medicinal purposes, as mentioned in a 1552 Aztec herbal, as well as to flavour the hot chocolate drinks valued among the Aztec nobles (Bruman 1948). The first vanilla plantations were only established from 1767 by the Totonac Indians (in the Veracruz region), particularly in the Papantla and Misantla areas (Fig. 1), marking the start of vanilla cultivation. According to Ecott (2004), Soto Arenas considers that Totonac Indians did not use manual pollination to produce vanilla pods. Indeed, no evidence of manual pollination has been reported before the 19th century. From 1841, the technique of manual pollination discovered in Europe was transferred to Mexico and Totonac Indians became the world most important producers (Bruman 1948), until the supremacy of Madagascar in 1924 (Lucas 1990).

Vanilla diversity in Mexico

Today, V. planifolia is cultivated in Mexico mainly in two areas: northern Veracruz and northern Oaxaca (Fig. 1). Crops in Oaxaca were established by the end of the 1980’s from regional spontaneous specimens. In Veracruz, cultivated populations were established two centuries ago and their origin is unknown. They are less variable and very different from the cultivated populations of Oaxaca (Soto Arenas 1999a).

Four major morphological types of V. planifolia are grown in Mexico (Soto Arenas 2003). The four types differ in their vegetative appearance. The type ‘Mansa’ (or ‘Dura’ (Soto Arenas 2003) or ‘Fina’ (Castillo Martinez and Engleman 1993)) is the most widespread. Diaz (1989) refers to two sub-types among ‘Mansa’ which can be distinguished based on stem and leaf colour (‘Amarilla’ and ‘Verde’). The ‘Acamaya’ type, also named ‘Rayada’ or ‘Variegata’ shows yellow stripes on leaves and stems. The ‘Albo-marginata’ type has leaves with a white margin and does not exist in the wild. The ‘Oreja de Burro’ type is distinguished from ‘Mansa’ by its reproduction mode. ‘Oreja de Burro’ is self-incompatible while ‘Mansa’ is self-compatible (Castillo Martinez and Engleman 1993; Soto Arenas 2003). According to Castillo Martinez and Engleman (1993), who have carried out a morphological study on these types from July 1987 to January 1989 in the Papantla area, morphological differences are found. ‘Oreja de Burro’ vines are more ramified than ‘Mansa’ plants, with longer internodes, larger leaf area with a slight striation on the upper face, and higher growth rate (Castillo Martinez and Engleman 1993). Soto Arenas (2003) does not confirm these morphological variations but distinguishes the two types with other characteristics. The ‘Oreja de Burro’ type has a better growth and a better resistance to fungi infection than ‘Mansa’ (Soto Arenas 2003). The existence of ‘Oreja de Burro’ is a problem in Mexico because its use has strongly hindered vanilla production (Soto Arenas 2003). Lopez ((1900), in: (Diaz 1989)) cites another type, ‘Mestiza’ which has larger leaves and fruits than ‘Mansa’, but Castillo Martinez and Engleman (1993) quote ‘Mestiza’ as the ancient name for ‘Oreja de Burro’. There is no evidence to suggest whether ‘Mansa’, ‘Acamaya’, ‘Albo-marginata’, and ‘Oreja de Burro’ represent different clones of the same species or are hybrids between V. planifolia and closely related species (Lubinsky 2003).

The first molecular data on specimens from crops in northern Veracruz, Oaxaca and other Mexican regions were obtained using iso-enzymes and showed low levels of total genetic variation. However, the data showed an important differentiation between plants from the two main regions, with on one hand a homozygote excess in Veracruz and on the other hand a higher genotypic diversity with some heterozygote individuals in Oaxaca (Soto Arenas 1999b). These results were recently confirmed by another iso-enzymatic study performed to characterize the cultivated and spontaneous vanilla in the south-eastern region of Mexico, with the highest heterozygosity and polymorphism levels found in natural populations from northern Oaxaca. Unique genotypes were found in wild accessions in Tabasco, Chiapas and Guatemala ((Cibrian Jaramillo 1999), in: (Lubinsky 2003)).

A molecular analysis using RAPD markers allowed a further genetic characterization of V. planifolia in Meso-America. Using these markers, Schlüter (2002) was able to differentiate V. planifolia from Costa Rica and Mexico. Among Mexican V. planifolia, two main groups were revealed: individuals from Oaxaca, Chiapas and Quintana Roo on one hand, and individuals from Veracruz, Federal District of Mexico, San Luis Potosi, Tabasco and Oaxaca on the other hand (Fig. 1). The individuals from Oaxaca and Tabasco present in the second group most probably correspond to specimens that were collected from the Veracruz region at the time of the establishment of new crops. As shown in the isozyme study, the specimens from Oaxaca were found more diverse than those from Veracruz.

Attempts to study some hypervariable regions, such as introns in three protein coding genes (alcohol dehydrogenase, calmoduline and glyceraldehyde 3-phosphate dehydrogenase) in order to detect intra-specific sequence variations were unsuccessful due to lack in polymorphism (Soto Arenas 1999b).

V. planifolia in introduction areas

V. planifolia dissemination from its area of origin

Following the discovery of the New World by C. Colombus in 1492, the earliest vanilla dissemination record from Mexico is the one by Father Labat who imported three V. planifolia vines into Martinique in 1697 (Lecomte 1901); and from there to Guadeloupe in 1701 (Table 2). There is then an early record of an introduction in Reunion Island in 1793 (Ridley 1912) (Table 2). Early in the 19th century, one major event is V. planifolia introduction by Marquis of Blandford into the collection of C. Greville at Paddington where it flowered in 1807. Greville then supposedly sent some cuttings to the botanical gardens of Antwerp (Belgium) and from there to Paris (Correll 1953; Bouriquet 1954). From the Botanical Garden of Antwerp, it was introduced into Buitenzorg in Java in 1819 by Marchal (Lecomte 1901; Bouriquet 1954; Purseglove et al. 1981), and in Reunion Island from the Jardin du Roi in Paris in 1822 by the ordinance officer of Bourbon, Marchant (Lecomte 1901, Bouriquet 1954) (Table 2). Another early introduction event is documented in India in 1835 (Table 2) but the plant died after flowering (Correll 1953).
Table 2

Historical records of vanilla dissemination around the world

Introduction date

Cultivation date record

Localization

From

By

Ref.

West Indies

1697

 

Martinique

 

Father Labat

(6)

1701

 

Guadeloupe

Martinique

Father Labat

(2, 6)

1900’s

 

St Vincent

Seychelles

 

(5)

<1900

 

Puerto Rico

Mexico

Parent of MM Utuano

(2)

1909

 

Puerto Rico

United States Plant Introduction Garden of Florida

Federal Experiment Station of Puerto Rico

(2, 5)

 

1898

French Guyana

  

(10)

Not specified

 

Surinam

  

(10)

 

1896

Trinidad

  

(10)

 

1910’s

Jamaica

  

(10)

Indian Ocean

1793

 

Reunion Island

  

(10)

1819

 

Reunion Island

Cayenne

Philibert & Perrotet

(2, 6)

1820

 

Reunion Island

Philippines

Philibert & Perrotet

(2, 6)

1822

 

Reunion Island

Jardin du Roi in Paris

Marchant

(2, 6)

1875

 

Reunion Island

Martinique

 

(6)

1827

 

Mauritius

  

(5)

1836

 

Mauritius

  

(6)

1880

 

Mauritius

  

(2, 4)

 

1865

Mauritius

  

(6)

1866

 

Seychelles

  

(2, 8)

1873

 

Comoro Islands

  

(2)

1893

 

Comoro Islands

  

(5)

1870

 

Madagascar (Nosy-Be)

  

(2)

1880

 

Madagascar (Nosy-Be)

  

(7)

1880

 

Madagascar (Ste Marie)

  

(7)

1840

 

Madagascar

  

(5)

1890

 

Madagascar

  

(7)

Africa

1852

 

French Congo (Gabon) near Libreville

 

Aubry-Lecomte

(2, 5, 6, 10)

1873

 

French Congo (Gabon) (Libreville)

 

Father Klaine Mission Sainte-Marie

(2, 5, 6, 10)

<1900

1901

Zanzibar

  

(10)

 

1894

Independent State of Congo

  

(10)

Not specified

 

Cameroons (present Nigeria and Cameroon), Sao Tome, Sierra Leone, Lagos (Nigeria), German Eastern Africa (present Burundi, Rwanda and Tanganyika (Tanzania))

  

(10)

1912

 

Uganda

Ceylon

 

(8)

 

1950’s

Uganda

  

(9)

Asia

1835

 

India

  

(5)

1853

 

Ceylon

 

Dr. Thwaites, manager of the Botanical Garden of Peradeniya

(6)

1819

 

Botanical Garden of Buitenzorg (Java)

Botanical Garden of Antwerp

Marchal

(2, 6, 8)

 

1846

Java

  

(5, 10)

 

20th c.

Bali, Sulawesi, Java, Sumatra

  

(1)

1861

 

New-Caledonia

 

Botanist Pierre

(2)

1865

 

Indochina

 

Botanist Pierre

(2)

1883

 

Australia (Queensland)

  

(5)

Oceania

<1900

 

Hawaii, Tahiti, Samoa, Fiji

Mexico

 

(5)

 

1903

  

Hawaii Agricultural Experiment Station

(5, 10)

References are: (1): Bernard (2005); (2): Bouriquet (1954); (3): Childers and Cibes (1948); (4): Childers et al. (1959); (5): Correll (1953); (6): Lecomte (1901); (7): Lucas (1990); (8): Purseglove et al. (1981); (9): Ranadive (2005); (10): Ridley (1912)

The lack of natural pollinators in the areas of introduction prevented sexual reproduction and pod production until the first half of the 19th century. Artificial pollination was discovered twice, independently and a few years apart, by C. Morren in 1836 in Liège and by J. Neumann in 1838 in Paris. However in all V. planifolia cultivating countries, pollination is made according to an easy manual technique discovered in 1841 by Edmond Albius in Reunion Island (Lecomte 1901; Childers and Cibes 1948). From 1841, when this simple manual pollination technique for vanilla was shared out among vanilla growers, vanilla vines were rapidly spread to the whole Indian Ocean area (Mauritius, Seychelles, Madagascar, Comoros) and to the rest of the world (Africa, Asia, Oceania, West Indies) during the 19th century (Table 2).

V. planifolia diversity in introduction areas

It is commonly agreed that V. planifolia cultivation started from a very narrow genetic pool (Soto Arenas 1999b; Lubinsky 2003), and that the extensive vegetative reproduction that followed might be responsible for the relatively high genetic uniformity observed to date in vanilla crops. Soto Arenas even suggested that most V. planifolia cultivated specimens around the world came from a unique clone (of the ‘Mansa’ type) exported from Mexico (Ecott 2004) to the C. Greville collection at Paddington and disseminated by cuttings via the botanical gardens of Paris and Antwerp, from where it would have been disseminated worldwide (Correll 1953; Bouriquet 1954). However, in many areas of introduction, some phenotypic variation is described. This will be discussed based on the case of Reunion Island as it was recently extensively studied, but this could be applied to other introduction areas.

Phenotypic diversity in Reunion Island

In Reunion Island, growers distinguish two main morphological types of V. planifolia. The major type named ‘Classique’ has a light green colour, flat leaves and pods progressively narrow. The second main type named ‘Mexique’ or ‘Bleue’ has darker and more bluish leaves than ‘Classique’, with a central gutter and curved sides and pods are cylindrical up to the stem. There are also other minor variants recognized by growers: the ‘Aiguille’ type which is distinguishable from the ‘Classique’ type by its more slender leaves and a thin pod; the ‘Grosse Vanille’ type which resembles V. pompona, with bigger and thicker leaves and stem than ‘Classique’; the ‘Stérile’ type with the same morphological characteristics than ‘Classique’ and which seems self-incompatible and lastly, the ‘Variegata’ type which exhibits leaves with white yellow stripes (Figure 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs10722-007-9260-3/MediaObjects/10722_2007_9260_Fig2_HTML.gif
Fig. 2

Morphological types from Reunion Island

Possible origins of this phenotypic diversity

Four hypotheses can be proposed to explain the morphological variability in cultivation areas: (i) the existence of different introduction events, (ii) the accumulation of somatic mutations, (iii) the possible role of sexual recombination, and (iv) the existence of epigenetic phenomena:

(i) In most regions, different introduction events are often reported in the literature (Table 2). In Reunion Island, five successive vanilla introductions can be identified (Table 2). Following a poorly documented first introduction in 1793 (Ridley 1912), a second introduction occurred on June 27th, 1819. Commander Philibert and the botanist Perrotet brought back from Cayenne a ‘big vanilla’, possibly V. pompona, but all plants apparently disappeared (Lecomte 1901; Bouriquet 1954). The third introduction occurred on May 6th, 1820. Philibert and Perrotet brought back from the San-Matteo forest in the Manilla region (Philippines) a vanilla species of small size, which seemed new, possibly V. planifolia. All plants also supposedly disappeared (Lecomte 1901; Bouriquet 1954). The fourth introduction occurred on September 25th, 1822 when Marchant brought back V. planifolia from the Jardin du Roi in Paris a Mexican vanilla (Lecomte 1901; Bouriquet 1954). This introduction event is thought to be the first successful one and the origin of most vanilla plants cultivated in Reunion Island today. However there is no evidence that the first introduced vines actually disappeared or that other introduction events did not occur later, as reported by Lecomte (1901) with a fifth introduction in 1875, of a Mexican vanilla from Martinique. The observed phenotypic variation could therefore be the result of these different historical introduction events.

Genetic diversity studies were conducted on V. planifolia cultivated in Reunion Island and related species using RAPD (Besse et al. 2004) and AFLP markers (Duval et al. 2006). A high variability was revealed within V. pompona and may be related to its large area of dispersion, from Mexico to Paraguay (Portères 1954). Conversely, an important intra-specific homogeneity was observed within V. planifolia. A unique origin of the disseminated material for this species may explain its high homogeneity. These results therefore confirm historical data indicating that the genetic secondary pool of V. planifolia in Reunion Island was probably constituted from one introduction event.

(ii) Vanilla is a vine with vegetative reproduction, and as such, can accumulate somatic point mutations, which may be responsible for morphological diversity, as observed for the colour and the shape of flowers and leaves (Dequaire 1976). These variants may then have been selected by vanilla growers and multiplied by cuttings. In an AFLP study (Duval et al. 2006), the level of genetic differentiation revealed in V. planifolia accessions cultivated in Reunion Island is indeed consistent with the hypothesis of an accumulation of somatic point mutations, as witnessed by the high number of rare alleles revealed. But this observed somatic point mutation diversity showed however no relationship with the observed morphological diversity, particularly for the two main morphological types cultivated on the island: ‘Classique’ and ‘Mexique’.

(iii) In introduction areas, such as Reunion Island, vanilla crop must be hand-pollinated. Seed germination may occur from pods dropped down on vanilla plots. For instance, this was observed in Guadeloupe (Ph. Feldmann, pers. comm.). Moreover, in Reunion Island, rare natural pollination events are reported on high flowers, such flowers being often visited by the bird Zosterops (Zosteropidae) or by ants (P. Fontaine, Jardin des Parfums et des Epices, pers. com.). Interestingly, Zosterops was recently shown to be involved in the pollination of an Angraecoid Orchid in Reunion Island (Micheneau et al. 2006). In Madagascar, Delassus (1960) also describes seeds germinating in the wild, but seedlings rarely reach adult age. Although it is clear, according to these arguments, that sexual recombination is expected to be a rare phenomenon in the areas of introduction, it is important to keep in mind that a single sexual reproduction event is able to generate numerous genotypes that can be vegetatively propagated rapidly. The importance of sexuality in vegetatively reproduced crops and its impact on genetic diversity has been previously shown for yam (Scarcelli et al. 2006) and cassava (Elias et al. 2001). In Reunion Island, AFLP analysis revealed a large majority of relatively homogeneous accessions (Duval et al. 2006), but a few individuals exhibited very different genotypes. It is therefore probable that sexual reproduction may have played a role in the apparition of morphological variability. The development of molecular tools with higher resolution and co-dominance may help to detect the occurrence of rare sexual multiplication and unravel additional genetic diversity thus explaining some of the observed morphological diversity.

(iv) Lastly, mechanisms of epigenetic origin may also be responsible for the observed phenotypic variations in the absence of congruent genetic variation. Such epigenetic variations may include various methylation states of DNA cytosines. These are responsible for the regulation of gene expression, and were shown to be transmitted through vegetative propagation (Xiong et al. 1999; Bretó et al. 2001; Imazio et al. 2002; Kalisz and Purugganan 2004). The study of cytosine methylation is now possible on a large scale using the Methylation Sensitive Amplified Polymorphism analysis (MSAP) (Reyna-López et al. 1997). This technique could be applied to cultivated Vanilla accessions to see if such epigenetic phenomena could explain morphological variation (in the absence of genetic variation). Other epigenetic phenomena have been described such as modifications on particular histone residues, histone variants in chromatin, DNA looping in chromatin conformation and higher-order chromatin structure (Mager and Bartolomei 2005). However, these are more difficult to analyze routinely and on a large sampling.

Our recent study revealed that polyploidization could also explain the morphological variation in cultivated species (Duval et al. 2006). Most of the data found in the literature indicate a basic chromosome number n = 16 and 2n = 32 for V. planifolia (Hoffmann 1929, 1930; Heim 1954; Chardard 1963; Martin 1963) but various numbers (from 13 to 53) are also commonly reported (Hurel-Py 1938; Nair and Ravindran 1994). Moreover, the occurrence of different ploidy levels is mentioned for V. tahitensis ((Tonier 1951), in: (Soto Arenas 2003) and was confirmed using flow cytometry, revealing two ploidy levels (diploid and tetraploid) in this species, with triploid artificial hybrids (Duval et al. 2006). Preliminary results obtained for cultivated V. planifolia in Reunion Island also seem to reveal a similar pattern of variation (Duval et al. 2006) and this is currently further investigated.

Preservation of vanilla genetic resources

The primary V. planifolia gene pool is severely threatened in its area of origin by deforestation, and overcollecting in order to establish crop, both leading to the extinction of wild populations, and sometimes by poor crop management (Soto Arenas 1999b, 2006a). For over 20 years, M.A. Soto Arenas, an orchid and vanilla botanical expert, has explored southern Mexico looking for wild V. planifolia specimens. He found less than thirty genuine wild specimens spreading naturally (Ecott 2004). These natural populations, source of new diversity, are about to disappear. In the main production area of Veracruz, no wild population is known but northern Oaxaca, near Tuxtepec (Fig. 1), is important as a centre for vanilla diversity (Lubinsky 2003). Some wild specimen were collected (1882–1955) in Guatemala and Belize, but none was found in recent prospections (Soto Arenas 1999b). The best way to preserve vanilla genetic resources should be in situ conservation in its natural habitat. Unfortunately, these areas are under high human pressure and in situ conservation cannot be a viable alternative if it is not connected to global area conservation strategies (Soto Arenas 1999b). If such a strategy was to be developed, it should be focused in the Oaxaca region in order to preserve the few remaining Vanilla genetic resources (Lubinsky 2003).

The secondary gene pool, represented by closely related species of V. planifolia, and the pool of local varieties maintained in V. planifolia introduction areas, must also be protected. Some desirable characters from the secondary gene pool can be used for breeding, such as self-pollination, root-rot and virus resistance, larger fruit set, less photoperiod dependence for flowering, better aromatic profile, and pod indehiscence (Soto Arenas 1999a; Bénézet et al. 2000; Leclerc-Le Quillec et al. 2001).

Ex situ conservation is in the present conditions the best strategy to preserve vanilla genetic resources. The establishment of germplasm banks in the field or in vitro is therefore essential to perpetuate the existent genotypes and to improve vanilla breeding and production. To date, several collections have thus been established worldwide.

In Madagascar, IRAM (Institut de Recherches Agronomiques à Madagascar) and IRAT (Institut de Recherches Agronomiques Tropicales) led vanilla breeding programs in the 1950–1980’s in the research station of Ambohitsara near Antalaha. A collection of approximately one hundred species was established in 1974 (FOFIFA 1990) from introduced exotic species and species collected locally. By the end of the 1990’s, only 22 species remained as well as about 250 hybrids resulting from former breeding programs (Grisoni et al. 1997). Nowadays unfortunately, this collection is abandoned.

Weiss (2002) mentions the existence of two germplasm collections. The first collection in CATIE (Centro Agronómico Tropical de Investigación y Enseñanza) of Turrialba (Costa Rica) includes Central American accessions, and has recently suffered from Fusarium infection (Lubinsky 2004). The second collection is in NRCS (National Research Centre for Spices) of Calicut (India), now IISR (Indian Institute of Spices Research), and comprises 28 vanilla accessions. Minoo (2006a) cites an in vitro gene bank at IISR with more than 300 genotypes including four Indian Vanilla species, exotic, indigenous, seedling progenies and somaclones. This collection has a long-term purpose of exchange.

In Princeton (New Jersey), wild Mexican species have been gathered in the 1990’s by Soto Arenas during a prospecting programme led by the Union Camp Corporation (now the International Paper company) (Lubinsky 2004). No more information can be found in terms of access and exchange.

Several botanical gardens possess some vanilla specimens in greenhouses, such as: Jardin des Plantes (Nantes, Montpellier), Jardin Botanique du Parc de la Tête d’Or (Lyon), Serres d’Auteuil (Paris), Jardin du Luxembourg (Sénat, Paris), Botanical Garden of the University of Copenhagen, Botanical Gardens of Antananarivo (Madagascar) and Royal Botanic Gardens of Kew. Moreover, there are collections by orchid hobbyists throughout the world that probably contain vanilla species. Since 1983, M. Pignal, curator at the Museum National d’Histoire Naturelle of Paris, has been setting up a vanilla collection of wild species in the Jardin des Plantes of Cherbourg, from prospections in Brazil, French Guyana and Costa Rica (Rongier 2006).

In the Pacific, Etablissement Vanille de Tahiti (EVT, French Polynesia) and the Secretary of the Pacific Community (SPC, Fiji) have handled a collection of about 250 accessions (Guarino 2004) covering 20 cultivars of V. tahitensis from French Polynesia and a few V. planifolia and hybrids (Wong 2005).

In Reunion Island, a collection of vanilla genetic resources started in 2002 and comprises today over 500 accessions including over 200 vines maintained in shade-houses. The collection contains a sample set of cultivated species and approximately thirty related species (Grisoni et al. 2006).

Botanical Gardens, SPC and Reunion accessions are available for exchange depending on their legal status (MTA).

Lastly, the CITRO (Centro de Investigaciones Tropicales) of the University of Veracruz sets out to initiate the creation of a Botanical Garden where the maximal amount of species (wild and cultivated) of the existing Vanilla types from Mexico and elsewhere will be collected. Many institutions contribute to this collection as the University of California, Herbario AMO (Asociación Mexicana de Orquídeas), the INIFAP (Instituto Nacional de Investigaciones Forestales, Agrícolas y Pecuarias) and the EVT (http://www.uv.mx/). At present, we do not know the status of this collection in terms of access and exchange.

Conclusions

Efforts are necessary to revise the taxonomy of the genus Vanilla, which remains unclear, and to clarify its place within the Orchidaceae family on the basis of molecular phylogenetics. One interesting proposition was made by Cameron (2005a) aiming to set up a bar code system for vanilla. He suggests the use of the ITS region and the psbA-trnH intergenic spacer that were sequenced for over 50 vanilla accessions collected worldwide. The bar code project is particularly interesting as it may allow routine identification of Vanilla specimens to the species level, and perhaps even to the accession level, and may further allow the specific identification of processed pods. To build a robust phylogeny for the Vanilla genus, reference herbarium specimens will need to be included. For this purpose, the development of plastid mononucleotide microsatellites should be considered for vanilla (particularly when using degraded DNA samples extracted from herbarium material), as they have already been successfully used for bio-geographical studies in orchids (Micheneau 2002; Fay and Krauss 2003).

The reproductive mode of Vanilla species in natural conditions is also a questioning topic. Despite its numerous primitive characters, vanilla flowers show a structure and pollination mechanisms as complex as the ones observed in derived orchids. Floral morphology evolution in Vanilla could be a model study to understand the evolution of orchids with one anther (Soto Arenas 1999b). Evolutionary inferences can also be made from further studies on fragrant flowers and fruits interactions with orchid bees, as Vanilla is the most primitive orchid genus to demonstrate this interaction (Lubinsky et al. 2006).

As detailed earlier, unravelling the mechanisms (sexual reproduction, somatic mutation, epigenetics...) responsible for the phenotypic variation observed for V. planifolia in introduction areas will be as well an exciting issue to pursue.

Finally, polyploidization is an additional phenomenon that may be at the origin of the morphological variation observed in cultivated species. This could also represent an important phenomenon in the evolution of the Vanilla genus.

A joint international effort is now essential to protect and study vanilla in its area of origin, where it is highly endangered, as well as in introduction areas, where it is considered as a patrimonial resource with traditional cultivation practices going back to the 19th century.

Copyright information

© Springer Science+Business Media B.V. 2007