Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

7.1 Introduction

The species of the genus Passiflora L., commonly known as passionflowers, are recognized for their tasty fruits, pharmaceutical properties and ornamental flowers. Although Passiflora species are numerous, few have significant commercial value. The vast majority are wild and poorly known. However, they constitute an important source of unexploited genetic diversity for the improvement of cultivated Passiflora species and the introduction of new crops. In addition, wild Passiflora species have potential useful characters, such as the resistance to particular pathogens, which can be incorporated by traditional breeding or molecular methods in crop species. The knowledge of the relatedness among Passiflora species is essential for the identification of wild crop relatives for future breeding programs. In this chapter, we review the advances in genetics, breeding and conservation of Passiflora species.

7.2 Botany

The pantropical family Passifloraceae comprises more than 600 species (Vanderplanck 2000) accommodated in 18 genera. Of this, Passiflora is the largest with ca. 520 species (MacDougal and Feuillet 2004). The first report of a Passiflora species was made in 1553 by Pedro Cieza de León during the Spanish colonial period in South America (Linnaeus 1753). In 1574, the Spanish botanist Nicolás Monardes suggested that flowers of Passiflora symbolized the Passion of the Christ. Linnaeus (1753) named the genus Passiflora, from the latin flos passionis that means “suffering flower”. Since Linnaeus’ publications, in which he described 22 and 24 Passiflora species (Linnaeus 1745, 1753), more than 500 new species have been described (MacDougal and Feuillet 2004) and their number continues to increase (Coppens d’Eeckenbrugge et al. 2001a; MacDougal 2001, 2003; Cervi 2002, 2006; Feuillet 2002, 2004; MacDougal and Hansen 2003; Jørgensen 2004; Jørgensen and Weigend 2004; MacDougal 2004, 2006; Vitta and Bernacci 2004; Krosnick 2005; Porter-Utley 2007; Vásquez et al. 2007; Azevedo 2008).

The center of diversity of Passiflora is located in the neotropics, with South America accounting for 95% of all the species (Fig. 7.1). A few species are also found in subtropical, and even temperate, regions of North and South America and 24 species are native from Southeast Asia, Australia and the Pacific Islands (Krosnick and Freudenstein 2005). Overall, passionflowers are adapted to a wide range of climates (0–40° latitude), altitudes (0–4,500 m) and ecosystems (from tropical humid forest to arid regions).

Fig. 7.1
figure 1_7

Distribution of Passiflora richness according to Ocampo et al. (2007) and Shawn Krosnick (personal communication)

Passiflora flowers are mostly pollinated by bees, hummingbirds and bats (MacDougal 1983, 1994; Endress 1994; Büchert 1998; Kay 2001), but other types of pollinators such as wasps, butterflies and moths have also been recorded (MacDougal 1983). Many species, such as P. edulis f. flavicarpa O. Deg. and P. incarnata L., are self-incompatible and their fertilization is done through cross-pollination. The coevolution between plants of Passiflora and larvae of butterflies of tribe Heliconiini has been the subject of several evolutionary studies (Gilbert 1971; Benson et al. 1975; Benson 1978; Gilbert 1980, 1982, 1983; Mitter and Brooks 1983; Spencer 1988). The caterpillars of these butterflies feed exclusively on Passiflora leaves.

7.2.1 Systematics

Based on previous classifications (Linnaeus 1753; Jussieu 1805a, b; De Candolle 1822; Roemer 1846; Harms 1923, 1925), Killip (1938) wrote the largest monograph on the genus including descriptions for 335 American species. His classification was based on floral characters and subdivided the genus into 22 subgenera. The Old World species were accommodated in the section Disemma of the subgenus Decaloba (Harms 1925). Escobar (1989) added the subgenus Porphyropathantus. The subgeneric classification of the genus has been recently challenged on morphological grounds by Feuillet and MacDougal (2003), who recognize only four subgenera which are subdivided into supersections, sections and series.

Cultivated species are essentially regrouped into four supersections of the largest subgenus Passiflora (see Sect. 7.4). At the subgenus level, the Feuillet and MacDougal classification is partly supported by recent molecular phylogenetic studies, as they clearly distinguish three major clades within Passiflora, corresponding to the three major subgenera (Muschner et al. 2003; Yockteng and Nadot 2004a; Hansen et al. 2006). The first problem is that this classification does not consider hundreds of species, such as the Old World species Passiflora tetrandra Banks ex DC. The second problem in the classification is the small subgenus Deidamioides, which regroups 13 species with dissimilar morphology. In fact, these species appear to be dispersed in phylogenetic reconstructions (Yockteng and Nadot 2004a; Krosnick and Freudenstein 2005). At lower levels, some supersections are supported by phylogenetic studies. For instance, the phylogeny based on nuclear sequences of glutamine synthase (ncpGS) supports mainly all supersections of subgenus Decaloba and the supersections Coccinea, Passiflora and Tacsonia of subgenus Passiflora (Yockteng and Nadot 2004a). However, the supersections Stipulata and Laurifolia of subgenus Passiflora do not form a clade in phylogenies. Members of the supersection Stipulata are dispersed in two non-related groups, in which section Dysosmia appears at the base of subgenus Passiflora and sections Tacsonioides and Granadillastrum are regrouped in one clade. Supersection Laurifolia is also divided into two well-supported but non-related clades. On one hand, the species of series Tiliifolia form a well supported clade closely related to species of supersection Stipulata except for the species in section Dysosmia. On the other hand, the species of series Laurifoliae and Quadrangulares form a clade closely related to supersection Passiflora. The results of this phylogenetic study point out the necessary revision of the classification by MacDougal and Feuillet. Molecular reconstructions are essential in this task; however, their robustness and resolution have not been sufficient for appropriate inferences on the divisions under the supersection level. Moreover, the interpretation of molecular data is highly complicated by the level of reticulation and the complexity of genome transmission in Passifloraceae. Interspecific hybridizations seem common in the genus, and chloroplast inheritance is often paternal or even biparental (see Sect. 7.5.3). Thus, phylogenetic inferences based on cpDNA sequences face the most severe problems anticipated by Harris and Ingram (1991) for classical phylogenetic methods, i.e., high intraspecific and even intraindividual variation and consequently lineage sorting, and also reticulate evolution of the chloroplast genome through hybridization and introgression. Studies based on nuclear ribosomal genes may suffer similar limitations. Many Passiflora species have originated from recent radiation, involving recent isolation and hybridization events, giving an incomplete concerted evolution of ribosomal gene copies in the genome (Lorenz-Lemke et al. 2005). Understanding the evolution of Passiflora is likely to be a long task that will impose many different molecular and morphological approaches as well as original methods for the interpretation of their data.

Although the classification of MacDougal and Feuillet (2004) needs to be revised and completed with the inclusion of many species, we have used it in the present text because it is the most recent to date.

  • 1. Subgenus Astrophea (DC.) Mast. – 57 species

  • 1.1. Supersection Astrophea Feuillet & MacDougal

  • 1.1.1. Section Astrophea Feuillet & MacDougal

  • 1.1.2. Section Capreolata MacDougal & Feuillet

  • 1.1.3. Section Leptopoda Killip ex Feuillet & Cremers

  • 1.2. Supersection Pseudoastrophea (Harms) Feuillet & MacDougal

  • 1.2.1. Section Pseudoastrophea (Harms) Feuillet & MacDougal

  • 1.2.2. Section Botryastrophea (Harms) Killip

  • 1.2.2.1. Series Botryastrophea (Harms) MacDougal & Feuillet

  • 1.2.2.2. Series Carneae Feuillet

  • 2. Subgenus Deidamioides (Harms) Killip – 13 species

  • 2.1. Section Deidamioides (Harms) Feuillet & MacDougal

  • 2.2. Section Polyanthea DC

  • 2.3. Section Tetrastylis (Bard. Rodr.) Harms

  • 2.4. Section Mayapanthathus MacDougal & Feuillet

  • 2.5. Section Tryphostemmatoides Harms

  • 3. Subgenus Decaloba (DC.) Rchb. – 214 species

  • 3.1. Supersection Pterosperma Gilbert & MacDougal

  • 3.2. Supersection Hahniopathanthus (Harms) MacDougal & Feuillet

  • 3.3. Supersection Disemma (Labill.) MacDougal & Feuillet

  • 3.3.1. Section Octandranthus Harms

  • 3.3.2. Section Disemma (Labill.) MacDougal & Feuillet

  • 3.3.3. Section Hollrungiella Harms

  • 3.4. Supersection Multiflora (Small) MacDougal & Feuillet

  • 3.5. Supersection Auriculata MacDougal & Feuillet

  • 3.6. Supersection Cieca (Medic.) MacDougal & Feuillet

  • 3.7. Supersection Bryonioides (Harms) MacDougal & Feuillet

  • 3.8. Supersection Decaloba (DC.) MacDougal & Feuillet

  • 3.8.1. Section Decaloba DC.

  • 3.8.2. Section Xerogona (Raf) Killip

  • 4. Subgenus Passiflora Feuillet & MacDougal – 236 species

  • 4.1. Supersection Passiflora Feuillet & MacDougal

  • 4.1.2. Series Passiflora Feuillet & MacDougal

  • 4.1.3. Series Palmatisectae Feuillet & MacDougal

  • 4.1.4. Series Pedatae Killip ex Cervi

  • 4.1.5. Series Setaceae Killip ex Cervi

  • 4.2. Supersection Stipulata Feuillet & MacDougal

  • 4.2.1. Section Granadillastrum Tr. & Planch

  • 4.2.2. Section Calopathanthus Harms

  • 4.2.3. Section Tacsonioides DC.

  • 4.2.4. Section Kermesinae (Cervi) Feuillet & MacDougal

  • 4.2.5. Section Dysosmia DC.

  • 4.3. Supersection Laurifolia (Cervi) Feuillet & MacDougal

  • 4.3.1. Series Laurifoliae Killip ex Cervi

  • 4.3.2. Series Quadrangulares Feuillet & MacDougal

  • 4.3.3. Series Tiliafolia Feuillet & MacDougal

  • 4.3.4. Series Marginatae Killip ex Cervi

  • 4.4. Supersection Coccinea Feuillet & MacDougal

  • 4.5. Supersection Distephana (DC.) Feuillet & MacDougal

  • 4.6. Supersection Tacsonia (Juss.) Feuillet & MacDougal

  • 4.6.1. Section Rathea (Karst.) Harms

  • 4.6.2. Section Insignes Feuillet & MacDougal

  • 4.6.3. Section Colombiana Esc.

  • 4.6.3.1 Series Leptomischae Esc.

  • 4.6.3.2. Series Colombianae Esc.

  • 4.6.3.3. Series Quindiensae Esc.

  • 4.6.4. Section Parritana Esc.

  • 4.6.5. Section Fimbriatistipula Esc.

  • 4.6.6. Section Tacsoniopsis Tr. & Planch

  • 4.6.7. Section Elkea Feuillet & MacDougal

  • 4.6.8. Section Tacsonia (Juss.) Feuillet & MacDougal

  • 4.6.9. Section Boliviana (Harms) Feuillet & MacDougal

  • 4.6.10. Section Trifoliata Feuillet & MacDougal

  • 4.6.11. Section Manicata (Harms) Feuillet & MacDougal

7.2.2 Morphology

Passiflora species exhibit a highly diversified morphology. Passionflowers are in general lianas, vines or herbs, mostly climbing with woody or herbaceous stems and axillary tendrils. Some species are arborescent and are regrouped in subgenus Astrophea. The stem can be cylindrical, angular or even winged. All species present stipules, sometimes deciduous, differing widely in size and shape. Leaves are alternate, simple, entire, lobed or palmate; rarely compound, such as in P. cirrhiflora Juss., P. deidamioides Harms, P. pedata L. and P. trifoliata Cav. The margins of the leaves range from serrate to entire. The extreme diversity of leaf morphology in Passiflora has been attributed to the pressure exerted in their coevolution with their main herbivores, the larvae of the Heliconiine butterflies (Gilbert 1982). The genus is also characterized by the presence of extrafloral nectaries that vary in form, position and number. They are common on the leaf petiole and along leaf margins, and can be also found on sepals or bract margins. In subgenus Decaloba, they are most often found in the lamina. These extrafloral nectaries (EFN) attract and reward ants that often have a protective role against herbivores, especially butterfly caterpillars, and nectar robbers (Smiley 1986; Apple and Feener 2001; Leal et al. 2006). Most passionflowers have three bracts, rarely less. Bracts can be foliar, elliptic, oblong or oval, persistent or deciduous with entire or dentate margin. The flowers are axillary, regular and hermaphroditic. Only P. tetrandra Banks ex DC. is dioecious. The calyx can be patelliform, campanulate or tubular. The perianth is composed of five sepals and five petals often similar in color and shape, and its color varies from flat green to showy violet or red. The most striking floral feature is the corona that is constituted of filaments in one or several series. These can be long, short or fused into a tube, such as in P. murucuja L. This character plays an essential role in the attraction and selection of pollinators (Endress 1994; Kay 2003; Yockteng 2003). A circular flower nectary is at the base of the corona. The reproductive organs are borne by an androgynophore holding the five stamens and three carpels, except in some Asiatic species (Krosnick et al. 2006). The ovary has one locule with multiple ovules disposed in three parietal placentas. The fruits are small, purple or blackish indehiscent berries with one or few seeds, commonly in subgenus Decaloba, or yellow or orange to red capsules with many seeds, mostly edible and measuring up to 30 cm long, commonly in subgenus Passiflora.

7.2.3 Cytology

Cytological data are essential to study the relationship between wild and cultivated species and to plan out interspecific hybridizations for future breeding programs.

Until now, chromosome counts have been reported for 94 Passiflora species (Table 7.1). The most common chromosome numbers are 2n = 12, 2n = 18 and 2n = 24, which correspond roughly to the three main subgenera. The most common chromosome numbers of subgenera Passiflora and Decaloba are 2n = 18 and 2n = 12, respectively. Few species of subgenus Astrophea have been examined but the results indicate that the predominant number in the group is 2n = 24. Divergent hypotheses about the base chromosome number of the genus have been proposed. A recent study based on a phylogenetic analysis proposed x = 12 as the base number (Hansen et al. 2006). Karyotype analyses have shown that several characters (number and position of satellites, number and length of chromosomes and the position of the centromere) are specific to each subgenus (Snow and Macdougal 1993). The majority of species studied so far are diploid, but some species display various ploidy levels such as P. misera Kunth. with 2n = 12, 36, P suberosa L. with 2n = 12, 24 and 36, P. subpeltata Ortega with 2n = 18, 72 and P. incarnata with 2n = 18, 36. Tetraploid hybrids (2n = 36) have also been obtained from somatic hybridization of diploid species (2n = 18) (Soares-Scott et al. 2005). However, the main commercial species and horticultural hybrids, obtained through sexual hybridization, are in general diploids (Vieira and Carneiro 2004).

Table 7.1 Chromosome numbers of Passiflora speciesa

Genome size values (2C) obtained for 10 diploid species of subgenus Passiflora range from 3.16 to 5.36 pg while the genome size of a tetraploid species of subgenus Decaloba is 1.83. Data of genome size can be useful in the assessment of somatic and sexual compatibility in Passiflora species. Further studies are, therefore, needed to increase data of genome size in Passiflora.

Examining the sites of 5S and 45S of rDNA, Melo and Guerra (2003) observed that species with x = 6 present only two 5S sites and two or four 45S sites whereas other species with x = 9 or 10 display more than two sites for both 45S and 5S (Table 7.1). These data suggest a diploid origin for the species with x = 6 and a polyploid origin for genomes with x = 9 or 10. In addition, the numbers of 45S and 5S sites should correspond to the ploidy level of the individual analyzed. In fact, they allow differentiating the two cytotypes of P. misera in which the number of 5S and 45S is three times larger in the hexaploid cytotype than in the diploid cytotype. In addition, interspecific hybrids derived from species of supersection Passiflora can be differentiated by their number of 45S rDNA sites (Melo and Guerra 2003).

Wild and cultivated Passiflora species generally present meiotic stability (Barbosa and Vieira 1997; Melo et al. 2001; Souza et al. 2003b). In contrast, hybrids exhibit meiotic instability, but sexual hybrids seem more stable than somatic ones (obtained by protoplast fusion) (Soares-Scott et al. 2003). Pollen viability in cultivated species is 80%, whilst in wild species it varies from 78.2% to 99.5% (Barbosa and Vieira 1997; Souza 2002).

7.2.4 Agricultural Status

Since prehistoric times, people have consumed fruits of Passiflora species. Passionfruits are rich in minerals, alkaloids, flavonoids, caretonoids and vitamins A and C, indicating their nutritional quality. More than 80 Passiflora species produce edible fruits (Coppens d’Eeckenbrugge 2003). About half of these species belong to subgenus Decaloba and their fruits are generally small, rarely more than 1.5 cm in diameter, so they are only collected on wild plants in rural areas. Although fruits produced by subgenus Decaloba are rich in qualitative terms, their potential for the development of economic fruit crops is very limited. Species with larger fruits belong to subgenus Passiflora. More precisely, species of economic potential for fruit production belong to supersections Passiflora, Laurifolia and Tacsonia (Fig. 7.2). Table 7.2 provides a list of 48 species of which more than 20 have been cultivated as fruit crops at very different scales.

Fig. 7.2
figure 2_7

Diversity of edible fruits of subgenus Passiflora. Photo from Geo Coppens d’Eeckenbrugge

Table 7.2 Cultivated species of Passiflora a

Although passionfruits were widely known and grown by Native Americans (Patiño 1963), their development into commercial crops is fairly recent. Passiflora edulis f. flavicarpa (Fig. 7.3) was long overshadowed by the purple maracuja, published as a distinct taxon only in 1932 in the Flora of Hawaii, as an introduction from Australia (Degener 1932). For this country, Winks et al. (1988) mention that P. edulis was introduced in 1861; however, commercial development started only 60–70 years later, the yellow form and its hybrids becoming important only in the late 1950s. In 1951, Hawaiian plantings of this form amounted to less than five acres, from which a few vines were selected. By 1958, 1,200 acres were devoted to this crop as a basis of a well-established Hawaiian passionfruit juice industry. It was introduced in Venezuela in 1954, and trials took place in the Cauca Valley (Colombia) in 1963 with materials from Hawaii, Brazil and Venezuela (Torres and Giacometti 1963; Morton 1967). Among Andean passionfruits, Castañeda (1956) only mentions sweet granadilla (P. ligularis Juss.) as a highly demanded fruit while the banana passionfruit (P. tripartita var. mollissima (Kunth) Holm-Nielsen and Jørgensen) is only mentioned for its organoleptic qualities and its absence from two of the three Colombian cordilleras. The first experiment for the commercial cultivation of this species was reported by Jaramillo (1957).

Fig. 7.3
figure 3_7

Passiflora edulis Sims. f. flavicarpa Deg. Photo from Geo Coppens d’Eeckenbrugge

Nowadays, 15 species are cultivated and a dozen are commercialized. The maracuja, P. edulis Sims, is the most important one. Its yellow form is cultivated in most tropical lowlands, particularly in tropical South America. Brazil is by far the main producer and consumer, with about 70% (half a million tons) of the world production. Ecuador is the main provider for the international market of passionfruit concentrate (Linden 2007). Colombia, the third main producer, has an appreciable home market and its contribution to international trade is variable. The purple form of P. edulis, P. edulis f. edulis (Fig. 7.4), originating from higher latitudes in southern South America, is confined to cooler climates. It has shown adaptation to tropical highlands, particularly in the Andes and in eastern Africa, as well as to subtropical climates, in Australia. The purple maracuja is present in small quantities on the international market, mostly provided by East African countries. While the yellow maracuja appears to be a cultigen, the purple form still exists in the wild in its birth place. Interestingly, the closest relative of P. edulis is P. incarnata, which was a common fruit crop of North America in Pre-Columbian times. The sweet granadilla, P. ligularis (Fig. 7.5), is cultivated for the fresh fruit market in the northeastern Andes, at moderate altitudes, and in Central America. Colombia is the main producer and consumer, and regularly provides small quantities at high price for the international market. The banana passionfruits, or curubas, are also widely cultivated in the Andes, above 2,000 m, and consumed in juices, sherbets and pastries. They mostly correspond to two cultigens, P. tripartita var. mollisssima (Fig. 7.6) and P. tarminiana Coppens and Barney. Passiflora cumbalensis (H. Karst.) Harms is commercialized at a very small scale around the city of Bogota in Colombia (Fig. 7.7). The ancient cultivation in home gardens of P. antioquiensis Karst. has been abandoned. Similarly, P. pinnatistipula (Harms) Killip is losing ground in High-Andean home gardens. The giant granadilla, P. quadrangularis L. (Fig. 7.8), is a cultigen present on the national markets in all tropical America. A close relative, P. alata Curtis, has been recently selected and developed in Brazil, for the fresh fruit market (Kavati et al. 1998). Among the section Laurifoliae of subgenus Passiflora, P. popenovii Killip (Colombia, Ecuador) (Fig. 7.9), P. nitida Kunth (northern South America) and P. riparia Mart. ex Mast. (Amazon) are sold in local markets. Passiflora nitida is often confused with P. laurifolia L., which occurs wild in northeastern South America, and is still, albeit rarely, cultivated in the Antilles.

Fig. 7.4
figure 4_7

Passiflora edulis f. edulis Sims. Photo from Geo Coppens d’Eeckenbrugge

Fig. 7.5
figure 5_7

Passiflora ligularis Juss. Photo from Geo Coppens d’Eeckenbrugge

Fig. 7.6
figure 6_7

Passiflora tripartita var. mollissima (Kunth) Holm-Niels. & Jørg. Photo from Geo Coppens d’Eeckenbrugge

Fig. 7.7
figure 7_7

Passiflora cumbalensis (H. Karst.) Harms. Photo from Geo Coppens d’Eeckenbrugge

Fig. 7.8
figure 8_7

Passiflora quadrangularis L. Photo from Geo Coppens d’Eeckenbrugge

Fig. 7.9
figure 9_7

Passiflora popenovii Killip. Photo from Geo Coppens d’Eeckenbrugge

The peculiar beauty of their flowers has made passionflowers particularly interesting for the industry of ornamental plants. The development of the ornamental potential of passionflowers is still confined to circles of amateurs, who have developed a number of interspecific hybrids. The earliest was obtained by Thomas Milne in 1819. Since then, more than 400 hybrids have been generated to obtain more showy flowers (Peixoto 2005). Breeders have also used tetraploidization to obtain larger flowers (Fischer 2004). This spectacular diversity of shapes and colors has not been translated into commercial success, mostly because potential market is essentially concentrated at high latitudes. Indeed, as a garden ornamental, the most successful species is P. caerulea L. Its winter hardiness has allowed it to take a relative importance under mild temperate climates. Thus, much remains to be done to develop passionflowers for the tropical garden (Abreu et al. 2009).

Passionflowers are also known for their medicinal properties (Table 7.3). In particular, P. incarnata is exploited in the pharmaceutical industry for its sedative, antispasmodic and analgesic properties (McGuire 1999; Dhawan et al. 2004).

Table 7.3 Theraupetic properties of Passiflora speciesa

Several common species of Passiflora have become weeds, as is the case of P. capsularis L., in the coffee growing zone of Colombia. Some have extended their range to other continents early after the great voyages of the XVI and XVII centuries, and some are even considered invasive. Passiflora foetida L. and P. suberosa have become pantropical weeds, for upland rice and other field crops. Nevertheless, the competitive ability of P. foetida has been used to control erosion and the grass Imperata cylindrica in coconut fields in Philippines and in sweet potato fields in Papua New Guinea, while their young leaves are used in Surinam and Java as a vegetable (Waterhouse 1994). Several species of supersection Tacsonia have naturalized in New Zealand (Heenan and Sykes 2003). The most surprising case is that of P. tarminiana, a cultigen showing limited potential of escaping cultivation in the northern Andes, but invading protected areas in Hawaii, New Zealand and Australia. The considerable impact of P. tarminiana on native floras generated prevention against the introduction of passionflowers in many tropical countries. The problem of invasive passionflowers is treated more thoroughly in Sect. 7.9.

7.3 Conservation Initiatives

7.3.1 Evaluation of Genetic Erosion

Passiflora species are essential components of natural and semi-natural habitats and are critical to maintain ecosystem health. Their conservation and sustainable use are vital for improving their agricultural production and preserving the environment. Furthermore, the disappearance of Passiflora species from the ecosystem would entail the loss of other organisms depending on them, such as the butterflies Heliconius and many nectar feeding insects and birds.

Characterization and evaluation of the genetic diversity of wild and cultivated Passiflora species are necessary to identify and prevent genetic erosion. The genetic diversity of different Passiflora species has been studied using various molecular tools as isoenzymes (Segura et al. 1998, 2003b, 2005; Fore and Spira 2002; Tague and Fore 2005), random amplified polymorphic DNA (RAPD; Fajardo et al. 1998, Aukar et al. 2002; Crochemore et al. 2003; Viana et al. 2003; Bellon et al. 2007; Junqueira et al. 2007), cpDNA (Sanchez et al. 1999), amplified fragment length polymorphisms (AFLPs; Segura et al. 2002; Ocampo et al. 2007) and sequences of chloroplast and nuclear regions (Lorenz-Lemke et al. 2005; Koehler-Santos et al. 2006a, b).

These genetic studies show that wild and cultivated Passiflora species have, in general, high intraspecific and interspecific genetic variability. Only one species, P. elegans Mast., has been reported to have low genetic variability (Lorenz-Lemke et al. 2005). The genetic diversity of some cultivated and wild species has been compared (Segura et al. 1998, 2002, 2003b, 2005; Bellon et al. 2007). In the group of banana passionfruits, cultigens (P. tripartita var. mollissima and P. tarminiana) show very limited variation, contrasting with the high diversity of wild species (P. mixta L. f.) (Segura et al. 1998, 2002, 2003b, 2005). In contrast, Bellon et al. (2007) report no differences in genetic diversity among wild and commercial accessions of P. edulis. Ocampo et al. (2007) emphasize a surprisingly high diversity in the same species. However, the susceptibility of cultivated species to many pathogens is a probable indication of genetic erosion (Vieira and Carneiro 2004). Furthermore, human activities can facilitate new encounters between plants and pathogens, causing damage in wild populations (Webster et al. 2007).

There are in situ and ex situ approaches to conserve biodiversity and different ways of using genetic resources. We review the different attempts to conserve germplasm of Passiflora.

7.3.2 In Situ Conservation

In situ approaches for the conservation of Passiflora species have been mainly limited to protected areas. These measures may be appropriate for particular species, such as those located in sparsely populated rain forests, but they may be insufficient when human pressure is high. In Colombia, Ocampo et al. (2007) showed that most of the genus diversity was particularly concentrated on the hillsides of the coffee growing zone (81% of the total 167 species), i.e., in a highly populated and perturbated region, out of protected areas. Several common species thrive in disturbed habitats, so their distribution even benefits from the human pressure. However, most of the 58 Colombian endemic species are found in these increasingly deforested hillside areas, which puts them at a high risk. Furthermore, the Passiflora diversity hotspots correspond strikingly well to particular ecotopes, hence the necessity of integrating biodiversity conservation measures in the general management of agricultural activities, recreational areas and water resources (Ocampo et al. 2007). Applying IUCN criteria, Ocampo et al. (2007) consider that 71% of the Colombian Passifloraceae are under some degree of threat, 10% being critically endangered, 6.1% vulnerable and only 16% of the species are placed in the two “least concern” and “near threatened” categories, indicating the necessity of establishing conservation measures, at least in this country.

A proper documentation of Passiflora genetic diversity in relation with the geographic distribution of species is essential. Jørgensen et al. (2009), from the Missouri Botanical Garden (USA), have conducted a study to establish the geographical distribution of Passiflora species based on herbarium data and subsequently estimated their threat status. Preliminary results focused on subgenus Decaloba, Jørgensen (2009) found that 40% of these species are vulnerable or endangered, while another 40% are not currently threatened.

For a same level of habitat fragmentation or perturbation, the erosion risk varies according to several factors, such as the situation of the species in the ecological succession, its reproductive biology and its dependence on particular pollinators and dispersors. For instance, high genetic diversity at the local level with little differentiation among regions in P. incarnata, an early successional species, suggests that this genetic structure is related to founder effects and long-distance pollen-mediated gene flow (Fore and Spira 2002; Tague and Fore 2005). This situation contrasts with that of P. mixta, a similar weed that is affected by the fragmentation of the habitat of its pollinator, the hummingbird Ensifera ensifera (Lindberg and Olesen 2001).

7.3.3 Ex Situ Conservation

Conservation of Passiflora genetic resources out of their natural habitat is mainly done in botanical gardens and germplasm collections, maintained by national institutes. More than 50 Passiflora germplasm collections are maintained around the world (Ferreira 2005) and the Passiflora accessions in collections have increased in recent years, from 524 accessions in 1994 to 1,235 in 2004 (Ferreira 2005).

Figure 7.10 illustrates the information available of Passiflora germplasm obtained from several databases. A total of 2,203 Passiflora accessions are currently reported for a total of 159 species; 19% of these accessions correspond to the widely cultivated P. edulis and 24.7% are unclassified. When compared to other collections of crop genetic resources, the proportion of wild species is much more important (28.3%). The global effort may appear important; however, it must be related to the number of species, and also to their potential usefulness for breeding. Thus, Fig. 7.10 should be taken with caution. Many accessions are usually exchanged among collections, so duplication may be relatively important. While this is positive in terms of safety, the genetic diversity may be less important than suggested by the numbers.

Fig. 7.10
figure 10_7

Percentage of accessions and number of species represented in germplasm collections in each country. The number of accessions per country is also indicated. The information was obtained from the Germplasm Resources Information Network of the USDA-ARS (GRIN) (USDA 2010), the New World Fruit Database of the Biodiversity International that save information of many ex situ germplasm collections worldwide (Biodiversity International 2009), the Kew Botanical Garden and the Missouri Botanical Garden. The information about Brazil collections have been obtained from Ferreira (2005). Only data from the first ten countries with more abundant collections are illustrated in this chart

The living plant collections of botanical gardens represent a significant source of conserved genetic resources. For instance, the Royal Botanic Gardens of Kew (UK) holds in their greenhouses 113 living passionflowers representing 72 different species and 21 hybrids or ornamental varieties. The Missouri Botanical Garden conserves 24 individuals of 15 species and 2 hybrids. In vivo plant collections with a scientific research purpose are also maintained by academic institutions, such as the Lawrence Gilbert’s laboratory at University of Texas in Austin (USA), which maintains 101 species and 30 hybrids (Gilbert 1998). Unexpected sources of ex situ genetic resources of Passiflora are the living collections maintained in greenhouses by plant societies and amateurs (Passiflora Society International 2008). Private societies regrouping amateurs are created for the sole purpose of studying, collecting and preserving Passiflora seeds and plants. Many members of these societies have large in vivo collections with more than a hundred species collected around the world. These societies have also seed banks available for their members. In France, the collection maintained by C. Houel, including 359 accessions from 241 species, has received the status of national collection (Houel 2009).

An appropriate documentation of germplasm is indispensable to make this stored diversity useful to everyone. The germplasm collection should be systematically classified and evaluated. For instance, Crochemore et al. (2003) used molecular markers to evaluate the samples of P. edulis obtained from the germplasm collection of the Instituto Agronômico do Paraná (IAPAR) in Brazil, facilitating by this way the identification of misclassified samples and the classification of unidentified ones. In addition, it is necessary to continuously increase the genetic variability of germplasm collection. This initiative requires identification of the most suitable populations. For instance, genetic studies in banana passionfruits showed that it would be preferable to collect samples of P. tripartita var. mollissima in southern populations (Ecuador and Peru) with high genetic variability, rather than in northern populations (Colombia and Venezuela), in order to enlarge the genetic diversity of these species in germplasm banks (Segura et al. 1998, 2002, 2003b, 2005).

7.4 Role in Elucidation of Origin and Evolution of Allied Crop Plants

The evolution of cultivated passionflowers has not been extensively investigated. However, despite the difficulties and inconsistencies in the taxonomic treatment of Passiflora, most of them show a close morphological relationship with several wild species, suggesting strong evolutionary affinities. The best approach is to identify the likely crop gene pools around these cultivated species. The closest relatives would be the taxa in the primary gene pool; slightly more distant relatives would constitute the secondary gene pool and so on.

The situation is different for ornamental plants, where vegetative propagation is the normal rule. Many crosses have been attempted according to the amateur creativity and potential genitors in the collection, with some surprising successes between what seem distantly related species. However, most of these hybrids set no fruits or empty fruits.

The present section is essentially devoted to the relatives of established fruit crops, all from subgenus Passiflora (Table 7.2). In each case, species exhibiting strong morphological affinities are considered. Elements from molecular studies are added at the end of each section, when available.

7.4.1 Series Passiflora (Supersection Passiflora)

P. edulis f. flavicarpa (yellow maracuja), P. edulis f. edulis (purple maracuja) and P. incarnata (maypop) can only be distinguished by a very limited number of morphological traits. They are much more clearly differentiated in their ecology. The first one is a vigorous tropical cultigen producing large yellow fruits. Although its origin is most probably Amazonian or periamazonian, the precise origin of yellow maracuja remains mysterious since it was developed out of South America. According to Vanderplanck (1991), the first documented cultivation was based on the seeds of a few fruits bought in a London market and sent to Argentina. Their descent would have been sent to the United States Department of Agriculture, and redistributed to Australia, New Zealand and later to Hawaii. This very narrow genetic base was somewhat widened through crosses with P. edulis f. edulis. The yellow maracuja shows strong self-incompatibility, even if a few pseudo-self-compatible clonal cultivars were selected in the 1950s. When reintroduced into South America, these selections were newly in contact with their natural pollinators and propagated again through seeds; so this trait was lost. Many cultivated populations still show the segregation for fruit color testifying that the yellow form was once crossed with the purple form. In contrast with our ignorance for the yellow maracuja, the origin of P. edulis f. edulis is clearly in southern South America, where it is still commonly found growing spontaneously, under subtropical conditions. Its adaptation to cool conditions allowed it to naturalize in other subtropical areas as well as in tropical highlands. The purple maracuja produces a fruit smaller than the yellow one, and appears at least partially self-fertile. The cross between the two forms of P. edulis is only fertile when the yellow form is used as the male genitor. Hybrids are intermediate and vigorous (Nakasone et al. 1967) with a regular meiosis, but have a lower chiasma frequency (Beal 1975).

The maypop was a common North American fruit crop in Pre-Columbian times. However, it is no more cultivated commercially, except for ornamental or pharmaceutical purposes. Crosses with P. edulis have shown limited success. Several authors report the obtention of fertile hybrids (Beal 1972; Anderson 1976; Winks et al. 1988), with more fertile F2 and backcross progenies (Kajewski 1941) while others had to double their hybrid genome to restore fertility (Knight 1991). Thus, independent of the strong affinity between the two species, P. incarnata must be considered to be part of the secondary or tertiary gene pool of P. edulis. Similarly, P. edulis has been crossed with P. cincinnata Mast., with variable success (Howell 1989). The hybrid of Howell did not flower, while those obtained by Ruberté-Torres and Martin (1974) were intermediate and produced an edible fruit. Afterwards, species of different supersections, such as P. caerulea and P. manicata (Juss.) Pers., have also been crossed successfully with P. incarnata or P. edulis (Escobar 1985; Vanderplanck 1991) showing that morphological and genetic affinity are poor predictors of crossability in Passiflora.

7.4.2 Series Tiliifolia (Supersection Laurifolia)

This series includes two groups of species showing particularly large morphological and genetic similarity. The first one involves the cultivated P. ligularis, as well as the wild P. tiliifolia L. (Fig. 7.11) that can be often observed in close vicinity, in the same highland conditions. Despite its striking resemblance with P. ligularis, P. tiliifolia is rarely cultivated. Other very similar species are P. palenquensis Holm-Niels. & Lawesson, and P. fieldiana, which grow under very different ecological conditions (low to mid elevations). A close relative (P. af. palenquensis) has been observed under cultivation in the Pacific lowlands of western Colombia. All these species produce fruits that are round to ovoid, tapering at apex, with a thin, smooth and brittle epicarp and sweet grayish arils, easy to consume in hand. Hybrids have not been reported between these species.

Fig. 7.11
figure 11_7

Passiflora tiliifolia L. Photo from John Ocampo

The second group involves the stone granadilla, P. maliformis (Fig. 7.12), and its close relatives, P. platyloba, P. serrulata Jacq. and P. multiformis Jacq. All of them exhibit yellow stipules, a same fruit shape and similar habit and habitat. In fact, intraspecific variation in these four species is so wide, as compared to presumed interspecific variation, that species delimitation should be re-examined objectively, on both morphological and genetic grounds. The only morphological trait that discriminates them is leaf lobation, but it may be highly variable within many species. The two groups are slightly differentiated in the molecular study of Yockteng and Nadot (2004a). Amplified fragment length polymorphism (AFLP) markers show a close affinity between P. ligularis and P. tiliifolia, but no particular affinity with P. maliformis (Segura et al. 2002; Ocampo et al. 2007).

Fig. 7.12
figure 12_7

Passiflora maliformis L. Photo from John Ocampo

7.4.3 Series Laurifoliae (Supersection Laurifolia)

The morphological similarity prevailing within this series made the identification and classification of species particularly difficult (Killip 1938). Several species may coexist in the same region such as P. nitida, P. laurifolia, P. crenata Feuillet & Cremers and P. cerasina Annonay & Feuillet in the lowland forests of French Guiana or P. nitida, P. nigradenia Rusby, P. riparia, P. fernandezii Escobar, P. chaparensis Vásquez and P. venusta Vásquez & Delanoy in Bolivia (Vásquez et al. 2007). The origin of P. laurifolia populations cultivated in the Antilles has to be clarified, and their relationship to wild conspecific materials from South America. In the northern South America, the most commonly cultivated Laurifoliae species is P. nitida, while P. laurifolia seems to occur only in the wild in the Guianas. No wild representative of P. popenovii has been reported, and the species is so difficult to propagate through seeds that vegetative propagation is often preferred. Its vernacular name “granadilla de Quijos” refers to an ancient Native American culture, and also a province of Ecuador, where it may have its origin (Patiño 1963). The situation is different for P. nigradenia, which is still found in both wild and cultivated conditions in Bolivia (Vásquez 1998).

The species of series Laurifoliae present a particular high potential. Indeed, they not only produce a highly aromatic fruit pulp, but also they exhibit at the same time a high genetic diversity and a high uniformity in most essential traits, so they could allow the breeder to create original and adapted cultivars without facing the problems of an excessive genetic segregation. However, their taxonomy should be first clarified, delineating species on an objective basis and establishing their affinities. The exceptional capacity to grow on flooded soils and their resistance to soil parasites is of outmost interest for developing rootstocks or transferring the corresponding genes to other passionfruit species. However, these species have several traits reflecting their lack of domestication. Their seeds tend to germinate very irregularly, often with marked latency, and they often allocate too many resources to vegetative growth. Thus, they cannot be accommodated on artificial supports, because they could invade other fruits trees in the orchard. In addition, their harvest season is short and impredictable. These problems should be addressed in order to fully develop their economic potential.

7.4.4 Series Quadrangulares (Supersection Laurifolia)

In contrast with the situation prevailing in other series of supersection Laurifolia, the high morphological similarity of P. alata and P. quadrangularis reflects their close genetic affinity and fertile hybrids are easily obtained (JC de Oliveira personal communication). Their progeny segregates for resistance traits (Oliveira et al. 1996). Both species have been used successfully in crosses with P. caerulea (Vanderplanck 1991).

7.4.5 Supersection Tacsonia

The supersection Tacsonia comprises 61 species, 11 of which are cultivated and known with the name of banana passionfruits. The fruits of P. tripartita var. mollissima and P. tarminiana are present all the year on the Andean markets and a small quantity is even exported to Europe. Their thin leathery pericarp and generous orange, succulent arils ensure the highest pulp yields among passionfruits (around 60% for the two cultigens). Tacsonia species are differentiated by their large flowers presenting a long floral tube mostly pollinated by hummingbirds. In most of them, the hypanthium length varies between 7 and 13 cm, restricting the normal access to nectar to only one species, the sword-billed hummingbird, Ensifera ensifera (Snow and Snow 1980; Lindberg and Olesen 2001). As for this bird, their native distribution is limited to the Andean mountains between 1,400 and 4,000 m, from Venezuela to Chile.

Although this has not been established systematically, interspecific barriers are much more labile in this supersection as compared to supersection Passiflora. Spontaneous hybrids have been relatively frequently reported, involving the cultivated P. arminiana, P. tripartita var. mollissima, P. pinnatistipula as well as the wild P. tripartita, P. mixta, P. cumbalensis and P. antioquiensis (Killip 1938; Escobar 1981; Geo Coppens d’Eeckenbrugge personal observations). The most distinct species, P. manicata, has also been involved in artificial crosses, giving fertile hybrids (Escobar 1985). However, Schoeniger (1986) in an attempt to exploit the intercompatibility of P. tripartita var. mollissima with P. mixta and P. cumbalensis for the genetic improvement of the former, observed phenomena of unilateral incompatibility, loss of fertility, seed germination and general viability in advanced hybrid and backcross generations.

Segura et al. (2002, 2003a) have studied the relationships between cultivated species and some of their wild relatives with isozyme and AFLP markers. The four most common species, P. tarminiana, P. cumbalensis, P. tripartita and P. mixta were differentiated from the others. In the isozyme study, the accessions of the two latter species showed particular affinity, as they clustered together, following a more geographic than taxonomic pattern. A wider study (Segura et al. 2005) confirmed the existence of a gene flow between P. mixta and P. tripartita, as well as a higher genetic diversity in Ecuador, as compared to Colombia and Venezuela, two countries where the genetic diversity of the two cultigens appeared very low. The closer genetic affinity between P. tripartita and P. mixta seems consistent with their closer morphological affinity (Villacis et al. 1998); however, P. tripartita var. mollissima hybridizes easily with both P. mixta and P. tarminiana. The observation of meiosis in a few hybrids indicates a better chromosome pairing between the two cultigens than with the wild P. mixta (Olaya 2002). However, the two parental phenotypes of the cultigens are recovered in less than four generations in the progenies of their hybrids, suggesting very limited exchanges at the genome level. The nature of interspecific barriers among species of supersection Tacsonia is not clear yet.

7.5 Role in Classical and Molecular Genetic Studies

The genetics of species Passiflora is mainly known from cultivated species. Therefore, the genetics of wild species is not known except for a few population genetic studies. Data from gene actions, physiological pathways and genetic mapping are only available for the cultivated species P. edulis. However, the studies based on crop species constitute a reference for future research in wild species. They can even provide new tools, as could be the case of transferable microsatellite markers.

7.5.1 Population Genetic Studies

Genetics studies conducted in wild and crop Passiflora species reveal their genetic diversity, the geographic structure of populations and the relationship among species. Passionflowers generally present high interspecific and intraspecific genetic variability except for P. elegans (Fajardo et al. 1998; Segura et al. 1998, 2002, 2003b, 2005; Sanchez et al. 1999; Aukar et al. 2002; Fore and Spira 2002; Crochemore et al. 2003; Viana et al. 2003; Lorenz-Lemke et al. 2005; Tague and Fore 2005; Bellon et al. 2007; Junqueira et al. 2007).

Species belonging to the same supersections are in general closely related. For instance, P. elegans and P. actinia of supersection Stipulata are close relatives (Lorenz-Lemke et al. 2005), as well as P. tarminiana, P. tripartita var. mollissima and P. mixta of supersection Tacsonia (Segura et al. 1998, 2003b). The genetic proximity among species has facilitated interspecific hybridization events. Genetic studies also permit to know how historical geological, climatic and ecological conditions have affected the distribution of species. For example, the high intraspecific variability of P. actinia is structured along a north–south gradient in Brazil, where northern populations are more diversified than the southern. This study gives evidence of migration of northern populations of P. actinia to the south probably caused by a change in temperature and humidity in southern region at the beginning of the Holocene (11,000–10,000 years bp). Populations of P. elegans show a low genetic variability not structured geographically, probably as a consequence of severe bottleneck events during Pleistocene glacial stage (Lorenz-Lemke et al. 2005). The genetic variation of banana passionfruit species, P. tarminiana, P. tripartita var. mollissima and P. mixta, is structured along a south–north gradient giving evidence that their center of diversity is probably in the southern region of their range (Segura et al. 2005). Passiflora nitida also has a high genetic variability with a geographical structure and a higher variation among populations than within populations (Junqueira et al. 2007). The opposite is found in P. alata in which the greater variation is found within populations while a geographical structure is absent, possibly due to high levels of gene flow between populations (Koehler-Santos et al. 2006a). Genetic variability among populations of P. incarnata is lower than within populations. The strong gene flow between geographically distant populations, enhanced by the combined action of self-incompatibility and long distance pollen transfer by pollinators, explain the low differentiation among P. incarnata populations (Tague and Fore 2005).

The study of geographical structure of genetic diversity is essential to prioritize the conservation of genetically diversified areas.

7.5.2 Genetics of Self-Incompatibility

The genetic control of self-incompatibility in Passiflora has been studied since 1959 only in cultivated species (Bruckner et al. 2005). In the yellow passionfruit species (P. edulis f. flavicarpa), self-compatibility is controlled by the combination of a gametophytic and a sporophytic systems (Rego et al. 1999, 2000; Suassuna et al. 2003; Souza et al. 2006). The gametophytic system is regulated by the S gene that has multiple alleles. A single allele is expressed in the haploid pollen grain that germinates only in a diploid pistil, which does not express the same allele. The sporophytic system is regulated by a cluster of three genes encoding for proteins with a role in the reception and recognition of pollen (Hiscock and Tabah 2003). Pollen will not germinate on the diploid stigma of a flower that contains either of the two alleles of the sporophyte parent that produced the pollen.

Self-incompatibility of Passiflora species facilitates the outcrossing of individuals and promotes the creation of new genotypes. Therefore, it must be taken into account in breeding programs to improve crop species, ensuring sufficient diversity of S haplotypes to provide fertilization and good fruit production.

7.5.3 Organellar Inheritance

Besides, mitochondria and chloroplast in Passiflora are inherited by different mechanisms. Studies present evidence that mtDNA is maternally inherited in passionflowers, while the inheritance of cpDNA can be maternal, paternal or biparental (Muschner et al. 2006; Hansen et al. 2007). The biparental plastid transmission, suspected by Corriveau and Coleman (1988) on the basis of observations in epifluorescence microscopy, was confirmed by Do et al. (1992) in crosses involving the two forms of P. edulis. Only the chloroplasts of P. edulis f. flavicarpa were found in the reciprocal hybrids. In a cross with P. coccinea Aubl. as female parent, P. edulis f. flavicarpa also transmitted its chlorotype to the progeny. Mrácek (2005) observed that the plastids of P. menispermifolia Kunth. and P. oerstedii Mast. are biparentally transmitted. The heteroplasmy of hybrids also permitted to evidence that the progeny’s plastome–genome was incompatible with the parental genome of P. menispermifolia (Mrácek 2005). The heteroplasmy could cause many problems in phylogenetic inferences based on chloroplast gene sequences. In species of subgenus Passiflora, the cpDNA is reported to be inherited paternally or biparentally whilst it appeared maternally inherited in two species of subgenus Decaloba studied so far (Muschner et al. 2006; Hansen et al. 2007). Although the differential inheritance of plastids could bring new evidence to support the subdivision of these two subgenera, it underlines that interpretation of phylogenies based on cpDNA sequences has to be very cautious. Moreover, the phenomenon of heteroplasmy could explain the incongruences between phylogenetic studies based on chloroplast sequences (Muschner et al. 2003; Hansen et al. 2006). In addition, it is not possible to infer the paternal origin of the pollen and seeds of hybrid individuals (Hansen et al. 2007).

7.5.4 Construction of Genetic Linkage Maps

Linkage maps are useful for identification of important plant genes controlling simply and quantitatively inherited traits. They could facilitate the identification of genomic regions that might affect the variation of important Passiflora traits involved in fruit production, fruit quality and disease resistance.

In self-incompatible species, like many Passiflora species, linkage maps are constructed using a strategy known as two-way pseudo-testcross, based on monoparental dominant markers that segregate in a 1:1 proportion (Grattapaglia and Sederoff 1994). The final result is the generation of two individual maps, one for each parental genotype. This method was used to generate genetic linkage maps of P. edulis f. flavicarpa, based on random amplified polymorphic DNA (RAPD) and AFLP markers in three different studies (Carneiro et al. 2002; Moraes 2005; Lopes et al. 2006). The linkage maps were obtained from the same F1 population derived from a single cross between two clones of P. edulis f. flavicarpa, “IAPAR 123” (female parent) and “IAPAR 06” (male parent) (Carneiro et al. 2002; Moraes 2005; Moraes et al. 2005; Lopes et al. 2006).

Moreover, the development of specific microsatellites markers to yellow passionfruit (Oliveira et al. 2005) and the use of a maximum likelihood approach (Wu et al. 2002) permitted to generate a first integrative linkage genetic map using the same F1 population. In this study, individual parental maps were integrated into one map based on the segregation of codominant markers (microsatellites) and dominant markers (AFLP) in one or both the parents (Oliveira et al. 2008). The integrated map is more saturated in markers and the linkage groups are longer than those in individual maps, which will probably facilitate the mapping of crucial traits for the crop species.

7.5.5 Mapping Genes and Polygenic Clusters

Once the genetic linkage map has been developed, the next step is to identify or map the genes controlling the traits of interest by quantitative trait loci (QTL) mapping.

Fruit phenotype is, in general, the result of the interaction of multiple genes as well as environmental factors. Therefore, QTL mapping is necessary to determine the genetic architecture underlying the fruit phenotype. Several fruit characters in yellow passionfruit (P. edulis f. flavicarpa) show a wide genetic variability and also a high coefficient of heritability (52.6–83%) (Moraes 2005; Moraes et al. 2005). Using an AFLP linkage map, Moraes (2005) mapped genomic regions associated with eight fruit-related traits (fruit yield, number of fruits, average fruit weight, average fruit length, average fruit width, percentage of pulp, soluble solids content and average of fruit size). A total of 41 QTLs were mapped with four to seven QTLs by trait (Moraes 2005). Most (90%) of the QTLs have small or medium effects on fruit traits explaining 15% of the trait’s phenotypic variance. These values are higher than those found in some recent studies conducted in other plant species such as cucumber (Yuan et al. 2008) and melon (Obando et al. 2008). The expression of some traits such as fruit yield and the number of fruits are correlated and probably controlled by the same genes or proximal genes. QTLs for important fruit quality traits in passionfruit were, therefore, located on the linkage map (Moraes et al. 2005). The key markers most closely linked to these QTLs can be further developed to provide tools for breeding and selection in Passiflora. A future QTL analysis in wild species would permit to locate genes coding desirable fruit traits that can be incorporated into cultivated species.

QTL mapping in Passiflora has also been used to identify genomic regions related to the resistance to one of the common pathogen Xanthomonas campestris pv. passiflorae (Lopes et al. 2006). This bacterial pathogen causes a disease in Passiflora that leads to a premature death of the plant. A unique quantitative resistance loci (QRL) was found in genetic linkage map explaining only 15.8% of the total phenotypic variance in the segregating population (Lopes et al. 2006). The results in Passiflora are not sufficient (Lopes et al. 2006) compared to the studies in other plant species that detected several QRLs explaining a large percentage of phenotypic variance for resistance to X. campestris (Studer et al. 2006; Soengas et al. 2007). However, it is necessary to continue this first effort in order to progress in the detection of genes controlling the resistance against pathogens in particular in wild resistant species.

7.5.6 Assessment of Gene Actions and Physiological Pathways

Some of the pathways studied in P. edulis are the ethylene pathway and its role in growth and development of the plant (Arjona and Matta 1991; Shiomi et al. 1996) and the jasmonate pathway and its role in defense response (Siqueira et al. 2008). The inositol phosphate biosynthesis pathway has also been studied. This pathway produces signal molecules having vital roles encompassing regulation of many processes indispensable to organism homeostasis (Abreu and Aragao 2007). One enzyme implicated in this pathway has been detected not only in P. edulis but also in P. eichleriana Mast., P. caerulea, P. nitida and P. coccinea Aubl.

Pelegrini et al. (2006) also described a new defense peptide Pe-AFP1 (Passiflora edulis antifungal peptide-1) that is a pathogen inhibitor. In vitro, this defense peptide inhibits the development of the filamentous fungi Trichoderma harzianum, Fusarium oxysporum and Aspergillus fumigatus but not of Rhyzoctonia solani, Paracoccidioides brasiliensis and Candida albicans (Pelegrini et al. 2006). The discovery of Pe-AFP1 can lead to the development of antifungal drugs against human and plant diseases and of transgenic plants resistant to fungal pathogens. Besides, these results are promising to conduct studies in the detection of new plant defense peptides in other Passiflora species, particularly in species already described as resistant to pathogens.

7.5.7 Host–Parasite Interactions

Species of Passiflora, especially crop species, suffer different diseases caused by virus, bacteria and fungi. In Table 7.4, the most common pathogens are listed. New descriptions of pathogens affecting Passiflora species appear frequently (Parry et al. 2004; Baker and Jones 2007; Tang et al. 2008; Villalobos et al. 2009). Some of these diseases produce considerable losses in passionfruit production. The harmful diseases are the fruit woodiness caused by the virus Passiflora woodiness virus (PWV) and cowpea aphid-borne mosaic virus (CABMV), the anthracnose, scab and septoriose caused, respectively, by the three fungi, Colletotrichum gloeosporioides, Cladosporium spp. and Septoria passiflorae and bacteriosis caused by the bacteria Xanthomonas axonopodis pv. passiflorae.

Table 7.4 Diseases affecting Passiflora speciesa

Larger transportation linked to human activity allows pathogens to encounter potential new hosts and can thus promote the emergence of new epidemics when pathogens could adapt to these novel hosts. Rarely these pathogens are specific to one species and thus they have the capacity to shift from a primary host to a new host species. For example, in Australia the Passiflora woodiness virus (PWV) described firstly in the indigenous species P. aurantia G. Forst. is now also infecting introduced species such as P. edulis, P. caerulea and P. foetida (Webster et al. 2007). Other examples are the infection by the Passiflora latent virus of the non-native species P. tarminiana in New Zealand (Tang et al. 2008) and the infection by fungus Phytoplasma spp. of introduced species P. edulis in Costa Rica (Tang et al. 2008; Villalobos et al. 2009). Pathogen species described initially in a phylogenetically distant host plant can also be found in species of Passiflora. For instance, the bean yellow mosaic virus infects now plants of P. caerulea in Italy (Parrella and Castellano 2002) and the worldwide Citrus tristeza closterovirus (CTV) infects the wild species P. gracilis J. Jacq. ex Link (Brunt et al. 1996). Closely related virus can develop the same disease symptoms in Passiflora plants. The fruit woodiness has been described to be caused primarily by only the PWV virus. However, Nascimento et al. (2006) found that the disease is also caused by the related CABMV potyvirus. While Brazilian plants are primarely infected by the CABMV, the Australian and Asian plants are mostly infected by the PWV.

Strategies of selection could be performed to choose the more resistant genotypes of cultivated species in order to reduce the susceptibility to diseases. Among the yellow passionfruit, Martins et al. (2008) found some genotypes moderately resistant to Colletotrichum gloeosporioides. Moreover, the resistance to the scab caused by the fungus Cladosporium cladosporioides has been enhanced by the selection of more resistant individuals of yellow passionfruit (Santos et al. 2008). However, the variability of resistance to different diseases of cultivated species appears in general to be very low (Junqueira et al. 2003).

Therefore, the identification of resistant species is essential to increase the variability of resistance in susceptible cultivated species by breeding strategies. The numerous wild species of Passiflora are a potential source of genotypes resistant to particular pathogens. Several wild species are resistant to the most common pathogens such as P. actinia that is non-susceptible to the PWV, the CABMV and the fungus Colletotrichum gloeosporioides (Junqueira et al. 2005) (Table 7.4). The feasibility to produce hybrids from interspecific crosses enhances the probability of obtaining resistant varieties of crop species (see Sect. 7.6.4).

7.6 Crop Improvement Through Traditional and Advanced Tools

7.6.1 On-Farm Selection and Genetic Resource Management

Passionfruits’ genetic improvement owes much to the efforts of the native growers in South America. For most species, even those of economic importance, many farmers still select a few good-looking fruits from higher yielding vines to establish the next orchard. In banana passionfruit, some growers select those seeds in the median part of the fruit. The intensity of selection is relatively weak particularly if the crop cycle is long (about 10-year-old banana passionfruit orchards). Growers may detect interspecific hybrids and favor them in their orchard. In Colombia, in a valley where P. tripartita var. mollissima cultivation has recently been established, the growers favor hybrids with P. mixta in an attempt to reduce the impact of anthracnose on fruit appearance. Other farmers bring spontaneous P. mixta × P. tripartita var. mollissima from the wild to grow them in their own home garden. Conversely, conservative farmers often discard those seedlings exhibiting particularly thin leaf lobes, indicative of hybridization with P. mixta, because hybrids tend to bear less fruits. Similar situations may exist in other cases where cultivated and wild species look very similar; some farmers tolerate wild vines of P. tiliifolia on the borders of their P. ligularis plots. Partial sympatry of similar cultivated and wild species also occurs in the series Laurifoliae in the lowlands.

As passionfruits have become more important commodities and even local markets have offered higher prices for more standardized products, plot size and planting densities have increased, favoring, unfortunately, the development of a wide cohort of pests and diseases. Although the shortening of crop cycles benefits the higher yield and better sanitary conditions of young vines, it increases both the costs of establishment of the crop and the need to get a higher and faster economic return. Such intensification has necessarily run parallel to the development of formal breeding in the passionfruit crops of wider importance. Local strains have been developed through mass selection for the sweet granadilla and banana passionfruit, while the development of yellow maracuja cultivars has been shared by the public and private sectors.

7.6.2 Development of Clonal Cultivars

Yellow maracuja breeding was initiated in Hawaii to improve yield and quality of the crop for the juice industry. Yields were multiplied fourfold, and juice yield improved from 25% to 35% (Morton 1967). To face the increasing phytopathological constraints of the crop, breeders attempted to widen its depleted genetic basis through hybridization with the purple form (Nakasone et al. 1967). Selection was most often carried out on an individual basis and the selected elite material was propagated clonally, which led to prefer pseudo-self-compatible genotypes (Fouque and Fouque 1980). When the elite material was strongly self-incompatible, cross-pollination had to be organized by interplanting distinct clones in the orchard (Knight 1972; Ito 1978).

In the context of clonal propagation of planting materials, another solution was attempted in Australia by grafting the plants on rootstock resistant or tolerant to the most severe soil pathogens. This practice imposed on the development of two parallel programs – a breeding program for scions and a rootstock program – without solving the problem of viral transmission via the planting materials (Coppens d’Eeckenbrugge et al. 2001b). Distinct species have been tested as rootstocks for P. edulis. Passiflora incarnata and P. caerulea were appreciated for their general resistance to soil pathogens as well as cool temperature at subtropical latitudes, but the latter was discarded because of its abundant suckering. Hybridization was attempted between both species and P. edulis f. flavicarpa for obtaining of better rootstock. Interspecific crosses involving P. incarnata were used as well for the scion, in order to improve the cold and virus resistance of commercial fruit types (Winks et al. 1988). However, their outcome is not clear, as the use of clones and pure P. edulis f. flavicarpa rootstock is still recommended in Australia (Department of Employment Economic Development & Innovation 2005).

7.6.3 Synthetic Population and Hybrid Breeding

When the Hawaiian selections of yellow maracuja came back to their home continent, they returned to sexual propagation, which had two major consequences. First, seed propagation reduced the impact of viral diseases as this planting material was not contaminated. Second, a higher level of agromorphological variation appeared in the crop, notably the segregation for fruit colors, still observed in many South American populations, and also in all important agronomical traits. Considerable genetic gains for yield and fruit size and quality were made possible, first through mass selection, then through the development of synthetic varieties (see Meletti et al. 2005 for a review), initially composed by selected clones (Maluf et al. 1989) and then by half-sib or full-sib progenies (Meletti 1998). Early producing cultivars were developed for the fresh fruit market and for the juice industry, showing yields up to 50 tons/ha, pulp yields close to 50% and soluble solids around 15–16°Brix (Meletti et al. 2000). However, no similar success was obtained in terms of disease resistances (Vieira and Carneiro 2004), which have a considerable impact on production, mostly because no sources of appreciable resistance/tolerance had been identified within P. edulis. A similar partial success allowed the development of P. alata from a home garden or from the wild to a commercial crop for the fresh fruit market, through selection and standardization of fruit shape (Kavati et al. 1998). As limitations in yield and quality had been overcome, the focus of breeding efforts shifted towards genetic resistances, through the characterization of other species (see Sect. 7.5.7) and introgression programs (Sect. 7.6.4).

In Brazil also, the yellow and purple forms of P. edulis have been crossed to generate varieties with better fruit qualities. Although these varieties are, in general, highly susceptible to diseases (Vieira and Carneiro 2004), they permitted the selection of plants with good fruit yield and also some degree of resistance to bacteriose and anthracnose (Junqueira et al. 2005).

7.6.4 Introgression of Traits Through Interspecific Hybridization

The search of resistance traits is the most frequent reason for attempting many different crosses and the subsequent introgression at the interspecific level. A few projects have been justified on other grounds, such as adapting the crop to temperate latitudes and/or to smaller pollinators through a reduction in flower size, or introducing self-compatibility to improve pollination efficiency (Knight 1991).

Some species seem to be amenable to hybridization more easily than others, as in the case for P. caerulea. Self-incompatibility of many cultivated species, such as P. edulis and P. incarnata, could be a functional component of interspecific incompatibility, explaining cases of unilateral incompatibility as in the cross between the two forms of P. edulis (Coppens d’Eeckenbrugge et al. 2001b). According to the difficulty of particular crosses, the production of interspecific seeds has involved controlled crosses or more sophisticated techniques, such as hormones, to retard floral abscission or/and intra- and interspecific double pollination (Payan and Martin 1975). The frequent sterility of the resulting F1 hybrids poses a further problem. Knight (1991) doubled the chromosome number of hybrids between P. incarnata and the two forms of P. edulis to restore their fertility. The subsequent tetraploid progenies were self-incompatible, which is consistent with the sporophytic system, remaining functional in polyploids. Their fruits gave a pleasant juice, despite a relatively clear color, with characteristics closer to maracujas than to maypops (Senter et al. 1993). Knight concluded that they had promise as a new fruit crop for warm–temperate zones. However, the only documented outcome today is the ornamental cultivar “Byron Beauty” derived from this hybrid (Knight et al. 1995).

Junqueira et al. (2005) attempted to transfer the resistance of P. setacea DC. against the anthracnose and PWV to P. edulis f. flavicarpa by introgression. The initial cross is fertile in both directions. The F1 hybrids appear vigorous, closer to the P. setacea parent in many respects, and resistant. However, they face pollination problems related to abnormal floral morphology. Four backcross generations were obtained using the recurrent genitor as pollen-donor, allowing the recovery of a P. edulis phenotype. However, genetic resistance to the PWV seems to be inherited in a quantitative manner, diminishing with successive generations. The BC3 still showed resistance to anthracnose; however, fruit yield was very low in comparison to susceptible cultivars, probably in relation to their very long androgynophore. The same authors report very similar results using P. coccinea instead of P. setacea, with hybrid vigor, general resistance transfer, abnormal floral proportions and very low fruit set in the F1 and BC1 generations. When exploring the potential of the cross with P. caerulea, they obtained sterile hybrids with good resistance to bacteriosis and anthracnose, but susceptible to the PWV and attractive to caterpillars of Dione butterflies. Fertility was partly restored in the BC2 generation. Many more such crosses within subgenus Passiflora have been attempted by the same group, including triple hybridizations. The compatibility of species has also been evaluated and some appear totally compatible, producing fertile seeds (Junqueira et al. 2005). For instance, the widely cultivated P. edulis f. flavicarpa seems totally compatible with P. glandulosa and very compatible with P. coccinea (79.2%) and P. setacea (85.7%) but it is totally incompatible to P. actinia and very incompatible with P. caerulea (8.6%) (Junqueira et al. 2005). Recently, 17 interspecific F1 hybrids were generated from the crosses P. laurifolia × P. nitida, P. edulis f. flavicarpa × P. coccinea, P. caerulea × P. amethystina J.C. Mikan, P. glandulosa Cav. × P. galbana Mast., P. coccinea × P. actinia, P. glandulosa × P. edulis f. flavicarpa, P. sidaefolia M. Roemer × P. actinia, P. galbana × P. actinia, F1 (P. coccinea × P. setaceax P. coccinea), F1 (P. coccinea × P. setacea) × P. mucronata Lam., P. eichleriana × P. gibertii N.E. Br., P. galbana × P. edulis f. flavicarpa, P. glandulosa × P. edulis edulis, P. glandulosa × P. sidaefolia, P. coccinea × P. setacea. Their success was confirmed using RAPD markers (Junqueira et al. 2008).

Schoeniger (1986) conducted a long series of experiments of interspecific crosses in the supersection Tacsonia, with the aim of introgressing genetic resistance to oidium and anthracnose from wild P. mixta and P. cumbalensis into P. tripartita var. mollissima. In this case, F1 hybrids also showed remarkable vigor, with larger leaves, stipules, bracts and flowers. According to Escobar (1981), pollen viability is equal to or higher in the hybrids, as compared to parental species. However, fertility strongly decreases in F2 and BC1 generations, following a lack of flowering and low fruit set, aggravated by poor seed germination and high mortality. Considerable variation appeared for all organs, including fruits (production, size, shape, succulence and flavor of arils), with cases of transgressions in the segregation and abnormal leaf and flower shapes. Selfing F2 and BC1 plants produced similar results, and some plants exhibited traits that were unknown in the parent species (Schoeniger 1986). Later, Coppens d’Eeckenbrugge crossed P. tripartita var. mollissima with P. tarminiana and P. mixta and characterized the resulting F1 hybrids; crosses were fully fertile. The hybrids were morphologically intermediate, with a tendency to be more similar to their maternal parent. They showed high vegetative vigor, with large leaves and larger, but somewhat rarer, fruits, of intermediate characteristics. Concerning resistance to anthracnose, which produces black dots depreciating the fruits, the hybrids between the susceptible P. tripartita var. mollissima and the resistant P. tarminiana expressed the symptoms at an intermediate level, suggesting a quantitative inheritance of the resistance and presaging difficulties for its effective introgression (Primot et al. 2005). These examples show that the introgression of resistance traits into cultivated passionfruits through interspecific crosses will be a long process. Sexual compatibility is not a synonym to genome compatibility and disease resistances are complex traits whose inheritance often seems to depend on a dosage effect. Lack of resistant germplasm in P. edulis f. flavicarpa leaves few other solutions than pursuing the effort. The situation may be different in species, such as P. tripartita, for which resistant conspecific materials may be explored, particularly in wild botanical varieties. P. mixta, a wild species, which shows spontaneous genetic exchanges with the cultigen, is also worth exploring.

The success of introgression at the interspecific level will depend not only on the number of crosses and generations obtained, but also probably more on the development of specific tools to understand and follow the transfer of targeted traits. A particular investment is necessary in cytogenetics, studying the chromosomal structure of the parents and recombination in the hybrids, as well as molecular tools, such as genetic maps and trait markers. Furthermore, cytogenetics could provide some keys for a better understanding of interspecific incompatibility in Passiflora.

7.6.5 Somatic Hybridization

Somatic hybrids of Passiflora species have been generated from protoplasts isolated from leaf tissues or callus derived from suspension cultures. These somatic hybrids are artificial polyploids and present, in general, chromosomal instability. For example, despite their high pollen viability, meiotic irregularities have been found in somatic allotetraploid hybrids (4n = 36) obtained from the cross between the diploid parents P. edulis f. flavicarpa (2n = 18) and P. amethystina (2n = 18) (Vieira and Carneiro 2004). In contrast, the somatic allotetraploid hybrids generated from protoplasts of P. edulis f. flavicarpa and P. cincinnata do not present aneuploidy. They have stable meiotic behavior and normal pollen viability suitable for the introgression of genes such as the resistance against Xanthomonas campestris pv. passiflorae present in P. cincinnata (Vieira and Carneiro 2004). Other desirable traits such as cold tolerance have been transmitted via somatic hybridization from P. incarnata to P. edulis f. flavicarpa (Otoni et al. 1995).

7.6.6 Genetic Transformation

To counteract the negative effects of pathogens, resistant transgenic plants of yellow passionfruit (P. edulis f. flavicarpa) have been generated. For instance, transgenic plants have become resistant to the virus passionfruit woodiness virus (PWV), introducing the gene codifying for the virus coat protein (CP) (Trevisan et al. 2007).

In another study, an antiapoptotic gene (p35) from a baculovirus, which is supposed to lead resistance to environmental stress and to a broad spectrum of diseases, was introduced to the genome of P. edulis f. flavicarpa by biobalistics (Freitas et al. 2007). Although it failed to confer resistance against the cowpea aphid-borne mosaic virus (CABMV) to transgenic passionfruit plants, it was able to increase their resistance against the bacterium Xanthomonas axonopodis pv. passiflorae, and their tolerance against the herbicide, glufosinate (Syngenta) (Freitas et al. 2007).

7.7 Genomics Resources Developed

7.7.1 Gene Sequences

In the NCBI genebank database (NCBI 2009), there are 1,648 Passiflora entries corresponding to 1,196 nucleotide sequences and 452 translated protein sequences. Approximately 95% of these entries were produced in phylogenetic studies and correspond to the sequences of plastid fragments, such as rcbL, the spacer trnH–psbA, the spacer trnL-F, and nuclear sequences, such as Internal Transcribed Spacer (ITS) and glutamine synthetase (GS2) (Muschner et al. 2003; Yockteng and Nadot 2004a, b; Krosnick and Freudenstein 2005; Hansen et al. 2006).

Population genetic studies have also generated molecular sequences. In a phylogeographical study, internal transcribed spacers (ITS) sequences of 32 individuals of P. actinia and 20 individuals of P. elegans from 41 different localities were obtained (Lorenz-Lemke et al. 2005). Besides, the study of genetic variation of populations of P. alata has provided new nuclear and chloroplast sequences (Koehler-Santos et al. 2006a). ITS of a fragment of glyceraldehyde 3-phosphate dehydrogenase (G3pdh) gene, the chloroplast intergenic spacers trnL–trnF and psbA–trnL and the intron trnL were sequenced to study 88 individuals of P. alata from 52 different localities of Brazil. For the first time, the second intron of gene LEAFY, which regulates the establishment of the floral meristem and flowering time in Arabidopsis, was sequenced in a Passiflora species. One novelty of this study was also the use of the first set of 12 microsatellite markers in a Passiflora species (Padua et al. 2005; Koehler-Santos et al. 2006a). A set of ten microsatellites specific to P. edulis f. flavicarpa, the yellow passionfruit, is also available (Oliveira et al. 2005) and was already used to construct a genetic linkage map (Oliveira et al. 2008) (see details in Sect. 7.5). Twenty-eight unpublished microsatellites specific to P. alata are also accessible in the NCBI database.

Several chloroplast and mitochondrial sequences of hybrids, product of interspecific crosses of Passiflora species, have been deposited in the database and used to determine the process of organellar inheritance in Passiflora (Muschner et al. 2006).

Besides, eight sequences of homolog isoforms of the allergenic compound of the pollen of Betulaceae (Bet v1) from Passiflora morifolia Mast., Passiflora sp., P. suberosa, P. misera and P. organensis Gardner are accessible in NCBI (Finkler et al. 2005). The aim of the study of Finkler and co-workers was to make a first characterization of Bet v1 homologs in Passiflora species and to identify the utility of this gene for future evolutionary and applied researches.

7.7.2 Expressed Sequence Tags

A few expressed sequence tag (EST) sequences of Passiflora are available in the NCBI database and all are obtained from the cultivated species P. edulis. Mita et al. (1998) were first to deposit five cDNA sequences of molecules playing an important role in the ethylene pathway. Their study generated two cDNA sequences of the ethylene receptor of P. edulis, called PE-ERS1 and PE-ETR1, which have a length of 2,297 bp with an open reading frame (ORF) encoding 637 residues of amino acids and 2,715 bp with an ORF encoding 738 amino acids. They also obtained cDNA sequences from two copies of the ACC synthase (PE-ACS1 and PE-ACS2) and one copy of the ACC oxidase (PE-ACO1), each having a sequence length of 1,054 bp with an ORF encoding for 351 amino acids, 1,063 bp with an ORF encoding for 354 amino acids, and 785 bp with an ORF of 261 amino acids. A subsequent study of ethylene pathway identified a third cDNA sequence of an independent copy of an ethylene receptor (PE-ERS2) measuring 2,357 bp that include an ORF codifying for 634 amino acids (Mita et al. 2002). The cDNA sequence of an antifungal protein (PE-AFP1) with a length of 25 amino acids residues was purified from seeds of P. edulis and constitutes the first defense peptide with antifungal properties identified in Passiflora (Pelegrini et al. 2006). Besides, the Myo-inositol-1l-phosphate synthase (MIPS) cDNA, component of Inositol phosphate biosynthesis pathway, as well as two cDNAs of translation elongation factor 1a-1 of P. edulis was sequenced (Abreu and Aragao 2007). MIPS cDNA isolated from developing passionfruit seeds measures 1,951 bp in length contains one open reading frame of 1,533 bp encoding 510 amino acids. Recently, the cDNA of the chloroplast-targeted allene oxide synthase from P. edulis (Pf-AOS), which is one of the components of the jasmonate pathway, was isolated. The coding region of Pf-AOS contains 1,512 nucleotides and encodes a protein of 504 amino-acids flanked by a 50-UTR of 155 nucleotides and a 30-UTR of 116 nucleotides (Siqueira et al. 2008). This study indicates that PfAOS may play an important role in systemic wound response against chewing insect attack (Siqueira et al. 2008).

7.7.3 Complete Genome Sequences

So far, no genome sequence is available from any Passiflora species. Instead, the genome of three pathogenic virus affecting Passiflora species has been totally sequenced (Iwai et al. 2006b; Song et al. 2006; Spiegel et al. 2007). The full maracuja mosaic tobamovirus (MarMV) genome consists of 6,794 nucleotides and contains four open reading frames (ORFs) coding for proteins (Song et al. 2006). The genome of carlavirus Passiflora latent virus (PLV) isolated from P. edulis plants measures 8,386 nucleotides and contains six ORFs (Spiegel et al. 2007). The genomic RNA of East Asian Passiflora virus (EAPV) isolated from P. edulis is composed of 10,046 nucleotides and contains a single ORF encoding a polyprotein of 3,220 amino acids (Iwai et al. 2006a). The availability of complete genome sequences of Passiflora virus will permit to understand better the mechanisms used by the pathogen to generate the disease in the host-plant. The continuous development of the molecular techniques will permit to obtain hopefully the first complete genome sequence of a Passiflora species.

7.8 Scope for Domestication and Commercialization

7.8.1 Wild Species as Resistant Rootstock

In passionflower, grafting appears to be a good and more immediate solution than breeding to reduce the heavy impact of soil pathogens. In French Guiana, Delanoë (1992) tested nine species as potential rootstocks for P. edulis f. flavicarpa. None showed incompatibility, but grafting success varied from 40 to 60% for P. coccinea, P. glandulosa and P. garckei, to 60–80% for P. nitida, P. laurifolia and P. cirrhiflora, or more than 80% for P. serrato-digitata, P. fuchsiiflora and P. candida. However, the plants grafted on P. fuchsiiflora and P. candida died after 17 months. In terms of quality and development of the rootstock and the scion, the best results were obtained with P. coccinea, P. glandulosa and P. laurifolia, followed by P. nitida and P. garckei. Junqueira et al. (2005) reported similar grafting experiments using P. nitida or a P. edulis f. flavicarpa × P. setacea hybrid as rootstock. The grafted materials yielded as much as seedlings of yellow maracuja, but only half of its production in cuttings. Grafted plants were effectively protected from soil pathogen attacks. However, the economic application of grafting is an issue because of the low availability of wild species germplasm, their erratic seed germination, and the elevated maintenance cost of a crop with short cycle (less than 3 years) (Meletti et al. 2005). In Colombia, Campos (1992) successfully used P. manicata as a rootstock for adapting P. tripartita var. mollissima to lower altitudes and dry conditions. However, the impact on economic parameters has not yet been tested.

7.8.2 New Commercial Fruit Crops

Among the very high diversity of Passiflora species producing an edible fruit, there are many potential candidates for developing new fruit crops. However, a more realistic approach would be giving priority to those species that have been partially domesticated, and thus rescuing their cultivated germplasm. These crops include P. incarnata, for temperate climates, species of series Laurifoliae such as P. maliformis and P. nitida, species of series Tiliifoliae such as P. platyloba and species of supersection Tacsonia such as P. cumbalensis, P. pinnatistipula and P. antioquiensis for tropical highlands. The gene pools for these species have been briefly presented in Sect. 7.4.

McGuire (1999) proposed to develop the maypop, P. incarnata into a fruit crop. This fruit was “abundant in Indian gardens” since its archeological remains are common in the southeast of the United States. However, the level of domestication reached in that time is uncertain (Gremillion 1989). The fruits of P. incarnata are good sources of vitamin A and niacin (Martin and Nakasone 1970).

The reputation of Passiflora maliformis suffers from its vernacular name of “stone granadilla.” However, the epicarp solidity is present only in part of its germplasm and the fruits are sold well on regional markets in Colombia. Some germplasm of this species shares interesting traits with the purple form of P. edulis. Indeed, the resemblance is such that they have received the same vernacular names of gulupa or chulupa. They share a yellow and highly aromatic pulp, similar fruit size and a deep red shell. But P. maliformis is superior to P. edulis in two aspects: the fruit shell does not wrinkle at maturity keeping its smooth appearance, and it is considered generally resistant or tolerant to most passionfruit pathogens. These interesting traits, found also in the closely related P. serrulata and P. multiformis, could make P. maliformis a strong competitor in the passionfruit market. Lowland species of series Tiliifoliae are already cultivated on a very small scale in forest regions of western Colombia. Their fruits could benefit from their resemblance with those of P. ligularis, which are appreciated in the national market. But their potential is wider since their fruits could be produced in many other rain forests. As for the highland sweet granadilla, their smooth and brittle epicarp makes them both attractive and easy to consume. The cultivated populations could currently provide an excellent basis for the commercial development of these species.

Species of the series Laurifoliae offer similar or even better perspectives. The genetic improvement should aim at equilibrating their resource allocation between generative and vegetative parts in order to reduce the overall development and promote a longer harvesting season. Breeding should easily maintain the general quality of the sweet and aromatic pulp, aril size, and fruit coloration, while the reduction of the fruit mesocarp would enhance pulp proportion.

The rosy passionfruit, P. cumbalensis, is cultivated only around Bogota, and this crop will certainly get lost in very near future if no particular attention is paid to it; this species is simply being forgotten. Indeed, it is very similar to its cousin P. tripartita var. mollissima, with which it shares the large and abundant orange arils of excellent flavor and aroma. The fruit is generally bright red, although some types give yellow fruits, and the plant seems less affected by anthracnose than P. tripartita var. mollissima. An effort should be done to promote the two fruits on the same market. The botanical varieties of supersection Tacsonia found from Colombia to Peru constitute a wide gene pool for the sustainable future of this potential crop (Holm-Nielsen et al. 1988). A similar case of a species that is becoming rare in home gardens is P. antioquiensis. Its germplasm is in high danger of extinction while it is still present in the wild. In cultivated materials, the flavor of the cream orange fruit pulp is a special and highly aromatic blend, matching the best fruits of supersections Passiflora and Tacsonia. But its fruiting may not be abundant enough for commercial production. The species should be restored not only as fruit but also as garden ornamental because of the exceptional beauty of its carmine and fusiform flowers hanging from 30 to 60 cm peduncles. Passiflora pinnatistipula is a very particular species of supersection Tacsonia because of its round fruits and its flowers, which have a relatively short hypanthium and a filamentose corona. The whitish pulp is extremely sweet; however, the seeds are relatively large, round and hard, so breeding could help to solve this problem. Another potential problem is that the species hybridizes very easily with P. tripartita var. mollissima, giving sterile hybrids (P. × rosea) that are difficult to distinguish from P. pinnatistipula until they flower. Therefore, these two species cannot be closely cultivated. Passiflora pinnatistipula may be grown at higher latitudes (up to 4,000 m), allowing fruit production in cold mountain climates.

7.8.3 Source of Therapeutic Compounds

Since prehistoric periods, Passiflora species have been used for therapeutic purposes (Dhawan et al. 2004). Although one species, P. incarnata, is the mostly recognized in pharmaceutics, many species have been used in traditional folk medicine in several countries (Table 7.3). After an exhaustive review, Dhawan et al. (2004) concluded: “This genus is a boon and blessing and a panacea for the ailing masses”. The medicinal properties of passionflowers are consequence of their diversity in phyto-constituents among which are flavonoids, glycosides, alkaloids and phenol compounds.

Several experimental studies performed in mice or rats give evidence of the anxyolitic, sedative, antioxidant and anti-inflammatory properties of Passiflora extracts such as P. incarnata, P. edulis, P. alata, P. quadrangularis extracts (Wolfman et al. 1993, 1994; Soulimani et al. 1997; Zanoli et al. 2000; Petry et al. 2001; Dhawan et al. 2003; Frode et al. 2004, 2005; Rudnicki et al. 2004; Santos et al. 2005; Bezerra et al. 2006; Garros et al. 2006; Gomes et al. 2006; Oliveira et al. 2006; Reginatto et al. 2006; Silva et al. 2006; Beninca et al. 2007; Castro et al. 2007; Montanher et al. 2007; Rudnicki et al. 2007; Vargas et al. 2007; Barbosa et al. 2008; Lorencini et al. 2008). Most of the studies have also shown an absence of secondary adverse effects of Passiflora extracts. No adverse effect of P. incarnata doses was found in female pregnant mice or their litter during gestational, fetal and post-natal periods (Mello et al. 2007). One experimental research in humans compared the effect of an oral dose of 500 mg of P. incarnata (Passipy™ IranDarouk) in a group of 30 patients with pre-operatory anxiety versus the effect of an oral dose of placebo in other 30 patients (Movafegh et al. 2008). The results show that anxiety was significantly lower in patients that took the Passiflora dose and the recovery of psychomotor function was comparable in both the groups. This study would support the efficacy of P. incarnata extracts as a remedy against anxiety. Other Passiflora medicinal uses are as antidote to counteract the venom of the scorpion Heterometrus laoticus (Uawonggul et al. 2006), and antimicrobial activity (Bendini et al. 2006).

The principal bioactive compounds described in this genus are the C-glycosil flavonoids (vitexin, isovitexin, orientin, isoorientin and apigenin) and β-carbolinic alkaloids (harman, harmin, harmalin, harmol and harmalol). The high concentration of isovitexin in P. actinia extracts would produce sedative and anxyolitic effects in animals (Santos et al. 2006). In addition, vitexin from various Passiflora species has an antithyroid effect in mice (Dhawan et al. 2004). Another active compound, the monoflavonoid chrysin purified from P. caerulea and P. incarnata, is possibly responsible for the anxyolitic and anticonvulsive effects on mice (Dhawan et al. 2004; Brown et al. 2007). This flavonoid is a ligand of benzodiazepine receptors and GABA receptors that mediates biochemical processes in the body. Passiflora incarnatas chrysin has also been reported to attenuate the suppression of activity of natural killer (NK) cells after a surgery (Beaumont et al. 2008). The NK cells are lymphocytes responsible for killing virus-infected cells and tumor cells during metastasis. Surgical procedures, anesthetics and pre- and post-surgery stress would affect the activity of NK cells. Therefore, Passiflora’s chrysin could have an indirect effect against cancer metastasis. Other anticancer activity was reported in extracts from P. edulis f. flavicarpa that influence the death of transformed foci, potential tumor cells (Rowe et al. 2004). In addition, extracts from fruit of P. edulis and P. foetida L. var. albiflora inhibit the activity of two enzymes, the gelatinase MMP-2 and MMP-9, two metallo-proteases involved in the tumor invasion, metastasis and angiogenesis (Puricelli et al. 2003). Chrysin has also reported to have some negative effects such as hyperalgesia, which is increased sensitivity to pain stimulation (Zhai et al. 2008). Another compound isolated from P. edulis is a defense peptide (Pe-AFP1) that inhibits the development of filamentous fungi Trichoderma harzianum, Fusarium oxysporum and Aspergillus fungigatus. The discovery of Pe-AFP1 could contribute, in a near future, to the development of biotechnological products as antifungal drugs against pathogenic fungi (Pelegrini et al. 2006). Besides, a hemolysin, probably a saponin, was isolated from the leaves of P. quadrangularis. Hemolysins are transmembrane receptors and are potential bactericidal and anticancer drugs (Yuldasheva et al. 2005).

The concentration of phytoconstituents in plants apparently depends on their mineral nutrition. The absence of minerals such as potassium, phosphor and nitrogen, in the nutrition of Passiflora plants increments their concentration in vitexin (Freitas et al. 2008). Populations of the same species located in different geographical areas would have different profiles in their phytoconstituents. For instance, the concentration of vitexin in P. foetida varies among samples collected in different geographical regions (Pongpan et al. 2007).

Despite all the beneficial pharmaceutical effects of passionflowers, precaution is needed about the presence of toxic compounds such as cyanogenic constituents. The plants of Passiflora produce secondary metabolites, especially cyanogenic glycosides as protection against herbivores. The unripe fruits of P. adenopoda DC. have already caused poisoning due to the presence of HCN (Dhawan et al. 2004). Passiflora manicata is called diablito (little devil) in Ecuador, for similar reasons.

Although a few studies are based on non-cultivated species, wild Passiflora species have been reported to be diverse in their phytoconstituents (Abourashed et al. 2002, 2003). Wild species also produced flavonoids and alkaloids. Sometimes, the concentration of these phytoconstituents is larger in wild species than in pharmaceutical ones. For example, the concentration of vitexin is seven times higher in the wild species P. holosericea L. than in P. incarnata (Abourashed et al. 2002). Wild species constitute, therefore, an important source of potential therapeutic compounds. Hence, it is necessary to conduct studies in order to detect their pharmaceutical properties and their active compounds.

7.8.4 Psychoactive Drugs

In certain tropical regions of South America, the dried leaves of P. edulis f. flavicarpa are sometimes smoked in the same way as those of Cannabis sativa L. Specialized websites indicate similar uses for P. incarnata, although the effect is said to be weaker than that of marijuana. Tea made with dried flowers also has a slight psychotropic effect. Indeed, P. incarnata may contain small amounts of harmala alkaloids, which are used in the preparation of a highly psychotropic drink, the ayahuasca (Recreational Drug Information Website 1999). Besides, harmala alkaloids reduce the breakdown of the psychoactive drug DMT (N,N-dimethyltryptamine) in the digestive system. In traditional medicine, P. incarnata has been used for drug deaddiction of morphine, cannabinoids, nicotine and alcohol, which was successfully tested on mice (Dhawan et al. 2004).

7.8.5 Dietary Supplements

Once again, dietary supplements are mainly reported in cultivated species. However, wild passionflowers constitute a latent source of dietary supplements waiting to be explored. The yellow passionfruit rinds, usually discarded after juice production, have been reported to be a source of variable dietary supplements. Firstly, their cell wall material rich in non-starchy polysaccharides constitute a possible source of dietary fiber, which could protect against cardiovascular diseases, diabetes, obesity, colon cancer and diverticular diseases (Yapo and Koffi 2008). Furthermore, yellow passionfruit rinds are also a potential source of pectin, a compound widely used in many foodstuffs as gelling agents (Yapo and Koffi 2006). The pectin isolated from P. edulis f. flavicarpa is comparable to the commercial citrus low-methoxyl pectin. Finally, the yellow passionfruit rinds have been also reported as a rich source of lycopene that is an antioxidant carotenoid which probably reduces risk of chronic diseases (Mourvaki et al. 2005). Few sources of lycopene are available in diet such as tomatoes.

7.8.6 Other Uses

Passiflora species could be a source of ecofriendly materials. For instance, a new polysaccharide, xiloglucan biopolymer, is extracted from peels of P. ligularis. This biopolymer could be used to produce a biodegradable film using the discards of passionfruit industry (Tommonaro et al. 2007). Seeds of P. edulis are also a source of vegetable oils useful to produce ecofriendly polymeric material (Lopes et al. 2008). Moreover, P. incarnata extracts have been described as useful to develop organic sunscreens with a protective defense against UV radiation. The use of plant compounds would reduce the concentration of chemical UV filters in sunscreens (Velasco et al. 2008). In addition, the enzyme system of P. edulis fruit barks has permitted the biotransformation of carbonyl compounds such as ketones and aldehydes (Machado et al. 2008). These compounds used in perfumery and in paints do not degrade quickly over time, hence the utility of passionfruit preparations to reduce compounds produced in the industry. In addition, peels of P. edulis have also permitted the reduction of synthetic dyes such as methylene-blue (Pavan et al. 2008).

Among the numerous alkaloids, phenolics, flavonoids and volatiles substances found in passionflowers, kavain, yagonin, dihydromethysticin and coumarin have properties that completely suppress the barnyardgrass growth. Hence, the use of the allelopathic plant P. edulis as a natural herbicide would reduce the dependency on synthetic herbicides and other agrochemicals (Khanh et al. 2006, 2008).

7.9 Invasive Species and Weeds

In a general sense, a weed is a native or exotic plant that grows and reproduces aggressively causing nearly always economic, environmental or ecological loss (Randall 2002). Instead, invasive species is an exotic species (e.g., plant or animal) that adversely affects economically, environmentally or ecologically the habitats that it invades. Human activities provoke the transportation of species beyond their native range. The introduced species are sometimes able to establish and spread in a foreign habitat. Among Passiflora species, 19 species have been reported as invasive in 37 countries (Table 7.5).

Table 7.5 Passiflora invasive species and native weeda

Islands, which usually present a great number of endemic species, are particularly susceptible to invasions because the introduced species lack natural competitors and predators that control populations in their native ecosystems. The probability of successful invasions is also increased in islands because of the availability of unfilled ecological niches. For instance, the entire island chain of Hawaii has been devastated by foreign insects, plants and animals among which 12 Passiflora species with two cause a documented impact. The introduced Andean species P. tarminiana (banana poka in Polynesian local language; often confused with P. tripartita var. mollissima) invaded at least 50,000 ha of upper-elevation rainforests of Hawaii (La Rosa 1992; Trujillo et al. 2001; Trujillo 2005), threatening the survival of the native Hawaiian koa (Acacia koa A. Gray) and other 20 native endangered Hawaiian plant species (Trujillo et al. 2001). To control this weed, the US government spent $90,000 per year for uprooting plants and $500,000 to find adequate biological control agent against it (Waage et al. 1981; Causton et al. 2000; Causton and Ranget 2002). A fungal pathogen Septoria passiflorae (Coelomycetes: Sphaeropsidales), noxious against P. tarminiana, was found in the plant native range (Trujillo et al. 1994). Inoculations of this biological agent permitted to eliminate 95% of the biomass of banana poka in 4 years and allowed the revitalization of the high-elevation forests of Kauai, Maui, and Hawaii (Trujillo et al. 2001; Trujillo 2005).

A native species can become a weed by the effects of natural events or human modifications to the habitat. For example, the Brazilian species, P. alata also known as maracuja doce, is colonizing unoccupied regions of southern Brazil (Koehler-Santos et al. 2006a, b). The colonization process of P. alata may have been facilitated by the presence of dispersion and pollination agents of other Passiflora species already present in the colonized area. In addition, high genetic diversity among and within the populations of P. alata probably promotes the success of the colonization process. Therefore, P. alata can, in these circumstances, be considered a native weed. In addition to P. alata, two other passionflowers are reported to be native weed, P. cinnabarina Lindl. in Australia and P. cincinnata also in Brazil (Randall 2002).

Weed and invasive species can be a threat for human health if the species are toxic. For instance, the weed P. alata induces occupational respiratory allergic disease (Giavina-Bianchi et al. 1997), and should be considered a risk because it could induce allergic problems in new colonized areas. Ten weedy Passiflora are also accounted as toxic (Table 7.5). Programs for the identification of Passiflora weeds and prevention of dispersal of these plants are essential to maintain the equilibrium of ecosystems.

7.10 Conclusion

Wild Passiflora species constitute a large source of genetic resource for the improvement of cultivated species and the introduction of new fruit crops and tropical garden ornamentals. Moreover, new pharmaceutical, organic and ecofriendly products can be developed from these wild passionflowers. Although the majority of studies have been conducted on cultivated species, in particular in P. edulis, they constitute a useful reference to conduct similar studies in wild species

The fragmentation and loss of habitat caused by human activities endanger many tropical plants, such as wild Passiflora species. Many passionflowers are endemic and certainly threatened; thus it is indispensable to ensure their conservation. The determination of their geographical distribution is required to further develop in situ conservation programs. In addition, ex situ Passiflora germplasm collections should be increased in number of species and individuals to ensure the conservation of a great range of the genetic variability of wild species and its economic valorization.