| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Systematics |
2Department of Molecular Biotechnology, Faculty of Bioscience Engineering, Ghent University (UGent), Coupure links 653, B-9000 Ghent, Belgium; 3Department of Biology, Faculty of Sciences, UGent, K.L. Ledeganckstraat 35, B-9000 Ghent, Belgium; 4Institute for Plant Biotechnology for Developing Countries (IPBO), UGent, K.L. Ledeganckstraat 35, B-9000 Ghent, Belgium; 5Naturaleza & Cultura Internacional, Mercadillo 18-10 y José Maria Peña, Loja, Ecuador
Received for publication April 29, 2004. Accepted for publication March 3, 2005.
ABSTRACT
Validity of the taxa currently recognized in the genus Vasconcellea was analyzed by investigating morphological and molecular data from 105 specimens of this genus and six specimens of the related genus Carica. Taxon identification of these specimens was compared with clustering in two phenetic dendrograms generated with 36 morphological characters and 254 amplified fragment length polymorphic (AFLP) markers. Moreover, cytoplasmic haplotypes were assessed using polymerase chain reaction restriction fragment length polymorphism (PCR-RFLP) of one mitochondrial and two chloroplast DNA regions. Results show that the morphological data set, containing mainly vegetative characteristics, merely reveals external resemblance between specimens, which is not directly associated with genetic relationships and taxon validity. Phenotypic plasticity and intercompatibility between several species are likely to confuse morphological delimitation of the taxa. Based on the results of our study, several specimens that could not be identified with the currently used identification key (1) could be attributed to a known taxon, which should be extended to include a higher range of morphological variability or (2) could be hypothesized to be of hybrid origin. Because of the high intraspecific variation within V. microcarpa and V. x heilbornii, revision of these taxa is recommended.
Key Words: AFLP • Carica • hybridization • morphological variability • PCR-RFLP • phenetic relationships • species separation • Vasconcellea
Vasconcellea Saint-Hilaire is by far the largest genus of the Caricaceae Dumortier, uniting 21 of the 35 taxa described for this dicotyledonous plant family (Badillo, 1971
, 1993
, 2001
). In the current classification (Badillo, 2001
) Vasconcellea comprises 20 species and 1 hybrid, Vasconcellea x heilbornii. In spite of the frequent absence of sexual reproduction, this hybrid is usually considered as a species.
Species of Vasconcellea are commonly referred to as highland papayas or mountain papayas (National Research Council, 1989
) because of their resemblance with papaya (Carica papaya) and their typical ecological preference for higher altitudes. Until recently (Badillo, 1971
, 1993
), Vasconcellea (also spelled as Vasconcella) was considered a section, sister to the section Carica, within the genus Carica L. Badillo (2000)
separated the monospecific section Carica (containing only Carica papaya) from section Vasconcellea, based on morphological and genetic (Aradhya et al., 1999
) evidence, by rehabilitating the section on generic level. Vasconcellea species are distributed throughout South America, with a concentration of diversity in the Andean valleys of Ecuador, where 16 of the 21 described species appear up to 3500 m a.s.l. (Badillo, 1993
, 1997
, 1999
; Romeijn-Peeters, 2004
). Five species of this genus have been placed on the International Union for the Conservation of Nature and Natural Resources (IUCN) Red List of Threatened Species: V. horovitziana, V. omnilingua, V. palandensis, V. pulchra, and V. sprucei (IUCN, 2003
). Personal observations in Ecuador suggest that even more species of Vasconcellea are endangered. Most significant threats are habitat destruction resulting from deforestation and conversion of forests into croplands or grasslands (IUCN, 2003
).
The most comprehensive review of the genus Vasconcellea, as section Vasconcella, is the monograph of Caricaceae by Badillo (1971
, 1993)
. This monograph contains an identification key, mainly based on characters of the staminate flowers, followed by comprehensive morphological circumscriptions of the different taxa. Species of Vasconcellea are wild, semi-domesticated, or domesticated plants, with a shrub- or treelike and pachycaulous habit. Plants are usually dioecious but sometimes monoecious or polygamous. The medullar stem is mostly simple or scarcely branched, in some species it is covered with spiny stipules, whereas the leaves are concentrated in a terminal crown. Leaves are large to very large and vary extensively in shape, from entire to compound. All organs produce a white latex, containing cysteine endopeptidases. Flowers are pentamerous with white, green, yellow, orange, or pink petals. The fruit is a berry with varying shape, dimension, and color (Badillo, 1993
).
In the course of previous ethnobotanical inventories of wild and semi-domesticated edible plants in southern Ecuador, an unrecognized variability among and within some species of the genus Vasconcellea was observed (Jiménez et al., 1998
; Scheldeman, 2002
). This high diversity is probably partly caused by the intercompatibility between several species (Jiménez and Horovitz, 1957
; Horovitz and Jiménez, 1967
; Mekako and Nakasone, 1975
) leading to the production of hybrids with varying degrees of fertility, which have been shown to occur spontaneously in areas where species distributions overlap (Badillo, 1971
). Interspecific hybridization can lead to fertile hybrids, which may cross with parental or nonparental species (Badillo, 1971
). Such complex hybrid populations in the so-called hybrid zone are a cline of morphological and genetic variability (Barton and Hewitt, 1985
).
Two naturally occurring hybrids of Vasconcellea with high introgressive potential have already been described by Horovitz and Jiménez (1967)
and Badillo (1971
, 1993)
: (1) V. x heilbornii, a taxon abundantly present in southern Ecuador and (2) an occasionally occurring hybrid between V. monoica and V. cundinamarcensis. Within V. x heilbornii, Badillo (1993)
recognizes the cultivar Babaco and the varieties chrysopetala and fructifragrans.
During our expeditions in Ecuador, we realized that many Vasconcellea specimens could not be identified at the specific level with the dichotomous key of Badillo (1993)
and that the high morphological variability within the genus is insufficiently understood (J. P. Romero-Motochi, E. Romeijn-Peeters, B. Van Droogenbroeck, and T. Kyndt, personal observation). Taxon identification is hard or even impossible when only vegetative plant parts are present, which is often the case during collection. Additional information about the genotype of the plants is very much needed to resolve taxonomical problems in this genus. Because the genotype is not influenced by environmental factors, evolution of closely related taxa can be investigated from an objective point of view with molecular techniques (Hillis, 1987
). In addition, some molecular marker assays, e.g., amplified fragment length polymorphism (AFLP) and random amplified polymorphic DNA (RAPD), reveal a large amount of characters per reaction, i.e., a high multiplex ratio, making them very useful in the assessment of botanical relationships and diversity (Karp et al., 1996
; McLenachan et al., 2000
). Recently, some molecular analyses have been performed in the Caricaceae family to clarify interspecific and intergeneric phenetic and phylogenetic relationships with the aid of fingerprinting techniques, such as RAPD (Jobin-Decor et al., 1997
), polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) (Aradhya et al., 1999
; Van Droogenbroeck et al., 2004
), and AFLP (Kim et al., 2002
; Van Droogenbroeck et al., 2002
). Their results verify the large genetic distance between the genera Carica and Vasconcellea, thereby validating their recent rehabilitation. Moreover, these studies revealed relationships between some Vasconcellea species on molecular level. The most extensive analysis, involving eight species of Vasconcellea, two species of Jacaratia A. DC. (Caricaceae) together with Carica papaya (Van Droogenbroeck et al., 2002
), suggested a close genetic relationship between the following species pairs: (1) V. stipulata and its putative hybrid V. x heilbornii, (2) V. weberbaueri and V. parviflora, and (3) V. palandensis and V. goudotiana. Aradhya et al. (1999)
described intraspecific variation in chloroplast PCR-RFLP haplotypes of V. microcarpa, V. quercifolia, and V. x heilbornii. Vasconcellea x heilbornii surprisingly did not share its haplotype with either one of its putative parent species, V. stipulata and V. cundinamarcensis. These results were confirmed and extended by the PCR-RFLP analysis of Van Droogenbroeck et al. (2004)
. Vasconcellea microcarpa and the hybrid V. x heilbornii again showed intraspecific variability, with specimens of V. x heilbornii having either the haplotype of their putative mother V. stipulata or, surprisingly, V. weberbaueri.
In general, the molecular studies mentioned (Aradhya et al., 1999
; Kim et al., 2002
; Van Droogenbroeck et al., 2002
, 2004
) demonstrate that AFLP fingerprinting (Vos et al., 1995
) and PCR-RFLP analysis of cpDNA and mtDNA are time- and cost-efficient methods to analyze inter- and intraspecific relationships and to investigate hybridization among Vasconcellea species.
The objectives of this study were (1) to verify the validity of the identification key of Badillo (1993)
by comparing the taxonomical grouping of the Vasconcellea specimens with detailed molecular and vegetative morphological data, (2) to evaluate the possible development of a vegetative identification key, (3) to investigate gaps in the current identification key by analyzing specimens that could not be identified unambiguously with this key, and (4) to evaluate possible hybridization events by comparing nuclear (AFLP) and cytoplasmic (PCR-RFLP) marker data.
MATERIALS AND METHODS
Plant material
A total of 105 Vasconcellea individuals and six individuals of the outgroup species Carica papaya were sampled (Table 1). Most specimens were collected in Ecuador by the authors. Leaf material or seeds from some individuals were kindly provided by other researchers (see Table 1 for details). For V. sphaerocarpa, V. sprucei, and V. glandulosa only herbarium specimens were available. Permission was given by the directors of these herbaria (HUA, BM, U) to use those specimens as sources of DNA.
|
) and named after Badillo (2001)
. After identification, a taxon-specific code was given to each specimen to increase the readability of this paper (Table 1). In total, the collected individuals represent 20 described taxa of the genus Vasconcellea. Sixteen specimens could not be readily identified due to the absence of staminate flowers or due to lack of resemblance with any of the described species. These unidentified specimens were given a taxon code starting with ‘sp.’ Specimens of V. x heilbornii that could not be identified either as one of the varieties chrysopetala or fructifragrans or as the cultivar Babaco were given a taxon code starting with "heil." Voucher specimens were deposited at GENT. Samples with no mature leaf material were omitted from the morphological analysis (see Table 1).
Morphological analysis
The identification key of Badillo (1993)
is mainly based on staminate flowers. As most species of Vasconcellea have a long vegetative state, this feature is often useless. Therefore this study focused on vegetative features, although also one generative characteristic, i.e., color of petals, was studied. Because leaf dimensions within Vasconcellea are very variable, quantitative data were considered rather undiagnostic. Consequently, mainly qualitative features were studied. All characteristics studied are listed in Table 2.
|
). Morphological data were converted into a similarity matrix, using the simple matching coefficient (Sneath and Sokal, 1973
), with the SIMQUAL function. A dendrogram was generated from the similarity matrix by the unweighted pair-group method using arithmetic averages (UPGMA) (Sokal and Michener, 1958
) with the SAHN function. The cophenetic correlation coefficient was calculated with a Mantel test (Mantel, 1967
) by comparing the matrix of cophenetic values with the similarity matrix, in order to estimate how well the dendrogram represents its corresponding pairwise distance matrix. This was done with the COPH and MXCOMP modules of NTSYS-pc (Rohlf, 2000
). Further, principal co-ordinate analysis (PCoA) (Gower, 1966
) was performed by extracting eigenvectors from the similarity matrix, using the EIGEN function. The data were projected onto the resulting eigenvectors using the PROJ function and two-dimensional plots of the data were achieved using MXPLOT.
Molecular analysis
DNA extraction
Leaf tissue was dried with silica gel and ground in liquid nitrogen. Total genomic DNA was extracted using the Qiagen Dneasy Plant Mini kit (Qiagen, Hilden, Germany). For herbarium material, an incubation in buffer AP1 (Qiagen) for 1 h at 65°C was necessary prior to extraction, and only small amounts of partially degraded DNA were obtained.
AFLP analysis
The AFLP analysis was carried out as previously described by Van Droogenbroeck et al. (2002)
. The following five primer combinations were used: E + GA/M + ACAA, E + GT/M + ACAA, E + GA/M + GCGT, E + GA/M + CTGT, and E + CG/M + CTGG. For each individual, DNA fingerprints were scored by visual inspection for presence (1) or absence (0) of specific AFLP fragments. Only distinct, major bands were scored. Data matrices were analyzed using Treecon 1.3b (Van de Peer and De Wachter, 1994
) and NTSYS-pc version 2.10L (Rohlf, 2000
). Genetic similarities were calculated using Jaccard's coefficient (Jaccard, 1908
) with the SIMQUAL module of NTSYS-pc or the DISTANCE ESTIMATION option of Treecon. Similarity matrices were analyzed using the UPGMA (Sokal and Michener, 1958
) clustering method in NTSYS-pc (SAHN module). Calculation of the cophenetic correlation coefficient was done as described. Reliability of clusters in each dendrogram was tested by bootstrap analysis (Felsenstein, 1985
) with 1000 replications using Treecon. Additionally, a PCoA analysis was performed based on the genetic similarity matrix.
CpDNA and mtDNA haplotype determination using PCR-RFLP
The PCR-RFLP data of two cpDNA regions (trnK1-trnK2 [K1K2] and trnM-rbcL [ML]) and one mtDNA region (nad4/1-nad4/2 [nad4/1–2]) were already available for some specimens included in our sample set (Van Droogenbroeck et al., 2004
). For the other specimens, additional PCR-RFLP data were generated with the eight PCR-fragment/enzyme combinations selected by Van Droogenbroeck et al. (2004)
: K1K2/EcoRV, K1K2/ScaI, K1K2/AfaI; ML/PstI, ML/ MseI for the cpDNA regions and nad4/1–2/HinfI, nad4/1–2/BstOI, nad4/1–2/ DdeI for the mtDNA region. Haplotypes were defined as a set of specific combinations of the observed variants for all detected mutations (for details, see Van Droogenbroeck et al., 2004
).
RESULTS
Morphological analysis
The UPGMA dendrogram based on morphological similarity values (simple matching coefficient), with cophenetic value of 0.86, is presented in Fig. 1. Three main clusters can be distinguished at 50% similarity level: cluster 1 consists of all specimens with parted leaves and (palin-)actinodromous venation; cluster 2 contains the only studied species that has compound leaves, i.e., V. palandensis; and cluster 3 contains specimens with simple leaves and pinnate venation.
|
Subclusters in cluster 1 are less clearly defined. Specimens belonging to the same taxon are clustered separately and intermingled with other taxa. Only V. x heilbornii Babaco (with its typical small number of lobules, 1C), C. papaya (the only taxon with nine primary veins, 1N) and V. weberbaueri (with typical serrate leaves, 1O) are grouped in clearly separated taxon-specific clusters. Also V. cauliflora (bearing white flowers and showing emergences on petiole and lamina, 1B), V. parviflora (showing wide secondary vein spacing and lobe bases typically narrowed above attachment, 1E), and V. monoica (narrow lobes with a remarkably low number of lobules, 1H) are clustered in species-specific clusters, but they appear less separated from the neighboring taxa. V. sphaerocarpa, V. sprucei, and V. glandulosa are isolated from the other taxa (clusters 1G and 1A), but because only one specimen was included for each of them, it is difficult to define these as species-specific clusters. All specimens of V. cundinamarcensis, except for cund193, are clustered with sp203 (cluster 1K). Two heterogeneous and remote clusters (clusters 1F and 1L) include all specimens of V. x heilbornii var. chrysopetala, V. x heilbornii var. fructifragrans, and most of the unidentified V. x heilbornii together with their putative parent species V. stipulata. Both described varieties of V. x heilbornii (var. chrysopetala and var. fructifragrans) are present in cluster 1F. Furthermore, both specimens of V. omnilingua are grouped in cluster 1I, together with unidentified specimens, sp239, sp240, and sp241, and specimens of V. microcarpa subsp. heterophylla and subsp. microcarpa (mich177 and micm212). Finally, all specimens of V. goudotiana, some of the remaining specimens of V. microcarpa (subsp. baccata, subsp. heterophylla, and subsp. microcarpa), and unidentified specimens are interspersed between and within the afore-described (sub)clusters.
A PCoA-analysis based on the simple matching coefficient was performed (results not shown). The PcoA supports separation of the three main clusters obtained with the cluster analysis (Fig. 1). However, separation in subclusters is poor. The first three principal coordinates account for 71.1% of the variation. The second principal coordinate (10.2%) separates more or less the specimens with different leaf division, i.e., simple leaves with pinnate venation, simple leaves with (palin-)actinodromous venation, and compound leaves (clusters 1, 2, and 3 in Fig. 1). Taxa are slightly separated by principal coordinates 1 (56.9%) and 3 (4.0%), but no clear spreading is present.
AFLP analysis
Specimens (Table 1) were analyzed with AFLP using five primer combinations, selected by Van Droogenbroeck et al. (2002)
for their high number of bands and polymorphism. Of a total of 254 scorable fragments, only nine (3.5%) were monomorphic in both Vasconcellea and Carica. When only Vasconcellea was considered, 19 (7.5%) monomorphic markers were found.
Unfortunately, the quality of the DNA obtained from herbarium material was too poor to be used in AFLP analysis, as might have been expected based on the age of the material and on similar experiences from other researchers (McLenachan et al., 2000
).
The AFLP data were used to make pairwise comparisons of the genotypes on both shared and unique amplification products to generate a similarity matrix using Jaccard's coefficient. The Jaccard coefficient of band matching is recommended for the analysis of DNA fingerprint data because it only takes into account positive band matching (Weising et al., 1995
).
Figure 2 shows the dendrogram with a cophenetic value of 0.85, generated using the UPGMA clustering method. On 55% Jaccard's diversity level seven clusters can be distinguished. Cluster 7 contains only C. papaya genotypes, clearly separated from all specimens of the genus Vasconcellea. Cluster 1 contains V. parviflora (1f), V. weberbaueri (1e), the hybrid V. heilbornii (1b, 1c, and 1d), and one of its putative progenitor species, V. stipulata (1a). Cluster 2 is separated into three subclusters. Cluster 2a holds V. palandensis (2a(IV)), V. cundinamarcensis (2a(I)), V. goudotiana (2a(V)), sp200 en sp203 (2a(III)), and subcluster 2a(II) (bootstrap value = 69%) containing sp101, sp205, sp312(I) and sp312(II). Within cluster 2b, individuals of the described species V. pulchra (2b(II)) and V. longiflora (2b(III)) are grouped together with V. microcarpa subsp. heterophylla and subsp. baccata and five morphologically unidentified individuals of Vasconcellea. Cluster 2c consists of V. monoica (2c(II)), V. omnilingua (2c(VI)), and V. microcarpa subsp. microcarpa and one specimen (mich266) of the subsp. heterophylla, together with five unidentified specimens. Individuals of each described species within clusters 2b and 2c are grouped together with high bootstrap values (bootstrap value = 81–100%), except for specimens belonging to V. microcarpa and its subspecies, which are scattered throughout these two clusters without any apparent affinity. Cluster 5 combines V. quercifolia (5a) and V. candicans (5b) with a bootstrap value of 69%. Clusters 3, 4, and 6 are species-specific, containing respectively V. cauliflora, V. chilensis, and V. crassipetala. Clusters 1–6 are combined in a Vasconcellea cluster supported by a 94% bootstrap value.
|
|
PCR-RFLP analysis
Eight fragment/enzyme combinations, selected by Van Droogenbroeck et al. (2004)
, were applied to analyze the complete sample set given in Table 1. The haplotype of each individual was determined based on a set of specific combinations of the observed variants for all detected mutations, as specified in Van Droogenbroeck et al. (2004)
.
Twelve different cp haplotypes (chlorotypes A–L) and four different mt haplotypes (mitotypes A–D) were found in our sample set. Specimens sharing the same chlorotype always had the same mitotype. Intraspecific haplotype diversity was only observed for individuals belonging to taxa V. x heilbornii and V. microcarpa.
Figure 2 gives an overview of the genetic information obtained with AFLP and PCR-RFLP. All individuals from AFLP cluster 1 share the same mitotype (C) and have three different chlorotypes (F, G, and J). Four other species also hold mitotype C, V. cauliflora (cluster 3), V. crassipetala (cluster 6), V. candicans (cluster 5b), and C. papaya (cluster 7), but have a different chlorotype (E, G, H, and L, respectively). These species are each grouped in a distinct cluster in the AFLP analysis. Vasconcellea crassipetala and V. parviflora share the same chlorotype and mitotype (G and C), although they are quite diverse according to nuclear AFLP results (mean similarity = 0.30). Cluster 2 contains the individuals with mitotypes A and B and chlorotypes A, B, C, and D. Chlorotype B is only represented in cluster 2b, while cluster 2c only includes chlorotype A. Cluster 2a comprises specimens with chlorotypes A, C, and D.
DISCUSSION
Comparing taxonomical grouping with molecular and morphological data
To verify validity of the current classification of Vasconcellea (Badillo, 1993
), molecular and morphological UPGMA dendrograms and cytoplasmic haplotype data were compared with taxonomical grouping. If taxa are valid, they are supposed to be defined within the molecular dendrogram. In the morphological study we have focused on vegetative characters to test if it would be possible to create an identification key based on vegetative morphological data. If this is the case, molecularly validated taxa should be morphologically distinguishable based on the used data. This means that specimens belonging to the same taxon should be clustered together in the morphological dendrogram, as long as the discriminating features are included in the analysis. Taxon validity will not be evaluated for taxa represented by only one specimen.
Legitimacy of the recent generic rehabilitation of Vasconcellea (Badillo, 2000
) is confirmed by the high AFLP-based genetic diversity (74%) between Carica and Vasconcellea. Comparable values of genetic diversity between both genera were reported before by Jobin-Decor (1997)
and Van Droogenbroeck et al. (2002)
, who presented values of 73% and 77%, respectively. In comparison with the high genetic diversity that exists between these two genera, our geographically very diverse sample set of Carica papaya (collected in South America, Africa, and Asia) and the 17 studied species of Vasconcellea showed quite limited intrageneric genetic variation (14% and 53%, respectively). In combination with the fact that Vasconcellea is reported to be slightly more closely related to Jacaratia, another genus of the Caricaceae, than with Carica (Aradhya et al., 1999
; Van Droogenbroeck et al., 2002
), we conclude that taxonomic delimitation of the genus Vasconcellea is unquestionably supported by several molecular analyses, including our data. On the other hand, morphological data used in this study do not adequately support the genetic divergence between Carica and Vasconcellea. This is not surprising because only one of the discriminating features listed by Badillo (2000)
, particularly the number of primary veins, is a vegetative characteristic. Consequently only one of the morphologic features discriminating Carica and Vasconcellea is included in this study.
Within the genus Vasconcellea only two taxa exhibit intraspecific variability in the cytoplasmic fragments analyzed in this study: V. x heilbornii and V. microcarpa each reveal two different chlorotypes. These two taxa also show very high AFLP-based genetic diversity, 33% and 43%, respectively. The results for these taxa will be discussed further in the following section.
All other taxa for which multiple specimens were included in this study are molecularly supported by both PCR-RFLP and AFLP analysis showing no intraspecific variability with PCR-RFLP and relatively low genetic diversity values with AFLP. AFLP-based clustering shows clearly delineated clusters that are confirmed by bootstrap values between 55% (V. cundinamarcensis) and 100% (several taxa). The following 10 taxa are also well-defined by morphological data: C. papaya, V. x heilbornii ‘Babaco’, V. quercifolia, V. candicans, V. weberbaueri, V. palandensis, V. parviflora, V. monoica, V. pulchra, and V. cauliflora.
Although molecular relationships do corroborate the validity of V. longiflora, V. cundinamarcensis, V. stipulata, and V. goudotiana, this is not reflected in an exclusive morphological delineation of their specimens. This lack of morphological association is probably caused by the fact that none of the studied features is discriminating for these taxa or by the reported presence of high variability within taxa (Jiménez et al., 1998
), possibly affected by phenotypic plasticity, introgression, or recent diversification. Rapid diversification and speciation is a typical characteristic of plant evolution on the South American continent that has been noticed since it became isolated during the late Cretaceous and early Tertiary Period, about 70 to 60 million years ago (Burnham and Graham, 1999
). Genetic associations between certain groups of taxa are likewise not mirrored in the morphological clustering. For instance, the genetic relationship between V. stipulata, V. x heilbornii, V. weberbaueri, and V. parviflora, which was previously also reported in other AFLP and PCR-RFLP studies (Van Droogenbroeck et al., 2002
, 2004
), is not reflected in the morphological dendrogram. On the other hand, the clear morphological clustering of the specimens on the basis of similarity in leaf division is not confirmed by genetic associations.
In conclusion, these findings indicate that our morphological data set, containing mainly vegetative characteristics, only reveals external resemblance between the specimens that is not directly associated with genetic relationships and molecular taxon validity. Therefore we can conclude that an identification key solely based on the studied vegetative characters is problematic for this genus. Phenotypic plasticity, recent diversification, and intercompatibility between several species may confuse morphological clustering based on the vegetative data. Possible hybridization events will be discussed further in this text.
Intraspecific diversity in V. x heilbornii and V. microcarpa
Only two analyzed taxa reveal variation at the cytoplasmic level as well as high AFLP-based diversity and morphological separation: V. x heilbornii and V. microcarpa.
A 33% mean genetic diversity value and two different chlorotypes were found in V. heilbornii, the supposed hybrid between V. stipulata and V. cundinamarcensis (Horovitz and Jiménez, 1967
; Badillo, 1971
). Of the two described varieties and the cultivar within this taxon, AFLP results only support the delimitation of V. x heilbornii Babaco with a bootstrap value of 100%. On the other hand, the varieties fructifragrans and chrysopetala are intermingled in both the molecular and morphological dendrograms. According to Badillo (1993)
, the subdivision between both varieties is only based on the size of the spiny stipules: chrysopetala should bear small and weak spiny stipules, while fructifragrans is characterized by large and firm spiny stipules. However, our results reveal that the genetic relationship is not reflected in the size of the stipules.
The AFLP results reveal that all analyzed specimens of V. x heilbornii are genetically related with V. stipulata and, although more distantly, with V. weberbaueri and V. parviflora. Moreover, two of the analyzed V. x heilbornii specimens hold the chlorotype of V. stipulata, while all others contain the same chlorotype as V. weberbaueri, as reported and discussed before by Van Droogenbroeck et al. (2004)
. Morphologically, V. stipulata and V. x heilbornii are difficult to distinguish based on the features studied, because pronounced variability in leaf morphology complicates identification. However, most specimens can be classified with certainty based on fruit shape, color of petals, and general habit.
Although they are not evidenced from this study, indications of the involvement of V. cundinamarcensis in the hybrid formation of V. x heilbornii cannot be denied. Van Droogenbroeck et al. (2002)
established that all individuals analyzed with AFLP clustered together with either V. stipulata or V. cundinamarcensis. Furthermore, the involvement of V. cundinamarcensis is reflected in the morphology of some specimens of V. x heilbornii. The (greenish) yellow flowers, the absence of stipules, and the slightly hairy petioles are the most pronounced features that indicate a relationship with V. cundinamarcensis.
A more detailed and extensive molecular and morphological analysis of V. x heilbornii and its putative parent species is being performed at this moment to clarify this problem (B. Van Droogenbroeck, E. Romeijn-Peeters, W. Van Thuyne, T. Kyndt, P. Goetghebeur, J. P. Romero-Motochi, and G. Gheysen, unpublished data). Based on all available evidence, a common maternal progenitor for V. x heilbornii and V. weberbaueri, or a possible triple hybridization event involving V. stipulata, V. weberbaueri and V. cundinamarcensis, can be hypothesized.
Intraspecific chlorotype variation in V. microcarpa has been revealed earlier by both Aradhya et al. (1999)
and Van Droogenbroeck et al. (2004)
considering relatively small sample sets (two and five specimens). It was again established in this study showing two different chlorotypes in 12 specimens belonging to three described subspecies: microcarpa, baccata, and heterophylla. In addition, nuclear DNA, which was never analyzed in detail before, also shows a very high within-species diversity level and accordingly very distinct grouping in the AFLP dendrogram. Subspecific classification is likewise not correlated with relationships revealed with molecular markers. Consequently, V. microcarpa and its subspecies do not appear to be valid taxa, as molecular data do not support their delimitation. Specimens identified as V. microcarpa are scattered throughout several genetically related groups, sometimes showing genetic affinities with unidentified specimens. As morphological diversity within these groups is very high, a possible hybrid origin is very plausible and will be discussed further in the following section.
Possible hybridization events and identification of unidentified specimens
Sixteen specimens involved in this study could not be identified with the key of Badillo (1993)
. Based on the results of this study, however, some of these specimens could be attributed to a known taxon or can be hypothesized to be of hybrid origin.
Intercompatibility between several Vasconcellea species has already been demonstrated by several studies (Jiménez and Horovitz, 1957
; Horovitz and Jiménez, 1967
; Mekako and Nakasone, 1975
), and natural interspecific hybrids have been observed in areas where species are sympatric (Badillo, 1971
). Of the eight species for which artificial crosses have been analyzed, six have shown to be compatible (V. microcarpa, V. cundinamarcensis, V. stipulata, V. cauliflora, V. monoica, and V. horovitziana), although not always reciprocal. V. goudotiana and V. parviflora are intercompatible but can only be crossed with a few of the abovementioned species. Based on the observed ease of hybridization, it can be hypothesized that interspecific hybrids between Vasconcellea species not analyzed to date might arise in nature. Taking into consideration that plant hybrid zones tend to occur more frequently in disturbed areas (Rieseberg and Ellstrand, 1993
; Rieseberg, 1995
), the human threat perturbing at least five species of Vasconcellea (IUCN, 2003
) might make them more disposed to interspecific hybridization. Hybridization events can be detected molecularly by checking for incongruence between nuclear and cytoplasmic data (Rieseberg, 1995
, 1997
) or by non-concordance between their position in genetic and morphological analyses (Arnold, 1997
). Because hybridization results in progeny with a mosaic of parental, intermediate, and extreme characters (Rieseberg, 1995
) and leads to populations with a wide range of different recombinants and segregating progeny (Barton and Hewitt, 1985
), morphological detection and identification of hybridization events is difficult (Rieseberg and Ellstrand, 1993
). Moreover, as the morphological data presented in this study do not clearly differentiate between some genetically well-confirmed taxa, these morphological results were not always useful in the characterization of possible hybrids.
A clear evidence of incongruence is found in a group of specimens, sp101, sp205, sp312(I), and sp312(II), which are genetically closely related to V. cundinamarcensis based on nuclear AFLP results, although cytoplasmic PCR-RFLP results reveal that they hold the same haplotype as V. palandensis, V. monoica, and V. omnilingua. Preliminary results of ITS sequences (T. Kyndt, B. Van Droogenbroeck, E. Romeijn-Peeters, J. P. Romero-Motochi, X. Scheldeman, P. Goetghebeur, P. Van Damme, and G. Gheysen, unpublished data) show intra-individual sequence heterogeneity for these specimens, suggesting a hybrid origin involving V. monoica and V. cundinamarcensis. Considering that these two species are reported to be compatible (Jiménez and Hovoritz, 1957
), we assume that these specimens belong to the group of superficially described (Horovitz and Jiménez, 1967
; Badillo, 1971
) interspecific hybrids between V. monoica and V. cundinamarcensis (m x c).
Because the haplotype of all other unidentified specimens is correlated with nuclear associations found with AFLP, no further incongruence between nuclear and chloroplast data has been observed. Nevertheless, taking into consideration that our chloroplast data do not reveal an exclusive haplotype for every species it is not unlikely that recently divergent intercompatible species, with the same haplotype, can lead to the production of hybrids with similar nuclear and cytoplasmic characteristics and high morphological variation. Because intercompatibility has not yet been investigated in all Vasconcellea species, it is difficult to draw solid conclusions from our results. In the following paragraph, we point out some different genetically related groups of specimens with high morphological variation as possible hybrids or introgressed individuals between co-occuring species. Though a much more extensive survey of these specimens is necessary to understand their origin and evolution, our results suggest some preliminary conclusions that will be helpful in guiding future research.
For instance, V. longiflora and V. pulchra could be the progenitors of mich177, sp183, micb186, mich190, and micb192, all of which were sampled near the collection site of these two sympatric species. Their extremely high variability revealed in the morphological analysis is probably a result of hybrid segregation. A second group of specimens for which a hybrid origin can be hypothesized reveals a similar high range of morphological variation, while a certain molecular similarity is established: micm065, micm067, micm265, micm266, micm273, sp271, and sp225. A sympatric group of specimens, mich255, sp256, sp257, mich258, sp259, and sp260, collected in Calvario (Azuay, Ecuador), is genetically and morphologically closely associated with V. longiflora (mich255, sp256, and sp257 were not included in the morphological analysis due to absence of mature leaves). Our results indicate that these specimens belong to the taxon V. longiflora or are a divergent population or species, resulting from introgression or recent radiation. Molecular data reveal a distant genetic affinity between specimens sp200 and sp203 and (1) V. cundinamarcensis, sharing the same cytoplasmic haplotype, and (2) the hybrid population between V. monoica and V. cundinamarcensis (sp101, sp205, sp312(I), sp312(II): m x c). As V. cundinamarcensis is known to be intercompatible with at least four other Vasconcellea species, sp200 and sp203 might be hybrids different from V. x heilbornii and the m x c hybrid population. Finally, a group of specimens involving sp239, sp240, and sp241 has a strong molecular and morphological association with the co-occuring species V. omnilingua. Although these three specimens show some variation in leaf morphology, not described by Badillo (1993)
, genetic analysis revealed a close relationship confirmed by a high bootstrap value (82%). It is possible that the taxon description of V. omnilingua (Badillo, 1993
) should be extended to include a higher range of morphological variation, but then again, their genetic association might also be a result of introgression or gene flow.
Based on the sufficient number of possible hybrids found among the analyzed samples, contemporary hybridization events leading to introgression between co-occuring plants are estimated to occur very frequently in the genus Vasconcellea. Because all plants identified as V. microcarpa by the key of Badillo (1993)
belong to one of the proposed hybrid groups, our data suggest that this taxon is actually a combination of several hybrids from diverse origins. Investigation of faster evolving chloroplast sequences that are able to differentiate between closely related species together with an extended morphological study of these suggested hybrid specimens is needed to further investigate their origin. Moreover, more extensive intercompatibility analyses could improve our understanding of hybridization in the genus Vasconcellea.
In conclusion, this study suggests that evolution in Vasconcellea is likely to involve reticulation, introgression, and recent speciation. Our results clearly demonstrate that the taxon descriptions of Badillo (1993)
are not complete and that they need a thorough revision for certain taxa. In general, we have shown that molecular marker techniques are very useful in resolving morphological identification problems in this genus.
|
1 The authors thank Prof. R. Drew, Dr. T. Fichet Lagos, Dr. J. Bigirimana, Dr. T. Thuan and Dr. F. Zee NPGS, USA for providing leaf material and the herbaria HUA, Columbia; BM, United Kingdom; U, The Netherlands for giving us authorization to use their specimens as source of DNA. The authors thank Dr. Ir. Xavier Scheldeman for valuable advice and useful discussions. Financial support for this research was provided by a grant to Tina Kyndt from the Institute for the Promotion of Innovation through Science and Technology in Flanders IWT-Vlaanderen, and by the FWO-Vlaanderen Project no. 3G005100. 
6 Present address: Flanders Interuniversity Institute for Biotechnology, UGent, Technologiepark 927, B-9052 Ghent, Belgium
7 Author for reprint requests (e-mail: Paul.Goetghebeur{at}ugent.be
)
8 Both authors equally contributed to this work
LITERATURE CITED
Aradhya M. K. R. M. Manshardt F. Zee C. W. Morden 1999 A phylogenetic analysis of the genus Carica L. (Caricaceae) based on restriction fragment length variation in a cpDNA intergenic spacer region. Genetic Resources and Crop Evolution 46: 579-586[CrossRef][ISI]
Arnold M. L. 1997 Natural hybridisation and evolution. Oxford University Press, New York, New York, USA
Badillo V. M. 1971 Monografía de la familia Caricaceae. Publicada por la Asociación de Profesores, Universidad Central de Venezuela, Maracay, Venezuela
Badillo V. M. 1993 Caricaceae, segundo esquema. Revista de la Facultad de Agronomía de la Universidad de Venezuela 43: 1-111
Badillo V. M. 1997 Neotipificación de Carica pubescens Lenné et Koch y de Carica quercifolia (St. Hil.) Hieron. y nuevos registros de la familia para Ecuador. Ernstia 6: 201-205
Badillo V. M. 1999 Nuevos registros para el Ecuador de especies del género Carica. Ernstia 9: 185-186
Badillo V. M. 2000 Carica L. vs. Vasconcella St. Hil. (Caricaceae) con la rehabilitación de este último. Ernstia 10: 74-79
Badillo V. M. 2001 Nota correctiva Vasconcellea St. Hill. y no Vasconcella (Caricaceae). Ernstia 11: 75-76
Barton N. H. G. M. Hewitt 1985 Analysis of hybrid zones. Annual Review of Ecology and Systematics 16: 113-148[CrossRef][ISI]
Burnham R. J. A. Graham 1999 The history of neotropical vegetation: new developments and status. Annals of the Missouri Botanical Garden 86: 546-589[CrossRef][ISI]
Felsenstein J. 1985 Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783-791[CrossRef][ISI]
Gower J. C. 1966 Some distance properties of latent root and vector methods used in multivariate analysis. Biometrika 53: 325-338
Hillis D. M. 1987 Molecular versus morphological approaches to systematics. Annual Review of Ecology and Systematics 18: 23-42
Horovitz S. H. Jiménez 1967 Cruziamentos interespecificos e intergenericos en Caricaceas y sus implicaciones fitotecnicas. Agronomía Tropical 17: 323-344
IUCN—The World Conservation Union. 2003 The IUCN Red List of Threatened Species. IUCN Species Survival Commission. IUCN, Gland, Switzerland and Cambridge, UK. Available at website, http://www.iucnredlist.org/
Jaccard P. 1908 Nouvelles recherches sur la distribution florale. Bulletin de la Societé Vaudoise de Sciences Naturelles 44: 223-270
Jiménez H. S. Horivitz 1957 Cruzabilidad entre especies de Carica. Agronomía Tropical 7: 207-215
Jiménez Y. J. Romero X. Scheldeman 1998 Colección, caracterización y descripción de Carica x heilbornii nm. pentagona B.; Carica pubescens (A.DC.) Solms-Laub y Carica stipulata B. en la provincia de Loja. Revista de Difusión Técnica y Cientifica de la Facultad de Ciencias Agricolas de la Universidad Nacional de Loja 29: 43-54
Jobin-Decor M. P. G. C. Graham R. J. Henry R. A. Drew 1997 RAPD and isozyme analysis of genetic relationships between Carica papaya and wild relatives. Genetic Resources and Crop Evolution 44: 471-477
Karp A. O. Seberg M. Buiatti 1996 Molecular techniques in the assessment of botanical diversity. Annals of Botany 78: 143-149
Kim M. S. P. H. Moore F. Zee M. M. M. Fitch D. L. Steiger R. M. Manshardt R. E. Paull R. A. Drew T. Sekioka R. Ming 2002 Genetic diversity of Carica papaya L. as revealed by AFLP markers. Genome 45: 503-512[Medline]
Mantel N. 1967 The detection of disease clustering and a generalized regression approach. Cancer Research 27: 209-220[ISI][Medline]
McLenachan P. A. K. Stöckler R. C. Winkworth K. McBreen S. Zauner P. J. Lockhart 2000 Markers derived from amplified fragment length polymorphism gels for plant ecology and evolution studies. Molecular Ecology 9: 1899-1903[CrossRef][Medline]
Mekako H. U. H. Y. Nakasone 1975 Interspecific hybridization among 6 Carica species. Journal of the American Society of Horticultural Science 100: 237-242
National Research Council. 1989 Highland papayas. In F. R. Ruskin [ed.], Lost crops of the Incas: little-known plants of the Andes with promise for worldwide cultivation, 252–261. National Academy Press, Washington, D.C., USA
Rieseberg L. H. 1995 The role of hybridization in evolution: old wine in new skins. American Journal of Botany 82: 944-953[CrossRef][ISI]
Rieseberg L. H. 1997 Hybrid origins of plant species. Annual Reviews in Ecology and Systematics 28: 359-389[CrossRef]
Rieseberg L. H. N. C. Ellstrand 1993 What can molecular and morphological markers tell us about plant hybridisation?. Critical Reviews in Plant Science 12: 213-241[CrossRef]
Rohlf F. J. 2000 NTSYS-pc. Numerical taxonomy and multivariate analysis system, version 2.1. Exeter Software, Setauket, New York, USA
Romeijn-Peeters E. 2004 Biodiversity of the genus Vasconcellea (Caricaceae) in Ecuador: a morphological approach. Ph.D. dissertation, Faculty of Sciences, Ghent University (UGent), Ghent, Belgium
Scheldeman X. 2002 Distribution and potential of cherimoya (Annona cherimola Mill.) and highland papayas (Vasconcellea spp.) in Ecuador. Ph.D. dissertation, Faculty of Agricultural and Applied Biological Sciences, Ghent University (UGent), Ghent, Belgium
Sneath P. H. A. R. R. Sokal 1973 Numerical taxonomy: the principles and practice of numerical classification. W.H. Freeman, San Francisco, California, USA
Sokal R. R. C. D. Michener 1958 A statistical method for evaluating systematic relationships. University of Kansas Science Bulletin 38: 1409-1438
Van De Peer Y. R. De Wachter 1994 TREECON for Windows: a software package for the construction and drawing of evolutionary trees for the Microsoft Windows environment. Computational Applications in Bioscience 10: 569-570
Van Droogenbroeck B. P. Breyne P. Goetghebeur E. Romeijn-Peeters T. Kyndt G. Gheysen 2002 AFLP analysis of genetic relationships among papaya and its wild relatives (Caricaceae) from Ecuador. Theoretical and Applied Genetics 105: 289-297
Van Droogenbroeck B. T. Kyndt I. Maertens E. Romeijn-Peeters X. Scheldeman J. P. Romero-Motochi P. Van Damme P. Goetghebeur G. Gheysen 2004 Phylogenetic analysis of the highland papayas (Vasconcellea) and allied genera (Caricaceae) using PCR-RFLP. Theoretical and Applied Genetics 108: 1473-1486[CrossRef][ISI]
Vos P. R. Hogers M. Bleeker M. Reijans T. Van De Lee M. Hornes A. Frijters J. Pot J. Peleman M. Kuiper M. Zabeau 1995 AFLP: a new technique for DNA fingerprinting. Nucleic Acids Research 23: 4407-4414
Weising K. H. Nybom K. Wolff W. Meyer 1995 DNA fingerprinting in plants and fungi. CRC Press, Boca Raton, Florida, USA
This article has been cited by other articles:
|
|
 |
|
  B. VAN DROOGENBROECK, T. KYNDT, E. ROMEIJN-PEETERS, W. VAN THUYNE, P. GOETGHEBEUR, J. P. ROMERO-MOTOCHI, and G. GHEYSEN Evidence of Natural Hybridization and Introgression between Vasconcellea Species (Caricaceae) from Southern Ecuador Revealed by Chloroplast, Mitochondrial and Nuclear DNA Markers Ann. Bot., May 1, 2006; 97(5): 793 - 805. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. E. ASSOGBADJO, T. KYNDT, B. SINSIN, G. GHEYSEN, and P. VAN DAMME Patterns of Genetic and Morphometric Diversity in Baobab (Adansonia digitata) Populations Across Different Climatic Zones of Benin (West Africa) Ann. Bot., May 1, 2006; 97(5): 819 - 830. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
