| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Systematics |
Centro de Conservación y Mejora de la Agrodiversidad Valenciana, Universidad Politécnica de Valencia, Camino de Vera 14, 46022 Valencia, Spain
Received for publication May 16, 2003. Accepted for publication July 24, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: amplified fragment length polymorphisms (AFLPs) • endangered species • Galápagos Islands • genetic resources • Lycopersicon cheesmanii • Lycopersicon esculentum • Solanaceae • tomato
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Nuez, 1995
). Consequently, conservation of wild species of Lycopersicon, not only ex situ in germplasm banks, but also in situ in the form of natural populations that conserve the genes in their place of origin, is a high priority.
Multiple molecular data sets indicate that tomatoes are deeply nested with Solanum and because of this, Spooner et al. (1993)
transferred them to Solanum sect. Lycopersicon and provided new combinations for many of the Lycopersicon species. However sound the phylogenetic interpretation, the taxonomic treatment of tomatoes within the genus Solanum is debatable, and both options are open (Lester, 1991
; Spooner et al., 1993
). The argument that because tomatoes have evolved from within Solanum and thus must be included in Solanum to produce a monophyletic phylogenetic classification and avoid paraphyly is subject to discussion; a Linnaean classification without paraphyletic taxa is a logical impossibility (Brummitt, 2002
; R. N. Lester, University of Birmingham, personal communication). For this reason, for the purpose of clarity, and because numerous historical and present references are cited within this report, we accept the paraphyly and use the original Lycopersicon names.
The genus Lycopersicon is divided into two subgenera, Eulycopersicon and Eriopersicon (Muller, 1940
). The former includes the red-, orange-, or yellow-fruited species L. esculentum, L. cheesmanii Riley, and L. pimpinellifolium (Jusl.) Mill. Eriopersicon comprises the green-fruited species L. chilense Dun., L. chmielewskii Rick, Kesicki, Fobes & Holle, L. hirsutum Humb. & Bonpl., L. parviflorum Rick, Kesicki, Fobes & Holle, L. pennellii (Corr.) D'Arcy, and L. peruvianum (L.) Mill. (Rick, 1979
; Taylor, 1986
; Warnock, 1988
). Although the cultivated tomato was probably domesticated in Mexico (Jenkins, 1948
), all wild species are naturally distributed in mainland South America (Warnock, 1991
; Esquinas-Alcázar and Nuez, 1995
), except for L. cheesmanii, which is endemic to the Galápagos Islands (Rick, 1956
). Like most of the Galápagos flora and fauna, L. cheesmanii has evolved in isolation from mainland species and has acquired a combination of morphological and physiological characteristics unique in the genus, such as orange-red or orange-yellow fruits, yellow-green foliage, minute seed size, and exceptional seed dormancy (Mackinney et al., 1954
; Rick, 1956
, 1963
).
Lycopersicon cheesmanii is of great interest for tomato breeding: it is easily crossed with the cultivated tomato to produce fertile offspring, and it is a source of variation for several traits of agronomic interest (Rick, 1979
; Nuez, 1995
). Some forms of L. cheesmanii that thrive near the coastline in extremely arid and saline environments are sources of tolerance to salinity and to drought (Rush and Epstein, 1981
). Lycopersicon cheesmanii is also of great interest in breeding for quality as it has high sugar (Hewitt and Garvey, 1987
; Poysa, 1993
) and ß-carotene contents (Mackinney et al., 1954
; Stommel, 2001
). The jointless pedicel of the fruit of some accessions, a trait conferred by the gene j – 2, has also been used in breeding processing tomatoes as it facilitates separation from the calyx during mechanical harvest (Rick, 1967
). Other L. cheesmanii accessions are resistant or tolerant to pathogens such as Alternaria alternata (Cassol and St. Clair, 1994
) or tomato yellow leaf curl virus (Hassan et al., 1984
; Picó et al., 1996
). Unfortunately, introduced plants and animals, particularly feral donkeys and goats, and human activity threaten L. cheesmanii populations (Rick and Bowman, 1961
; Jackson, 1994
). Therefore, conservation of L. cheesmanii is urgently needed.
Within L. cheesmanii, one well-defined infraspecific form (L. cheesmanii forma minor), characterized by having a distinctive syndrome of morphological traits, can be distinguished (Rick, 1983
). Plants of L. cheesmanii f. minor can be easily identified by their three- to four-pinnately compound leaves. Rick (1980)
demonstrated that this trait is under the control of a single gene (Pts) with incomplete dominance. In the remainder of the species (one- to two-pinnately compound leaves), which is considered as the typical or straight L. cheesmanii, the assortment of many traits of taxonomic interest is too heterogeneous to permit classification into subgroups (Rick, 1983
). However, two different forms without formal taxonomic designation can be distinguished; they vary greatly for a trait (internode length) that has a large impact on the plant morphology. The two forms also have different habitat preferences (coastal or inland). One, which we have labeled L. cheesmanii ‘short,’ corresponds to forms with short internodes (2–4 cm) and usually thrives near coastal areas. The other, which we have called L. cheesmanii ‘long,’ has longer internodes (5–8 cm) and usually grows inland. Distinction between these two forms, despite not corresponding to different formal taxa, may be of interest to breeders. In addition to these endemic forms, typical cultivated tomato (L. esculentum) as well as weedy cherry tomato (L. esculentum var. cerasiforme (Dun.) Gray), have also been reported from the islands (Rick, 1956
). Recently, we also found wild and weedy populations of plants that bear small (<15 mm) red fruits and seem to be derivatives of the introduced L. esculentum var. cerasiforme. We have given them the provisional name L. esculentum ‘Gal cer’ to distinguish them from the typical feral forms of tomato L. esculentum var. cerasiforme, which have larger fruits (>15 mm).
Studies on the molecular diversity of Galápagos tomatoes and their implications for conservation have been limited. The only comprehensive work was performed by Rick and Fobes (1975)
with allozymes. They found that diversity among L. cheesmanii accessions was high, but that within-accession allozyme variation in L. cheesmanii was very low and comparable to the degree of genetic uniformity characteristic of pure lines of domesticated L. esculentum. This is expected as L. cheesmanii stigmas are inserted or only slightly exserted, which leads to autogamy, and because of the consequent rapid fixation of alleles and differentiation between populations (Rick, 1983
). Since the valuable contributions of Rick and co-workers (Rick, 1956
, 1963
, 1971
; Rick and Bowman, 1961
; Rick and Fobes, 1975
) on the biosystematics, diversity, relationships with the continental species, and evolution of Galápagos tomatoes, there have been no further studies on these subjects for L. cheesmanii. Although L. cheesmanii has been included in many studies on the molecular diversity of the genus Lycopersicon, in general, only one or a few accessions have been studied at a time (Palmer and Zamir, 1982
; Miller and Tanksley, 1990
; Bretó et al., 1993
; Williams and St. Clair, 1993
; Álvarez et al., 2001
; Marshall et al., 2001
; Peralta and Spooner, 2001
; Nesbitt and Tanksley, 2002
).
The optimal use of genetic resources in present and future breeding requires implementing strategies to conserve wild germplasm. Because of the unique characteristics of Galápagos tomatoes, their utility for tomato breeding, and their susceptibility to genetic erosion, it is necessary to have updated studies on their characteristics, distribution, relationships, origin, and diversity. Amplified fragment length polymorphisms (AFLPs) give a high level of polymorphism in Solanaceae (Milbourne et al., 1997
; Kardolus et al., 1998
), allow a large number of loci to be scored in a single reaction, and have a much better repeatability among laboratories than other markers such as random amplified polymorphisms of DNA (RAPDs) (Jones et al., 1997
). Therefore, AFLPs may be useful to address these issues in Galápagos tomatoes.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
, 1963
, 1971
, 1983
; Rick and Fobes, 1975
; C. M. Rick Tomato Genetics Resource Center [TGRC] passport data website: http://tgrc.ucdavis.edu). Apart from the search for coastal populations, we also prospected inland on Santa Cruz Island.
|
). Longitude and latitude of each population were determined using a global positioning system (GPS; Magellan Meridian XL, San Dimas, California). Altitude was measured either with the GPS or with a barometric altimeter.
Plant materials used in the analysis of variation
The plant materials used to study both morphological and molecular variation consisted of 43 accessions. Of these, 15 were wild accessions we collected in the Galápagos, selected to represent the whole range of different morphological forms found (L. cheesmanii ‘short,’ L. cheesmanii ‘long,’ L. cheesmanii f. minor, and L. esculentum ‘Gal cer’) and the different collection sites; eight were L. cheesmanii accessions from the TGRC, representative of the different forms stored in this gene bank (L. cheesmanii and L. cheesmanii f. minor); three were accessions of weedy L. esculentum var. cerasiforme collected in the Galápagos; 11 were accessions from the continental L. pimpinellifolium, the wild species genetically closest to L. cheesmanii (Rick and Fobes, 1975
; Esquinas-Alcázar and Nuez, 1995
), and one accession of cultivated L. esculentum, and of each of the wild subgenus Eriopersicon species: L. chilense, L. hirsutum, L. parviflorum, L. pennellii, and L. peruvianum (Table 1). Pairwise geographic distances of selected accessions were calculated either from the GPS position, if available, or from the approximate position of the collection site.
|
Growing conditions
Seeds were subjected to thermotherapy (24 h at 80°C) to avoid infection by seed-transmitted pathogens. Seeds were germinated on moistened filter paper in petri dishes. Seed testa of some Galápagos accessions were excised to ensure successful germination (Rick, 1956
; Rick and Bowman, 1961
).
Germinated seeds were transferred to seedling trays and were kept in a growth chamber until their transplantation to a glasshouse. Plants were spaced at 0.3 m within the row and 1 m between rows, trained with vertical strings, and watered by drip irrigation.
To improve fruit set, inflorescences were shaken three times a week with a mechanical vibrator to favor the release of pollen from anthers. For each of the accessions of the self-incompatible species L. chilense, L. hirsutum, L. pennellii, and L. peruvianum, hand pollinations using a mixture of pollen of several plants from the same accession were performed to ensure some fruit set.
Morphological characterization
Six plants per accession were characterized using the International Plant Genetic Resources Institute (IPGRI) descriptors list for tomato (IPGRI, 1996
). Characters evaluated included qualitative and quantitative traits of the plants, flowers, inflorescences, and fruit. Mean values for each accession were obtained as the mean of six plants, for which a variable number of flowers/fruits per plant was evaluated.
DNA extraction and AFLP analysis
For each accession, genomic DNA was isolated from a mixture of young leaves from six plants using the DNeasy Plant Mini Kit (Qiagen, Valencia, California, USA) following the protocol of the manufacturer. DNA concentration was quantified on agarose, and a 0.25-µg DNA sample was digested by the enzyme combination EcoRI and MseI at 37°C for 2 h. Ligation was performed with the AFLP Core Reagent Kit (Invitrogen, Carlsbad, California, USA) following the instructions of the manufacturer. After ligation, the reaction mixture was diluted 1 : 10 in Tris-EDTA (TE) buffer.
For the preselective amplification, a 5-µL aliquot from the DNA dilution was added to a 25-µL solution containing 2.5 µL of 10x buffer, 0.5 µL of primer EcoA (10 µmol/L), 0.5 µL of primer MseC (10 µmol/L), 1.0 µL of dNTPs (10 mmol/L), and 0.8 units of Taq polymerase (Roche, Basel, Switzerland). After preamplification, DNA was diluted again 1 : 10 in TE buffer. The selective amplification was performed on 2-µL aliquots using four combinations of primers (Table 2). DNA fragments were separated in an ABI Prism 310 genetic analyzer (Applied Biosystems, Foster City, California, USA). Resulting fragments were scored as binary traits (1 = present, 0 = absent) using Genographer 1.6 software (Montana State University, Bozeman, Montana, USA).
|
test.
The resulting genetic similarity matrices were used to generate unweighted pair group method using arithmetic means (UPGMA) phenograms using the NTSYSpc2.0 software package (Rohlf, 1996
). The reliability and robustness of the phenograms were tested by bootstrap analysis with 1000 replications to assess branch support. A limit of 50% was used to indicate statistical support for the topology at a node (Highton, 1993
). A principal coordinate analysis (PCA) based on genetic similarity matrices was performed using the DCENTER and EIGEN algorithms of the NTSYSpc2.0 software package.
Genetic diversity was estimated with the proportion of polymorphic fragments (Pp) and the total diversity (HT) (Nei, 1973
). Total diversity was partitioned into diversity among (DST) and within (HS) groups. The relative magnitude of gene differentiation among groups (GST) was calculated as the ratio DST/HS (Nei, 1973
). The estimates of genetic identity (J) and standard genetic distance (D) were calculated as indicated in Nei (1974)
.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). However, information collected during the expedition, together with subsequent observations of plants in cultivation and previously published information on the subject, allowed us to distinguish three forms of L. cheesmanii with different gross morphology (Fig. 2) and/or ecological preferences. Because these differences among the forms of L. cheesmanii might affect their adaptation and genetic differentiation, we have studied them separately. This separation will give more information of relevance regarding the infraspecific variation, diversity, and conservation strategies than if all L. cheesmanii accessions were forced into a single group. Similarly, the newly reported L. esculentum ‘Gal cer’ has been studied as a separate entity of L. esculentum var. cerasiforme to provide more information on its origin and diversity. The main characteristics and differences between these wild forms are the following:
|
, 1963
, 1971
; Rick and Bowman, 1961
; Rick and Fobes, 1975
). Morphologically, these plants are very similar to L. esculentum var. cerasiforme, but fruits are much smaller (always <15 mm). They are found inland, in areas of high soil humidity in highly disturbed areas (road and trail margins, works areas, dumps, gardens). Like domesticated L. esculentum and feral L. esculentum var. cerasiforme, they do not need long (or bright) days to flower. The ripe fruit is deep red and is attractive to the Galápagos mockingbird (Nesomimus parvulus), which we observed carrying ripe fruits in their beaks.
Apart from these four wild forms, on the inhabited islands we also found cultivated (L. esculentum) and feral (L. esculentum var. cerasiforme) tomato. The feral tomato is found as a weed in vegetable gardens and trail margins. Galápagos accessions present the typical traits of this taxon (Fig. 2e).
There are important differences among L. cheesmanii, L. esculentum, and the wild continental species L. pimpinellifolium in the quantitative morphological traits measured in this work (Table 3). Internode length is much shorter in L. cheesmanii ‘short’ and L. cheesmanii f. minor than in the other taxa. Although there are not absolute differences in the size of the flower among the taxa, L. cheesmanii has a lower petal length to sepal length ratio than L. esculentum ‘Gal cer’ or L. pimpinellifolium (Table 3). Another remarkable feature of L. cheesmanii is that stigmas are only slightly exserted (always less than 0.4 mm), much less than stigmas of L. esculentum or L. pimpinellifolium (Table 3). There is also an important difference between L. cheesmanii and the other Eulycopersicon taxa in the flowering requirements. Although in their native environment and in other places at lower latitudes than Valencia, such as Davis, California (R. Chetelat, TGRC, personal communication), L. cheesmanii flowers all the year round if the conditions are favorable for plant growth and development, in the latitudes of Spain (40° N) they do not flower from mid-November to mid-February.
|
Survey of Lycopersicon populations in the Galápagos Islands
A total of 37 populations of Lycopersicon were found in distinct geographical locations in our field work in 2000 (Fig. 1). Of these, 12 were coastal populations, with three of these belonging to L. cheesmanii ‘short’ and nine to L. cheesmanii f. minor found in environments undisturbed by humans. All of the L. cheesmanii ‘short’ populations that we found were situated in the same geographical area (northern channel of Baltra Island). Lycopersicon cheesmanii f. minor was more widespread. The largest coastal populations were found on the small islands of Bartolomé and Sombrero Chino, where there are no introduced herbivorous animals. On these islands, the plants were found in lava flow crevices, and in general, populations consisted of many plants (>50), most young and vigorous. On the tiny island of Corona del Diablo (the crown of a volcano in the sea) we found only two very old plants. On islands with feral goats or donkeys, such as Santiago and Baltra, we could only find plants on cliffs and other places inaccessible to these animals. In these cases, populations are formed by a few individuals, in many cases by a single plant. Most of the plants are very old, and new shoots sprout from the enlarged base of the stem.
Regarding coastal populations that were found on collecting expeditions from 1950 to 1974 and that are maintained in the TGRC, we found again populations in the north of Baltra, Bartolomé, north of Bahía James (in Santiago), and Corona del Diablo. However, despite exhaustive searches we found neither living plants nor dead remains of the previously reported L. cheesmanii populations on Santa Cruz (in the vicinity of the village of Puerto Ayora), Gardner, Pinzón, or Rábida. On the other hand, we discovered an important population not reported previously on Sombrero Chino Island.
Inland Santa Cruz yielded many Eulycopersicon populations (21), in many cases formed by hundreds of individuals spreading over tens of meters. In general, in contrast to the coastal populations, these populations are found in environments highly disturbed by humans. Most of the populations (18) corresponded to L. esculentum ‘Gal cer,’ while only three corresponded to L. cheesmanii ‘long.’ The few L. cheesmanii ‘long’ populations that we found occurred in the vicinity of or were intermingled with L. esculentum ‘Gal cer.’ On Santa Cruz Island, we also collected two samples of L. esculentum var. cerasiforme, present as a weed in vegetable gardens, and two accessions of cultivated L. esculentum.
AFLP analysis
A total of 593 fragments were scored, of which 534 (90.1%) were polymorphic. Of these, 222 were specific to subgenus Eriopersicon. Specific to subgenus Eulycopersicon were 112 fragments: of these 31 were only found in one or more accessions of L. cheesmanii, 21 were present only in L. esculentum, and 54 in L. pimpinellifolium, but none were limited to L. esculentum ‘Gal cer.’ When considering L. cheesmanii alone, 28 fragments were unique to L. cheesmanii ‘short,’ six to L. cheesmanii ‘long,’ but none were specific to L. cheesmanii f. minor. Within L. esculentum, we found 16 fragments in domesticated L. esculentum but not in L. esculentum var. cerasiforme and 45 fragments in L. esculentum var. cerasiforme but not in cultivated L. esculentum. Regarding fragments consistently present in all the accessions of one taxon of subgenus Eulycopersicon but not in any accession of the other taxa, we found one in L. cheesmanii, three in L. pimpinellifolium, but none in L. esculentum or any of its varieties.
Two fragments were shared by some L. esculentum ‘Gal cer’ and L. esculentum var. cerasiforme but not by the rest of members of the genus. One of these fragments was present in L. esculentum var. cerasiforme LA3123 coming from the Santa Cruz Island and also in all of the accessions of L. esculentum ‘Gal cer.’ Another fragment was present in the two L. esculentum var. cerasiforme from Santa Cruz (GLP-26 and LA3123) and in six of seven L. esculentum ‘Gal cer’ accessions. This shows that L. esculentum var. cerasiforme is genetically very similar to L. esculentum ‘Gal cer.’
The total diversity (HT) value for subgenus Eriopersicon (HT = 0.248) was more than three times higher than that for Eulycopersicon (HT = 0.072), indicating that the former is much more diverse than the latter (Table 4). Within subgenus Eulycopersicon, the highest gene diversity is found in L. pimpinellifolium (HT = 0.071). Lycopersicon esculentum (domesticated plus L. esculentum var. cerasiforme) and L. cheesmanii have similar values (Table 4). In contrast, L. esculentum ‘Gal cer’ has a very low value for gene diversity (HT = 0.014). Gene differentiation among L. esculentum var. cerasiforme (plus cultivated L. esculentum), L. esculentum ‘Gal cer,’ L. pimpinellifolium, and L. cheesmanii is the highest found in this work (GST = 0.462). When considering the three L. cheesmanii forms, the highest diversity is found within L. cheesmanii ‘short’ (HT = 0.060) with a value almost twice that of L. cheesmanii ‘long’ and L. cheesmanii f. minor. These three forms have relatively low gene differentiation (GST = 0.235), although higher than the differentiation between subgenera Eriopersicon and Eulycopersicon (GST = 0.190). This low value is due to the huge variation within subgenus Eriopersicon more than to similarity between subgenera Eriopersicon and Eulycopersicon.
|
|
The cluster analysis shows a clear separation between the much more divergent subgenus Eriopersicon and subgenus Eulycopersicon (Fig. 3). Within Eulycopersicon, four main clusters, supported by a bootstrap value of 99%, were formed by L. cheesmanii, L. pimpinellifolium, cultivated L. esculentum, and L. esculentum var. cerasiforme plus L. esculentum ‘Gal cer.’ A phenogram constructed using only Eulycopersicon accessions was very similar in topology to that obtained when using accessions of Eriopersicon and Eulycopersicon together.
|
Principal components analysis (PCA)
The first PCA gave results very similar to those obtained with the cluster analysis. However, because the first component (61.6% of total variation) basically accounted for the separation from subgenus Eulycopersicon of subgenus Eriopersicon, a new PCA was performed including only Eulycopersicon accessions (Fig. 4). In this analysis, the first, second, and third components explain 28.5, 25.0, and 16.7%, respectively, of the total variation.
|
Correlations between geographical and genetic distances
Because no clear pattern of differentiation by geographical origin was observed from the cluster or PCA for the wild Galápagos tomatoes, within forms with a greater number of representatives (L. cheesmanii f. minor and L. esculentum ‘Gal cer’) we studied the correlation between geographic and genetic distance between pairs of accessions. In neither case did we find a significant correlation between the two variates, indicating that there is no relationship between the geographic distance that separates the place of origin of each accession and the genetic distance between them (Fig. 5).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
At the molecular level, the Galápagos L. cheesmanii can be unambiguously separated from the other species of the genus. However, there is no clear molecular distinction among the three L. cheesmanii forms that we have considered here, indicating that all native Galápagos tomatoes (L. cheesmanii) share a common genetic background. This supports the idea of Rick (1983)
and R. Chetelat (TGRC, personal communication) that the heterogeneity for many traits among the different accessions of L. cheesmanii does not permit a rational classification into genetically differentiated subgroups. Perhaps a few genes are responsible for the gross morphological differences (leaf subdivision and internode length) among the forms we have considered. For instance, the highly dissected leaf trait characteristic of L. cheesmanii f. minor is controlled by a single gene (Rick, 1980
). In addition, some point mutations result in short internode phenotypes (Zeewart, 1984
; Nadzhimov et al., 1988
), a trait that distinguishes L. cheesmanii ‘short’ and L. cheesmanii f. minor from L. cheesmanii ‘long.’ Nonetheless, these forms have important differences in some traits, such as soluble solids concentration, that have a complex inheritance (Saliba-Colombani et al., 2001
) and that are very interesting for tomato breeding. Therefore, at least from a breeders' point of view, separating these three forms is of clear utility.
Lycopersicon esculentum ‘Gal cer’ is morphologically and genetically very similar to L. esculentum var. cerasiforme. However, the fruit size is much smaller than that of typical L. esculentum var. cerasiforme and similar to L. pimpinellifolium. In this respect, we agree with the observation of R. Chetelat (TGRC, personal communication) that fruit size alone should not be used to distinguish between red-fruited Eulycopersicon taxa. Because the diameter of the fruit of L. esculentum var. cerasiforme and L. pimpinellifolium may overlap and is subject to environmental influence, a combination of traits should be used instead.
Origin of Galápagos tomatoes
The molecular and morphological analyses show that, as Rick (1956)
proposed and subsequent studies confirmed, L. cheesmanii is indeed a member of the subgenus Eulycopersicon (Palmer and Zamir, 1982
; Miller and Tanksley, 1990
; Bretó et al., 1993
; Williams and StClair, 1993
; Álvarez et al., 2001
; Marshall et al., 2001
; Peralta and Spooner, 2001
; Nesbitt and Tanksley, 2002
). Lycopersicon cheesmanii shares most molecular markers with L. pimpinellifolium and L. esculentum, i.e., they are obviously sister species and therefore appropriately placed in the same subgenus. The fact that genetic diversity within L. cheesmanii is almost as great as that found in L. pimpinellifolium suggests that Lycopersicon migrated to the Galápagos some time ago. It is unknown when the successful establishment that resulted in the present day L. cheesmanii occurred. However, Nesbitt and Tanksley (2002)
suggest, on the basis of DNA sequences, that the three Eulycopersicon species (L. cheesmanii, L. esculentum, and L. pimpinellifolium) diverged from a common ancestor approximately one million years ago. Therefore, this is the minimum limit for the establishment of the Eulycopersicon that resulted in the present day Galápagos L. cheesmanii. Because L. cheesmanii f. minor is characterized by a derived trait (conferred by gene Pts), it is probably more recent in origin. Lycopersicon cheesmanii f. minor is adapted to a very specific environment, which would account for its lower diversity despite its relatively greater abundance. Whether the Pts gene is directly involved in the greater adaptation to coastal conditions or whether it has simply become fixed as a result of inbreeding remains to be determined.
Regarding L. esculentum ‘Gal cer,’ the lack of references to this type of tomato in previous studies (Rick, 1956
, 1963
), and its low genetic diversity suggest an extremely recent origin. Although their fruit size is similar to L. pimpinellifolium, the molecular markers for L. esculentum ‘Gal cer’ are very different from the accessions of L. pimpinellifolium we used. In fact, the genetic distance of L. pimpinellifolium from L. esculentum ‘Gal cer’ is greater than that separating L. pimpinellifolium from L. cheesmanii, two very clearly different species. Our work suggests that L. esculentum ‘Gal cer’ appeared after introduction and naturalization of L. esculentum. The presence of the weedy L. esculentum var. cerasiforme has been documented in the inhabited island of Santa Cruz since 1952 (Rick, 1956
). R. N. Lester (University of Birmingham, personal communication) suggested that natural selection by frugivorous animals might have exerted a strong selection pressure favoring a reduction in the fruit size of L. esculentum var. cerasiforme, which would have resulted in the L. esculentum ‘Gal cer’ form.
Present distribution and genetic erosion of Galápagos tomatoes
Isolation from the continent and the singular ecological and climatological conditions under which the Galápagos tomatoes have evolved give them a high value both as genetic resources for the improvement of cultivated tomato and for the study of the evolution of an island endemic from continental ancestors. Unfortunately, the recent introductions of animals and plants, as well as human pressures, have put the endemic Galápagos L. cheesmanii at risk of extinction.
We located populations of Galápagos tomatoes conforming to L. cheesmanii and L. cheesmanii f. minor (Rick, 1963
). Because intrapopulation variation in these forms is negligible (Rick and Fobes, 1975
) but interpopulation variation is relatively high, extinction of local populations results in the irreversible loss of genetic material. Several populations reported from the 1950s to the 1970s have not been found since then despite exhaustive searches. Other populations have been rediscovered, but in some cases, they are endangered; e.g., in Santiago, L. cheesmanii plants are found only in areas out of reach of feral goats and donkeys, and the populations usually consist of a few old plants. In other cases, such as in the vicinity of the village of Puerto Ayora in Santa Cruz, where several populations were previously reported (Mackinney et al., 1954
; Rick, 1956
), the increase in human activity has probably been responsible for their complete disappearance. Our observations point to feral goats and donkeys and expanded human activity (e.g., habitat loss) as the main factors responsible for elimination of individual populations. We have found only one population (on Sombrero Chino) not reported previously, which is further evidence of elimination of populations, particularly coastal forms.
Some inland populations on Santa Cruz, which we have called L. esculentum ‘Gal cer,’ are made up of many individuals, even hundreds in some cases. These L. esculentum ‘Gal cer’ populations, not previously described, are found in highly disturbed areas, on occasions intermingled with L. cheesmanii ‘long.’ Our observations suggest that native L. cheesmanii ‘long’ could be displaced by this recently introduced L. esculentum ‘Gal cer,’ because L. esculentum ‘Gal cer’ is a highly efficient invader of disturbed open spaces that produces large quantities of fruits and seeds dispersed by different animals, including the Galápagos mockingbird.
In addition, all the members of Eulycopersicon can intercross and the offspring is fertile, suggesting that they have not completed the speciation process that results in crossing barriers. Although we did not find direct evidence in our data, it is still possible that L. cheesmanii ‘long’ populations have already suffered a loss of genetic integrity through hybridization with L. esculentum ‘Gal cer,’ because both taxa can be found intermingled or physically very close. Although L. cheesmanii has inserted or weakly exserted stigmas, which leads to autogamy (Rick, 1963
), styles of L. esculentum ‘Gal cer’ are longer and the stigmas are exserted, thus some cross-pollination is likely (Rick et al., 1978
). In addition, the native Galápagos bee, Xylocopa darwinii, has been observed by Rick (1963)
to visit L. cheesmanii flowers and by us to visit the L. esculentum ‘Gal cer’ flowers. This suggests that occasional hybridization between adjacent L. cheesmanii ‘long’ and L. esculentum ‘Gal cer’ is possible, although it would probably favor pollen transfer from L. cheesmanii to L. esculentum.
Conservation considerations
The lack of correlation between genetic and geographic distances in L. cheesmanii probably is a consequence of the strict autogamy throughout the species, leading to rapid, and often fortuitous fixation of alleles (genetic drift) and differentiation between populations, even if they are physically very close (Rick, 1963
, 1983
). In fact, the genetic distances among the four L. cheesmanii f. minor populations from the small island of Bartolomé are similar to the genetic distances among other L. cheesmanii f. minor populations from other islands. This means that conservation of substantial diversity might be achieved if a few protected areas were established where several populations are found. It also has important implications for conservation ex situ, because high diversity could be collected by visiting just a few places where many scattered populations exist. Furthermore, within L. cheesmanii, the greatest DNA diversity was found in L. cheesmanii ‘short’ and the lowest in L. cheesmanii f. minor. Rick and Fobes (1975)
also found that L. cheesmanii f. minor was the least variable form. This indicates that most genetic variation could be conserved by preserving the more variable forms of L. cheesmanii.
Because an important part of the great DNA diversity can be found in a single area, establishment of a few protected areas could enable the conservation of most of the natural diversity of native Galápagos tomatoes. These areas should include protection of the native populations from introduced herbivorous animals and removal of all L. esculentum ‘Gal cer’ plants from the surroundings. Collecting Galápagos tomatoes in order to conserve them ex situ can also provide a means to prevent the irreversible loss of germplasm. In this way, accessions that were collected from the 1950s to the 1970s by C. M. Rick and others could be useful to reestablish some extinct natural populations.
|
|
FOOTNOTES |
|---|
2 This paper is dedicated to the memory of Dr. Charles M. Rick, who made the greatest contributions to the biosystematics of Lycopersicon.
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Bretó M. P. M. J. Asins E. A. Carbonell 1993 Genetic variability in Lycopersicon species and their genetic relationships. Theoretical and Applied Genetics 86: 113-120[ISI]
Brummitt R. K. 2002 How to chop a tree. Taxon 51: 31-41[CrossRef][ISI]
Cassol T. D. A. St. Clair 1994 Inheritance of resistance to blackmold (Alternaria alternata (Fr.) Keissler) in two interspecific hybrids of tomato (Lycopersicon esculentum x L. cheesmanii f. typicum). Theoretical and Applied Genetics 88: 581-588[CrossRef][ISI]
Esquinas-Alcázar J. F. Nuez 1995 Situación taxonómica, domesticación y difusión del tomate. In F. Nuez [ed.], El cultivo del tomate, 13–42. Mundi-Prensa, Madrid, Spain
Hassan A. A. H. M. Mazayd S. E. Moustafa S. H. Nassar M. K. Nakhla W. L. Sims 1984 Inheritance of resistance to tomato yellow leaf curl virus derived from Lycopersicon cheesmanii and Lycopersicon hirsutum. HortScience 19: 574-575
Hewitt J. D. T. C. Garvey 1987 Wild sources of high soluble solids. In D. J. Nevins and R. A. Jones [eds.], Plant biology, vol. 4, Tomato biotechnology, 45–54. Liss, New York, New York, USA
Highton R. 1993 The relationship between the number the number of loci and the statistical support for the topology of UPGMA trees obtained from genetic distance data. Molecular Phylogenetics and Evolution 2: 337-343[CrossRef][Medline]
IPGRI [International Plant Genetic Resources Institute]. 1996 Descriptors for tomato (Lycopersicon spp). IPGRI, Rome, Italy
Jackson M. H. 1994 Galápagos: a natural history. University of Calgary Press, Calgary, Alberta, Canada
Jenkins J. A. 1948 The origin of the cultivated tomato. Economic Botany 2: 379-392
Jones C. J. et al 1997 Reproducibility testing of RAPD, AFLP and SSR markers in plants by a network of European laboratories. Molecular Breeding 3: 381-390[CrossRef][ISI]
Kardolus J. P. H. J. van Eck R. G. van den Berg 1998 The potential of AFLPs in biosystematics: a first application in Solanum taxonomy (Solanaceae). Plant Systematics and Evolution 210: 87-103[CrossRef][ISI]
Lester R. N. 1991 Evolutionary relationships of tomato, potato, pepino and wild species of Lycopersicon and Solanum. In J. G. Hawkes, R. N. Lester, M. Nee, and N. Estrada [eds.], Solanaceae III: taxonomy, chemistry, evolution, 283–301. Royal Botanic Gardens, Kew, London, UK
Mackinney G. C. M. Rick J. A. Jenkins 1954 Carotenoid differences in Lycopersicon: hybrids of an unusual race of L. pimpinellifolium. Proceedings of the National Academy of Sciences, USA 40: 695-699
Mantel N. A. 1967 The detection of disease clustering and a generalised regression approach. Cancer Research 27: 209-220[ISI][Medline]
Marshall J. A. S. Knapp M. R. Davey J. B. Power E. C. Cocking M. D. Bennett A. V. Cox 2001 Molecular systematics of Solanum section Lycopersicum (Lycopersicon) using the nuclear ITS rDNA region. Theoretical and Applied Genetics 103: 1216-1222[CrossRef][ISI]
Milbourne D. R. Meyer J. E. Bradshaw E. Baird N. Bonar J. Provan W. Powell R. Waugh 1997 Comparison of PCR-based marker systems for the analysis of genetic relationships in cultivated potato. Molecular Breeding 3: 127-136
Miller J. C. S. D. Tanksley 1990 RFLP analysis of phylogenetic relationships and variation in the genus Lycopersicon. Theoretical and Applied Genetics 80: 437-448[ISI]
Muller C. M. 1940 A revision of the genus Lycopersicon. United States Department of Agriculture Miscellaneous Publication No. 382. USDA, Washington, D.C., USA
Nadzhimov U. K. S. C. Jupe M. C. Jones I. M. Scott 1988 Growth and gibberellin relations of the extreme dwarf dx tomato mutant. Physiologia Plantarum 73: 252-256[CrossRef]
Nei M. 1973 Analysis of gene diversity in subdivided populations. Proceedings of the National Academy, USA 70: 3321-3323[CrossRef]
Nei M. 1974 A new measure of genetic distance. In J. F. Crow and C. Denniston [eds.], Genetic distance, 63–76. Plenum Press, New York, New York, USA
Nesbitt T. C. S. D. Tanksley 2002 Comparative sequencing in the genus Lycopersicon for the evolution of fruit size in the domestication of cultivated tomatoes. Genetics 162: 365-379
Nuez F. 1995 Desarrollo de nuevos cultivares. In F. Nuez [ed.], El cultivo del tomate, 625–669. Mundi-Prensa, Madrid, Spain
Palmer J. D. D. Zamir 1982 Chloroplast DNA evolution and phylogenetic relationships in Lycopersicon. Proceedings of the National Academy of Sciences, USA 79: 5006-5010
Peralta I. E. D. M. Spooner 2001 Granule-bound starch synthase (GBSSI) gene phylogeny of wild tomatoes (Solanum L. section Lycopersicon (Mill.) Wettst. Subsection Lycopersicon). American Journal of Botany 88: 888-1902
Picó B. M. J. Díez F. Nuez 1996 Viral diseases causing the greatest economic losses to the tomato crop. II. The tomato yellow leaf curl virus (TYLCV)—a review. Scientia Horticulturae 67: 151-196[CrossRef]
Poysa V. 1993 Use of Lycopersicon cheesmanii and L. chmielewskii to increase dry matter content of tomato fruit. Canadian Journal of Plant Science 73: 273-279[ISI]
Rick C. M. 1956 Genetic and systematic studies on accessions of Lycopersicon from the Galápagos islands. American Journal of Botany 43: 687-696[CrossRef][ISI]
Rick C. M. 1963 Biosystematic studies on Galápagos tomatoes. Occasional Papers of the California Academy of Sciences 44: 59-77
Rick C. M. 1967 Exploiting species hybrids for vegetable improvement. Proceedings of the XVII International Horticultural Congress 3: 217-219
Rick C. M. 1971 Lycopersicon Mill. In I. L. Wiggins and D. M. Porter [eds.], Flora of the Galápagos Islands, 468–471. University of Stanford Press, Stanford, California, USA
Rick C. M. 1979 Biosystematic studies in Lycopersicon and closely related species of Solanum. In J. G. Hawkes, R. N. Lester, and A. D. Skelding [eds.], The biology and taxonomy of the Solanaceae, 667–677. Linnean Society of London, London, UK
Rick C. M. 1980 Petroselinum (Pts), a new marker for chromosome 6. Tomato Genetics Cooperative Report 30: 32.
Rick C. M. 1983 Genetic variation and evolution of Galápagos tomatoes. In R. I. Bowman, M. Berson, and A. E. Leviton [eds.], Patterns of evolution in Galápagos organisms, 97–106. American Association for the Advancement of Science, San Francisco, California, USA
Rick C. M. R. I. Bowman 1961 Galápagos tomatoes and tortoises. Evolution 15: 407-417[CrossRef][ISI]
