Moscone, E.A., Scaldaferro, M.A., Grabiele, M., Cecchini, N.M., Sánchez García, Y., Jarret, R., Daviña, J.R., Ducasse, D.A., Barboza, G.E. and Ehrendorfer, F. 2007.

THE EVOLUTION OF CHILI PEPPERS (CAPSICUM - SOLANACEAE): A CYTOGENETIC PERSPECTIVE. Acta Hort. (ISHS) 745:137-170

VIth International Solanaceae Conference
Eds.: D.M. Spooner etal. 137Acta Hort. 745. ISHS 2007

Introduction

Capsicum L.[tribe Solaneae, subtribe Capsicinae. according to Hunziker (2001)] is a small genus comprising ca. 30 species, some with varieties of great economic importance native to the New World. Five species (C. annuum var. annuum, C. chinense, C.frutescense, C.baccatum, C.baccatum varieties pendulum and umbilicatum, and C.pubescense) were domesticated by American natives. After Columbus, they became widely exploited in tropical to temperate regions because of their fruits, which have high nutritional contents, especially in vitamins. They are constituent of the human diet, the pungent cultivars as spice ("ajies", "paprika", "chillies", "hot peppers") and the sweet types as vegetables ("sweet pepper", "bell pepper", "pimiento"). Moreover, the genus has medical and ornamental applications (IBPGR, 1983; Pickersgill, 1991; Esbaugh, 1993; Bosland and Votava, 2000; Hunziker, 2001).

Several authors have focused on the taxanomy of this genus during the past 50 years (Hunziker, 1950, 2001; Heiser and Smith, 1953; Eshbaugh, 1979, 1980; Pickersgill et al., 1979; Pickergill, 1988; Barboza and Bianchetti, 2005), but a complete treatment is still lacking. A list of 31 recognized species according to Barboza (unpubl. data) and some relevant taxonomic traits are presented in the table. The predominant growth forms in the genus are perennial shrubs, although several species also display biennial herbaceous growth, particularly those of the Capsicum annuum complex (C.annuum, C.frutescense and C.chinense). A few taxa can develop into trees, e.g., C. parvifolium and C. rhomboideum. Capsicum caballeroi and C.eximium are unique in having herbaceous, shrubby and arborescent forms. Corollas are stellate (most frequently), rotate or campanulate. 

Corolla color in particular has been utilized to characterize the cultivated species and their wild relatives, and to divide them into "white-" and a "purple-flowered group." However, this concept becomes inadequate when describing the flower color variation found in more distantly related wild species and the genus as a whole. The reason is that some species exhibit single-colored flowers (e.g., C.chacoense, C.fruburgense, C.rhomboideum), i.e., white, cream, yellow, ochre, pink, lilac and purple-violet, while others (e.g., C.coccineum, C.hunzikerianum, C.parvifolium) have different color combinations in lobules, throat and tube, often including spots of various colors, which complicates species delimination (Hunziker, 2001; Barboza and Bianchetti, 2005). In the wild peppers, fruits are usually spherical (in some cases also ovoid or elliptic), their color being red or yellowish-green, rarely orange. In the cultivated taxa, fruits traits are highly variable within species due to human selection (Pickergill, 1988). Seeds are reniform (except in C. pubescense where they are irregular) and yellowish, brownish or blackish. Self-incompatibility is the rule in the genus. From the 11 species examined for this trait, only C. cardenasii was found to be self-incompatible (Eshbaugh, 1979; Onus and Pickersgill, 2004).

List of recognized Capsicum species

Fruit pungency is characteristic of the genus due to substances unique to peppers in a mixture known as "capsaicinoids." The presence of pungency is controlled by a single dominant major gene (Pun1 or C) demonstrated in the Capsicum annuum complex, although its expresiion varies among species and cultivars depending on other modifier genes epistatically affected by Pun1 and environmental conditions (Lippert et al., 1966; IBPGR, 1983; Bosland and Votava, 2000; Hunziker, 2001; Lefebvre et al., 2002; Stewart et al., 2005). At least two wild species, C.lanceolatum and C.rhomboideum (Bosland and Zewdie, 2001) are reported to be completely free of pungency, and pungency is also absent in some accessions of C. chacoense (Eshbaugh, 1980), in one C. cornutum (as C.dusenii Bitter; Hunziker, 1971), and in cultivars of C. annuum var. annuum after human selection.

Distribution

According to Hunziker (2001), four centre’s of distribution can be recognized for Capsicum : 1) southern USA and Mexico to western South America (Peru;12 spp.), 2) northeastern Brazil and coastal Venezuela (1 sp.), 3) eastern coastal Brazil(10 spp.), and 4) central Bolivia and Paraguay to northern and central Argentina (8 spp.).The greatest number of species (16) is concentrated in Brazil (Barboza and Bianchetti,2005). It is noteworthy that 16 species are endemics of different regions of South America(see

1). The original place of domestication for each cultivated species is still underdiscussion (see below). Peppers are among the oldest cultivated plants in the Americas,and archeological remains indicate that C. annuum in particular was used by man evenbefore the advent of agriculture (Pickersgill, 1969). The domesticated species includewild populations, except C. pubescens, which is known only in cultivation. In the case of C. annuum and C. baccatum, the wild forms are distinguished as taxonomic varieties, i.e.,glabriusculum and baccatum, respectively. The spontaneous forms of the C. annuum complex intergrade in a way that makes it difficult to separate one from the other, and to distinguish between weedy forms escaped from cultivation and true wild forms(Pickersgill et al., 1979; Pickersgill, 1988; Eshbaugh, 1993).

Karyotype

Research from different fields has contributed to the knowledge of the genetic diversity in the genus and relationships among species, e.g., morphological analyses(Pickersgill et al., 1979), crossing experiments (Tong and Bosland, 1999; Onus andPickersgill, 2004), classical (cf. Pickersgill, 1977, 1991; Moscone et al., 1993a, 1996a; Pozzobon et al., 2006) and molecular cytogenetic analyses (Park et al., 1999, 2000;Scaldaferro et al., 2006), systematic biochemical studies (Ballard et al., 1970; Boslandand Zewdie, 2001), and protein electrophoretic studies (McLeod et al., 1983; Loaiza-Figueroa et al., 1989). In addition, DNA sequence analyses (Walsh and Hoot, 2001) and studies on restriction fragment length polymorphisms (RFLP), polymerase chain reaction(PCR)- amplified fragment length polymorphisms (AFLP), randomly amplifiedpolymorphic DNA (RAPD), and polymorphic plastid DNA (cpDNA) markers have been published (cf. Prince et al., 1995; Paran et al., 1998; Rodriguez et al., 1999; Buso et al.,2002; Votava et al., 2005).

From a practical point of view, important achievements have been the identification of many genes contributing to crop quality in peppers, such as those responsible for resistance to different diseases, and the production of capsaicinoids andbeta-carotene (cf. Lippert et al., 1966; Daskalov and Poulos, 1994). It has been possible to obtain genetically improved cultivars by induced mutagenesis (cf. Daskalov, 1986;Daskalov and Baralieva, 1992), and to construct a relatively saturated linkage map based on molecular markers (Livingstone et al., 1999; Ben Chaim et al., 2001; Lefebvre et al.,2002). Furthermore, one should mention the efforts to construct bacterial (BAC) and yeast artificial chromosome (YAC) libraries to better handle the pepper genome (Kim etal., 1998; Tai and Staskawicz, 2000) and to create complementary DNA (cDNA) libraries as a strategy to generate expressed sequence tags (EST) for the discovery of new genes(Lee et al., 2002).

For crop improvement and the success of breeding programmes in particular, it is essential to obtain more basic information on the genetic diversity of the available germplasm, and on the genomic affinities between the possible donors of valuable allelesand the crops to be improved. In this respect, the cytogenetic characterization of the germplasm is important for the conservation and utilization of plant genetic resources, aswell as for hybridization and biotechnological approaches, including transformations.Therefore, we are facing a broad programme of genome characterization in the pepper germplasm by classical and molecular cytogenetics to evaluate its inter- and intraspecific variability. Such a programme will ultimately enhance our knowledge of the genomeorganization and evolution in Capsicum, with reference to the origin of the crop species and their improvement.

What is Karyotype?

A karyotype is the number and appearance of chromosomes in the nucleus of a eukaryote cell.The term is also used for the complete set of chromosomes in a species, or an individual organism. Karyotypes describe the number of chromosomes, and what they look like under a light microscope. Attention is paid to their length, the position of the centromeres, any differences between the sex chromosomes, and any other physical characteristics. The preparation and study of karyotypes is part of cytogenetics.

The study of whole sets of chromsomes is sometimes known as karyology. The chromosomes are depicted (by rearranging a microphotograph) in a standard format known as a karyogram or idiogram: in pairs, ordered by size and position of centromere for chromosomes of the same size.

Karyotypes can be used for many purposes; such as, to study chromosomal aberrations, cellular function, taxonomic relationships, and to gather information about past evolutionary events.

The karyotype analyses with a variety of technical approaches used for the present contribution demonstrate that cytogenetics provides powerful tools for species characterization and taxonomic grouping in Capsicum. This becomes particularly evident when cytogenetic data are combined with methodological approaches from other disciplines such as morphological and phytogeographic studies (Eshbaugh, 1979;Pickersgill et al., 1979), crossing experiments (Pickersgill, 1988, 1991; Onus and Pickersgill, 2004), isozyme analyses (McLeod et al., 1983; Loaiza-Figueroa et al., 1989), chloroplast and nuclear DNA sequencing studies (Park et al., 2000; Walsh and Hoot,2001), and plastid DNA fragment polymorphism analysis (Buso et al., 2002).
Chromosome numbers reported for 23 of the 31 recognized species allow us to distinguish two species groups: one with 2n=2x=24 (13 species) and another with2n=2x=26 (10 species; see Table and references therein). Among the 2n=24 species here examined, Capsicum chacoense, C. galapagoense, C. annuum, C. chinense and C. frutescens appear as closely related taxa according to their karyotype features.

These species are placed in the "white-flowered group" as they have uniformly white,stellate corollas (sometimes cream in the latter three species; Pickersgill, 1991). Inaddition, they exhibit red fruits (various colors in the cultivated forms of the domesticated species) of different shapes, and yellowish seeds. The last three species are included inthe C. annuum complex, where poorly developed crossing barriers and great similarities with respect to morphology, isozymes, and DNA sequences suggest that they could be conspecific (cf. Pickersgill, 1988, 1991; Park et al., 2000; Walsh and Hoot,. 2001).

Group with 24 chromosome

Another 2n=24 species group is formed by Capsicum eximium, C. cardenasii, C. pubescens, and C. tovarii, which display rather large chromosome complements and high DNA contents, except the latter species (cf. Belletti et al., 1998). In addition, theyhave two active NORs in the haploid complement (one of them placed in a long arm in C. eximium and C. cardenasii), a sm instead o fan St chromosome (except C. pubescens),high heterochromatin amounts (more than 12%) of the GC-rich type (C. pubescens alsowith a single AT-enriched band), and complex banding patterns (except C. eximium).These species form the "purple-flowered group," in which C. pubescens is the coremember (Pickersgill, 1991). Their flowers are partly (or totally) purple-violet; C. tovarii also has white-cream forms (sometimes with greenish spots). The corolla shape is stellatein C. eximium and C. tovarii, campanulate in C. cardenasii, and rotate in C. pubescens. Fruits are spherical and red, and seeds reniform and brownish, except in C. pubescens,where fruit shape and color are variable, and seeds are irregular and blackish. The shrubby growth form is typical, although C. eximium also has herbaceous and arborescent variants. The close relationships between C. eximium, C. cardenasii and C. pubescens (particularly the first two taxa) is supported by crossing experiments, which produce fertile hybrids in any of the combinations, although none of these species can be successfully crossed with C. tovariii (Pickersgill, 1991). Further support for this grouping comes from molecular data, which show C pubescens in particular to be distant from the white-flowered taxa (cf. McLeod et al., 1983; Park ci al., 2000; Walsh and Hoot, 2001).In any case, links between the "white-" and the "purple-flowered group" are evident fromthe fact that C. eximium and C. cardenasii give viable progeny in crosses with C. frutescens and C. baccatum (cf..McLeod et al., 1983; Pickersgill. 1991).

The position of Capsicum tovarii is still not clarified. Crossing experiments demonstrate unilateral incompatibility with the other members of the "purple-flowered group" and successful hybrids only with C. baccatum and C. praetermissum (cf. Tongand Bosland, 1999; Onus and Pickersgill, 2004). On the other hand, isozyme studies donot show this species closely related either to the "white-" or the "purple-flowered group"(McLeod ci al., 1983). In addition, phylogenetic trees from DNA sequences areconflicting, as Walsh and Hoot (2001) suggested a relatively close relationship with C. pubescens and both species separated from C. exiinium and C. cardenasii, whereas Choong (1998) found C. tovarii not related to the remaining "purple-flowered species."The present results indicate that C. tovarii is a member of the "purple-flowered group,"supported by karyotype features common to the whole assemblage.

The remaining two species with 2n=24, Capsicum parvifolium and C. flexuosum, appear in isolated positions (somewhat uncertain for the first) according to their karyotype features. Their genomes are among the largest of the species studied. Their heteroehromatin is GC-rich, although the latter taxon has much more heterochromatin anda more complex banding pattern than the former. Moreover, C. jiexuosuin differs inhaving predominantly intercalary bands and two active NORs in comparison with C. parvifolium, which has mostly terminal bands and only one active NOR in the haploidcomplement. The latter species is unique in the genus in having only m chromosomes incytotype 2. Both taxa possess spherical red fruits, but they are separated by othermorphological traits, with C. Jlexudsüin having the same corolla shape and color asC. baccatuni in addition to blackish seeds, whereas C. parvifolium exhibits a white rotatecorolla with purple spots and greenish inner parts, together with brownish seeds. Both species are shrubs, but in addition, C. parvifolium also has arhorescent forms. Although cpDNA fragment polymorphism data place C. flexuosum into the "white-flowered group"(Buso ci al., 2002), karyotype analyses suggest that it could be the 2n=24 species closest to the taxa with 2n=26. The lack of crossing experiments and isozyme and DNA sequencedata for both species prevent further speculations.

Group with 26 chromosome

In the group of taxa with 2n=26, relationships between species based onkaryotype comparisons should be considered with caution. The striking chromosome variation found at the intraspecific level (where more than one accession was examined)makes the whole group very heterogeneous. Nevertheless, Capsicum mirabile and C. schottianurn appear to be closely related in sharing only one active NOR in the basicchromosome set, large karyotypes, high amounts of GC-enriched heterochromatin, andvery complex banding patterns, mostly with terminal bands. On the other. 'hand,C. pereirac and C. canipylopodiurn are distinct from each other and from the remaining 2n=26 species. The former possesses two active NORs in the haploid complement andless heterochromatin, which is GC-rich and partly of mottled appearance (AT-rich only insome bands of one cytotype), whereas the latter has large amounts of AT-richheterochromatin and very complex (predominantly intercalary) banding patterns. Finally, C. recurvatum and C. rhomboideum stand apart by having low heterochromatin amounts(5%) and simple banding patterns, with the latter species deviating because of its verysmall chromosomes. Although all 2n=26 taxa are shrubs, the former species can also beherbaceous like C. mirabile, whereas the latter is the only member of the group with arborescent growth forms. The isolated position of C. rhomboideum among - the2n=26 species here examined is also evident by its rotate and entirely yellow corolla,- the remaining taxa having stellate and white corollas (sometimes with a green-yellowish tube) and with spots of various colors in the lobules or throat (purple in C. mirabile and C. pereirae, violet or brownish in C. schottianum, golden in C. campylopodium, and greenish in C. recurvatum).

According to their fruit shape and color and seed color, the 2n=26 species can be separated in two subgroups. One of them, with red spherical fruits and brownish seeds,includes Capsicum lanceolaturn and C. rhomboideum. The other subgroup, with green-yellowish spherical fruits (compressed in C. campylopodium and depressed in C. cornutum and C. friburgense) and blackish seeds, includes all the other species knownto have 2n=26. In any case, the links between the 2,7=26 species remain quite uncertain,because of lack of crossing experiments and isozyme and DNA sequence data. The latter are available only for C. rhomboideum and demonstrate that it is far apart from the species with 2n=24 (Walsh and Hoot, 2001). All peppers with 2n=26 examined by epDNA fragment polymorphism analyses [C. campylopodium, C. cornutum (as C. dusenii), C. mirabile (as C. buforum) and C. villosum] are grouped together, although the first species stands more apart and appears linked to C. chacoense with 2n=24 (cf.Buso et al., 2002).

Karyotype evolution

Polyploidy is not a relevant phenomenon in the evolution of peppers and has so far been found in only one accession of Capsicum annuum var. glabriusculum (as C. annuum wild), with 2n=4x=48 (Pickersgill, 1977). In contrast, the presence on the diploid level of intraspecific cytotypes in half of the species here studied indicates very active processes of chromosome diversification in Capsicum. This is most conspicuous in the wild species,particularly those with 2n=26. As our chromosome analyses were carried out on primary root tips of germinating seeds, correlations with phenotypic differentiation between intraspecific cytotypes were limited to voucher comparisons of the collected samples,which demonstrate no major distinctive macro-morphological traits corresponding to the cytotypes. Molecular analyses and crossing experiments between cytotypes should help inevaluating the importance of karyotype differentiation for species delimitation and the appearance of new taxa.

In the following paragraphs a synthetic model of the possible chromosome evolution in Capsicum is depicted (see Fig). This is based on a working hypothesisfrom previous contributions (Moscone et al., 1993a, 1995, 1996a, 2003) and modifiedaccording to the new karyotype information presented here.

DNA sequence data recently obtained in our laboratory (Sehr et al., in prep.) confirm the monophyly of Capsicum. Thus, the 2n=24 and 2n=26 species are taxonomically correctly placed in Capsicum and should have originated from a common ancestor

The diagram of possible karyotype relationships. in the evolution of Capsicum(Fig. 3) indicates that the C. annuum complex has been differentiated comparativelyearly, just after C. chacoense, by a moderate increase in heterochromatin amount and banding pattern complexity (less evident in some cytotypes of C. annuum). Additionalderived features in the complex are the loss of one active NOR in C. chinense andC. annuuni (two cytotypes), the change of position of one NOR from a short to a longchromosome arm and of another NOR from an st to an m chromosome in C. annuum var.glabriusculum (one cytotype), the change of a NOR-bearing m to an Sm chromosome, andthe increase in karyotype asymmetry in C. annuum (most cytotypes). Thus, the corespecies of the "white-flowered group" appear as relatively primitive. A reduction of genome size and a loss of one active NOR evidently has accompanied the differentiation of C. galapagoense and C. rhomboideum.

Among the more advanced and recent pepper lines, the group with Capsicum baccatum and C. praetermissum is characterized by derived features like a moderate increase in genome size, heterochromatin amount and banding pattern complexity. Incomparison with the more basal groups discussed above, C. baccatum has acquired one o rtwo additional active NORs, whereas C. praetermissum has developed a banding pattern of mostly mottled appearance, indicative of a different satellite DNA composition(moderately GC-rich). The main derived traits of the three most advanced branches in ourdiagram are considerable increases in genome size, heterochromatin amounts and bandingpattern complexities. The first of these branches gave rise to the "purple-flowered group,"composed of C. eximium, C. cardenasii, C. pubescens and C. tovarii, with a change ofone St to an sin chromosome. In the former two species of this group, one active NOR haschanged location from a short to a long chromosome arm.

Exclusive changes in C. eximium are a decrease in heterochromatin amount and banding pattern complexity,whereas one chromosome has changed from sm to St in C. pubescen.s'. Furthermore, thelatter species acquired an AT-enriched satellite DNA in addition to the GC-rich one. Thesecond line has produced C. flexuosum, where a change of place of one active NOR froman st to an m chromosome and the appearance of conspicuous intercalary heterochromaticbands are advanced karyotype features. Finally, a third branch of great evolutionaryrelevance has led to the origin of the most highly evolved group with the majority of the2n=26 species. This is linked to a Robertsonian fission and has resulted in increasedkaryotype asymmetry and the loss of an active NOR. The taxa of this group, with C. mirabile as the core species, are in an active phase of striking intraspecific karyotype repatterning and speciation. Main derived traits are the change of the active NOR from an in an sm chromosome in C. schottianum, C. campylopodium and C. recurvatum and adecrease in heterochromatin amount and banding pattern in C. recurvatum. Furthermore,significant heterochromatin differentiation through the conspicuous appearance of moderately GC-rich (mottled) and some AT-rich heterochromatic bands occurred in C. pereira (the latter bands in cytotype I), whereas C. campylopodium has developed conspicuous intercalary banding, mostly composed of AT-rich satellite DNA and somemixed GC- and AT-rich bands.

The occurrence of two independent centric fission events in the history of peppers,as postulated here, is supported by the morphological features and geographical distribution of the two 2n=26 species subgroups. One of them is formed by Capsicum lanceolatum [the available karyotype information only indicates the chromosome number(Tong and Bosland, 1997)] and C. rhomboideum, with red nonpungent fruits and brownish seeds, growing in Mexico, Central America, and northwestern parts of SouthAmerica. The second subgroup includes eight species (C. campylopodium, C. cornutum, C . friburgense, C. mirabile, C. pereira, C. recurvatum, C. schottianun and C. villosum) with yellowish-green pungent fruits and blackish seeds, growing in eastern coastal Brazil. Considering the karyological features, the geographical distribution and the monophyly of the extant Capsicum species, one can speculate that a pepper ancestor firstappeared in a semiarid region of south-central Bolivia. In this "nuclear area,' delimited by the cities of Cochabamba, Sucre and Santa Cruz, the ancestor of C. chacoense could havefirst differentiated. Subsequent migrations accompanied by radiation and speciation might then have led to the present distribution of Capsicum, according to the hypothesis by McLeod et at. (1982) with some modifications. A strong argumentfor the basal placement of C. chacoense is its pivotal position in the genus, as shown bythe karyological data already discussed and, particularly, by the fact that its cytotypes (characterized by the localization of active NORs) are also found in other species, i.e.,cytotype I in C. frutescens and C. baccatum (With one/two additional NORs), andcytotype 2 in C. annuum. In addition, C. chacoense appears nested either with white-flowered species, i.e., C. baccatum (and C. praetermissum) according to isozyme and DNA sequence data (McLeod et al., 1983; Walsh and Hoot, 2001) or with C. campylopodium, one of the most advanced species in the genus with 2n=26, according to cpDNA fragment polymorphism analyses (Buso et al., 2002).

We assume that one primary migration to the lowland selvatic regions of northern Bolivia and western Brazil led to the origin of Capsicum annuum, a species growing wild in these areas (Barboza, unpubl. data). From there, it could have reached northern parts of South America and, mainly, Central America and Mexico, where nowadays it has its greatest diversity. Further movements to Amazonia gave rise to the other members of the C. annuum complex, i.e., C. chinense and C. frutescens. This speculative pathway is supported by the presence in the "nuclear area" of a river system flowing into the Amazon basin, which may have favored the primary dispersion as proposed by McLeod et al.(1982). From the C. frutescens-like ancestors C. parvifolium could have arisen in dry areas of northeastern Brazil, although its direct origin from the ancestral pool in the"nuclear area" also appears possible. Furthermore, the C. annuum line most probabl ygave rise to the endemic C. galapagoense, as well as to C. rhomboideum in northwesternSouth America. After the latter species had colonized Central America, C. lanceolatum could have appeared. A second early migration to eastern lowland subtropical regions ofBolivia may have led to the origin of C. baccatum and later to the "purple-flowered group" (C. eximium, C. cardenasii, C. pubescens and C. tovarii), differentiated in the dry western Andean regions at middle elevations. In addition, another branch of a C. baccatum-like ancestor may have extended to the eastern and southern subtropicalareas to produce C. praetermissum and C. flexuosum, respectively. Finally, the ancestor of the latter species could have migrated further east to the coastal rain forests of Brazil toestablish here the most active centre of diversification in Capsicum, with the 2n=26 group of C. campylopodium, C. cornutum, C. friburgense, C. mirabile, C. pereirae, C. recurvatum, C. schottianum and C. villosum.

Origin of the cultivated taxa

Results from morphology, crossing experiments, enzyme and DNA analyses(Pickersgill et al., 1979; McLeod et al., 1983; Pickersgill, 1988, 1991; Loaiza-Figueroa etal., 1989; Park et al., 2000; Walsh and Hoot, 2001) support an evolutionary scenario fo rthe domesticated peppers (Eshbaugh et al., 1983a, Fig. 2; Eshbaugh, 1993), whichproposes three independent ancestral lines leading to the cultivated Capsicum taxa. The first includes wild (sub)shrubby members with herbaceous forms of the white-flowered           C. annuum complex where species definition by different methodological approaches is complicated (see above). Although the present observations also indicate a close relationship between the three species of this complex, they provide enough karyological information to clearly distinguish each species as an independent taxon. Capsicum chinense and C. frutescens, centred in north Amazonia, are apparently closely related,whereas C. annuum, centred in Mexico, stands somewhat more apart. Our data do notexclude the possibility that cultivars in these two areas could have originated independently.

The cultivated Capsicum baccatum varieties pendulum and umbilicatum evidently have originated from the wild C. baccatum var. baccatum centred in Bolivia (Eshbaugh,1993). These three taxa differ from the former group by having considerably higheramounts of heterochromatin and are karyologically very homogenous. A third line to be considered consists of shrubby purple-flowered taxa. Their greatest diversity occurs in thecentral Andes and encompasses C. eximium, the Bolivian endemic C. cardenasii, the Peruvian endemic C. tovarii and the cultigen C. pubescens. With the exception of C. eximium, high heterochromatin amounts and complex banding patterns are distinctive traits of this group. Our data suggest that C. cardenasii is most likely involved in theorigin of C. pubescens, although the participation of C. tovarii cannot be excluded. Inconclusion, the triple origin of the cultivated Capsicum taxa is clearly confirmed by our present data and previous papers (Moscone etal., 1993a, 1995, 1996a, 2003).

Where are the original places of domestication for the cultivated peppers? In thecase of Capsicum annuum, it has been postulated that its cultivation first occurred in Mexico, which is the major centre of diversity for this species and where archeological remains from about 7000 B.C. are found (Pickersgill, 1969). Support for this hypothesiscame from cytogenetic data of Pickersgill (1971), which demonstrate that the typicalkaryotype of the cultivated C. annuum var. annuum, with two "acrocentric" chromosomes(one sm and one st) is found only in accessions of the wild C. annuum var. glabriusculum from southern Mexico and Guatemala (cf. Eshbaugh et al., 1983a). Patterns of geneticvariation obtained by enzyme electrophoretic studies suggest a primary centre of domestication in eastern Mexico and a second one in a western part of the country (Loaiza-Figueroa et al., 1989). Our results show that wild C. annuum var.glabriusculum from Florida (USA) and var. annuum cultivars from Argentina and Austria are quite similar and share cytotype 1. Although these data are still fragmentary, they open the possibility of tracking the origin of domestication by extending karyotype analyses to additional samples.

With respect to Capsicum chinense and C. frutescens, which have a similar distribution in South America, domestication supposedly began in the lowland Amazonbasin, their area of greatest diversity (IBPGR, 1983). The oldest archeological remains of C. frutescens have been found in Peru and date from 1200 B.C. (Pickersgill, 1969). Ourcytological data in both species are still not conclusive.
Archeological remains from Peru date the cultivation of Capsicum baccatum by man back to 2500 B.C. (Pickersgill, 1969). However, the cultivated varieties of thisspecies could have been domesticated in lowland subtropical regions of southern Bolivia,as the wild variety baccatum has mainly an Andean distribution with the main centre of diversity in Bolivia (McLeod et al., 1982; Eshbaugh, 1993). Our findings in this speciesof the same cytotype (with two NORs in the haploid complement) in wild and cultivated accessions from Andean areas support the speculations made above.

Finally, for Capsicum pubescens, a mid-elevation region of Bolivia had been proposed as the original place of domestication, as its closely related species grow there and plants with fruits of smaller size occur there also, which approach the more primitive size (McLeod et al., 1982; Eshbaugh, 1993). The karyotype affinities we found in the"purple-flowered group" suggest that a region comprising southern Peru and northern Bolivia, which includes the distribution areas of the extant C. tovarii and C. cardenasii, could be the most probable place of domestication.

Read More: http://www.actahort.org/books/745/745_5.htm

GO UP