The material studied of A. porphyrocalyx corresponds to accession J.F.M. Valls, J.P. Moss and G.P. Silva 7303, collected in Brazil, Minas Gerais state, municipality of Uberaba, in the gardens of the Uberaba Country Club, on the edge of highway BR-050, next to Río Grande river, 20 km southeast of Uberaba, 19°58'S 47°47'W, in 1983. Germplasm from this original collection has been conserved at the Wild Arachis Genebank of Embrapa, in Brasília, Distrito Federal, and increased seed has been distributed to partner institutions. Seeds used in this study were obtained from the peanut germplasm collections of the Instituto de Botánica del Nordeste in Corrientes, Argentina. The voucher materials of the original accession are deposited in the herbaria CTES and CEN, and are paratypes of the species name. The holotype and isotypes of A. porphyrocalyx were collected nine years later from exactly the same site (J.F.M. Valls, C.E. Simpson, R.N. Pittman, D.E. Williams and G.P. Silva 13271).

Karyotype measures were obtained by the analysis of five individuals and four Feulgen-stained metaphase plates per individual and using the computer application MICROMEASURE version 3.3 (Reeves and Tear 2000). Karyotype description is based on the nomenclature by Levan et al. (1964). Chromosomes were classified in three categories according to the centromeric index (CI = short arm × 100/total length of chromosome): metacentric (m) when CI = 50–37.5 and submetacentric (sm) when CI = 37.5–25 and subtelocentric when CI = 25–12.5. Classification of SAT chromosomes was performed on the basis of the satellite relative size and position of the centromere (Fernández and Krapovickas 1994). The total chromosome length (TCL) was obtained by summing the average length of each chromosome in the four metaphase samples of each individual, and then the average of the five individuals was performed. The chromosome mean length was calculated by dividing the TCL by the number of chromosomes of the species. The karyotype asymmetry indices were estimated using the intrachromosomal (A₁) and interchromosomal (A₂) indexes by Romero Zarco (1986).

Data from homologous chromosomes were combined first to obtain mean values of different pairs of chromosomes in the same metaphases and, subsequently, of the same chromosome pair in different metaphases. Haploid complements were represented in the ideogram. Chromosomes were ordered first by morphology and then by decreasing size.

The chromosome number of A. porphyrocalyx, previously determined by Peñaloza and Valls (2005), was confirmed by the mitotic analysis, 2n = 2x = 18 (Fig. 2a). Likewise, in all analysed metaphases, a chromosome pair with characteristics of ‘A’ chromosomes was observed (Fig. 2b). This chromosome pair was easily identified by showing a lower condensation level of the euchromatin regions of their arms in comparison with the same regions in the remaining chromosomes of the complement and by corresponding to the smallest chromosomes of the karyotype. The ‘A’ chromosome pair corresponds to pair 7 (Fig. 3).

The karyotype consisted of seven pairs of metacentric chromosomes and two subtelocentric pairs (14m + 4st; Fig. 3). These data do not agree with those reported by Peñaloza and Valls (2005) who observed four pairs of submetacentric chromosomes, but this discrepancy can be due to the fact that the formula published by these authors is based only on a visual analysis. Chromosomal size ranged between 1.24 and 2.08 µm with a mean length of 1.63 µm, belonging to the category of small chromosomes according to Lima de Faría (1980), and the mean length of diploid complement was 29.37 µm (Table 1). The indexes of asymmetry revealed a moderately high degree of intrachromosomal asymmetry (A1 = 0.30) but low interchromosomal asymmetry (A2 = 0.16). Only one pair of SAT chromosomes was found in all analysed metaphases. These chromosomes were the longest metacentric chromosomes of the complement (pair 1) and, as previously reported by Peñaloza and Valls (2005), correspond to the SAT chromosomes type 8 described by Fernández and Krapovickas (1994).

The metaphases of some individuals showed two or three chromosome pairs with extended primary constrictions (centromeres) and the chromosome arms separated. These chromosomes had the centromere unusually large or stretched during prophase or prometaphase, and consequently the number of chromosomal elements increased up to 25. Similar behaviour has been observed in some chromosomes of other species of the genus Arachis, such as Arachis cardenasii (Krapov. & W.C. Gregory, 1994), Arachis helodes (Mart. ex Krapov. & Rigoni, 1958), Arachis valida (Krapov. & W.C. Gregory, 1994), Arachis duranensis (Krapov. & W.C. Gregory, 1994) and Arachis correntina ((Burkart) Krapov. & W.C. Gregory, 1994) all belonging to the section Arachis (Fernández and Krapovickas 1994), and of other genera such as Antirrhinum majus (Linneaus, 1753), Allium sphaerocephalum (Crome ex Schltdl, 1824) (Lima De Faria 1956), Libocedrus chilensis (Endlicher, 1847) (Hunziker 1961), Pisum (Linneaus, 1753) and Lathyrus (Linnaeus, 1753) (Neumann et al. 2015). Just recently, Neumann et al. (2015) have classified this type of centromere as ‘intermediate’ between the two types of centromeres, monocentric and holocentric, and having an organization characterized by multiple Cen-H3 domains.

DAPI staining after FISH revealed C-DAPI+ centromeric bands in all chromosomes of the karyotype (Fig. 2c). These bands had similar sizes along karyotype, except in the ‘A’ chromosomes, where they were most conspicuous (Fig. 3). The results of in situ hybridisation showed one pair of 45S rDNA loci in proximal position on the long arm of the longest metacentric chromosomes (pair 1) and one pair of 5S rDNA loci in proximal or interstitial positions on the second longest chromosome pair of the karyotype (pair 2; Figs 2c and 3).

Like most species of the genus Arachis, the karyotype of A. porphyrocalyx consists of small size chromosomes, mainly metacentric. The smallest chromosome pair showed all features that define the ‘A’ chromosomes: a chromosome length 54% smaller than the largest chromosomes of karyotype, and showing allocycly in somatic prophases and pro-metaphases (Fernández and Krapovickas 1994). Chromosomes with these features have not been reported until now for other species not assigned to the A genome of the section Arachis (Fernández and Krapovickas 1994, Robledo et al. 2009). Arachis porphyrocalyx has a moderately asymmetric karyotype due to the presence of two pairs of subtelocentric chromosomes. This structure contrasts with the more symmetric karyotypes of the other x = 9 species of Arachis that are composed by eight metacentric pairs and only one submetacentric (Lavia 1998, 1999). Besides, the presence of subtelocentric chromosomes distinguishes A. porphyrocalyx from the other species of the section Erectoides, which have karyotypes formed only by metacentric and submetacentric chromosomes (Fernández and Krapovickas 1994, Lavia 2001, Lavia et al. 2009, Ortiz et al. 2013). Indeed, the presence of subtelocentric chromosomes in the karyotypes is uncommon within the genus Arachis, a feature that A. porphyrocalyx only shares with Arachis batizocoi (Krapov. & W.C. Gregory, 1974) and Arachis glandulifera (Stalker, 1991) (Fernández and Krapovickas 1994), both species of the section Arachis.

Even though A. porphyrocalyx owns a unique pair of SAT chromosomes in metaphase, as the other species with x = 9, these chromosomes correspond to a different type according to the classification proposed by Fernández and Krapovickas (1994). Thus, A. porphyrocalyx has SAT chromosomes type 8, while the remaining x = 9 species show type 3 (Lavia 1998, Silvestri et al. 2015). Similarly, it differs from those observed in Arachis douradiana (Krapov. & W.C. Gregory, 1994), Arachis hermannii (Krapov. & W.C. Gregory, 1994), Arachis major (Krapov. & W.C. Gregory, 1994), Arachis paraguariensis (Chodat & Hassler, 1904) and Arachis stenophylla (Krapov. & W.C. Gregory, 1994) from the section Erectoides that have satellites of type 2, 3A or 4 (Fernández and Krapovickas 1994, Lavia 2001, Lavia et al. 2009, Ortiz et al. 2013).

Until now, four distribution patterns of centromeric C-DAPI+ heterochromatin have been identified in the karyotypes of Arachis species (Seijo et al. 2004, Robledo and Seijo 2008, 2010, Robledo et al. 2009, Silvestri et al. 2015, Ortiz, unpublished). One of them, proper to A. glandulifera of section Arachis (D genome), is characterised by a markedly heterogeneous distribution, with chromosomes showing large bands, and others with faint bands or devoid of them. The second pattern, with large bands of similar size in all or almost all chromosome pairs, is present in the x = 9 species and A and K genome species of the section Arachis. The third pattern, with small blocks at most chromosomes, which are revealed as faint bands or like-dot bands on the flanks of the centromeres, is shown in the F genome species of the section Arachis and in species of sections Erectoides and Procumbentes. And the fourth possibility, with no detectable bands in the entire karyotype, is proper to B genome species of the section Arachis. Arachis porphyrocalyx has a pattern that is different from those species of the section Erectoides and is similar to that observed in the x = 9 and in some x = 10 species of the section Arachis. It even resembles that observed in the A genome species, since the A chromosomes have pericentromeric bands with relative size greater than those in the rest of karyotype; which also in turn strengthens the identity of these chromosomes.

Regarding the number and location of ribosomal loci (45S rDNA and 5S rDNA), A. porphyrocalyx has the same number of sites as other x = 9 species, that is one pair of each loci (Silvestri et al. 2015). However, the 45S rDNA loci in A. porphyrocalyx are located on a metacentric pair as in A. praecox, while in A. palustris and A. decora, they are located on a submetacentric pair. Meanwhile, 5S rDNA loci in A. porphyrocalyx are on the long arms of a large-size metacentric pair and in the remaining x = 9 species on the short arms of a small-size metacentric pair. Until now, the number of rDNA loci has been characterized for two species of the section Erectoides, A. stenophylla and A. paraguariensis (Raina and Mukai 1999). The number of ribosomal loci observed in A. porphyrocalyx agrees with what has been detected in those species, except that A. stenophylla has two pairs of 45S rDNA loci.

The fact that the karyotype of A. porphyrocalyx has distinct distribution pattern of heterochromatin, conformed by large bands of the similar size in all chromosome pairs, and has SAT chromosomes type 8 suggests that it corresponds to a distinct genome from that present in Erectoides species. On the contrary, its banding pattern is most related to that present in x = 9 species of the section Arachis. However, due to the presence of a pair of A chromosomes, different SAT chromosomes, different location of the 5S rDNA loci and a more asymmetric karyotype than that of the other x = 9 species, it is suggested that A. porphyrocalyx also does not have the G genome.

Although the four x = 9 species share the chromosome number, the karyotypic differences between A. porphyrocalyx and the remaining three species are evident. Therefore, and as was proposed (Peñaloza and Valls 2005), the reduction in the number of chromosomes might have occurred more than once in the Arachis genus.

All Arachis species with x = 9 are naturally distributed in Brazil (Fig. 1), and their evolutionary history probably is associated with watercourses. The northernward distribution corresponds to A. palustris and comprises both sides of Tocantins River in the states of Maranhão and Tocantins, between 7°22'S and 12°33'S. Arachis decora is distributed in the northeast of Goiás and in the south of Tocantins state, separated by approximately 150 km from A. palustris. In contrast, the only two populations known to A. praecox are located in the Mato Grosso state, separated by approximately 900 km of any of the other two x = 9 species of the section Arachis. Meanwhile, A. porphyrocalyx has been located in Minas Gerais state, municipality of Uberaba, near the Rio Grande, some 20 km southeast of Uberaba (19°58'S, 47°47'W). Thus, the four x = 9 species of the genus Arachis, at least in the present, belong to three different basins (Fig. 1). That is, A. decora and A. palustris share the Tocantins River Basin and are the closest species from the geographic standpoint; A. praecox belongs to the basin of the Paraguay River, while A. porphyrocalyx to the Paraná River. Notoriously the last species is located in the same basin that lodges the species with A chromosomes of section Arachis (Robledo et al. 2009).

The three x = 9 species of section Arachis are annuals and morphologically constitute different entities. Arachis praecox differs from A. palustris and A. decora by the short central axis from 2 to 3 cm, while in the other two it has about 15 cm of length. Arachis decora and A. palustris are morphologically very similar although they are distinguished because the former has bristles in the stipules, while the latter lacks them (Krapovickas and Gregory 1994). Fruit shape also distinguishes these two species. Contrarily, A. porphyrocalyx is perennial, has thickened secondary roots, a central axis between 5 and 15 cm of length, flowers concentrated at the base of the plant, lateral branches procumbents, epiphyllum with hairs, petioles and rachis with hairs and bristles (Valls and Simpson 2005), a series of features that lead to its initial allocation in the Erectoides section.

As previously discussed, A. porphyrocalyx does not share the same genome of the other x = 9 species. The presence of A chromosomes would be a strong reason for the assignment of A. porphyrocalyx to A genome, but the fact that it has two subtelocentric chromosomes, SAT chromosomes type 8, a single pair of 45S ADNr sites, and as the most significant trait, the basic number x = 9 distinguishes it from the three karyotype types established for A genome species (Robledo et al. 2009).

Taking into account the geographic distribution, morphological and cytogenetic features, the hypothesis of Peñaloza and Valls (2005), which suggests the basic chromosome number x = 9 would have originated at least twice in the evolutionary history of the genus, becomes relevant.

Some years ago, when the existence of a diploid x = 9 species with a pair of A chromosomes was not yet known, it had been proposed that a diploid x = 10 species, belonging either to the A genome group (Lavia 1998) or to a non-A genome group (Tallury et al. 2005), was the ancestor of all species with x = 9 by reduction of chromosome number. In this work, it has been demonstrated that A. porphyrocalyx has not the same genome type of the other x = 9 species, suggesting that the reduction of the number of chromosomes must have occurred more than once in the evolution of the genus Arachis; therefore, the proposed hypotheses must be updated.

In this sense, we propose that a diploid x = 10 entity, without A chromosomes and with large bands of the similar size in all or almost all, chromosome pairs could be the common ancestor of all x = 9 species as well as the x = 10 species with A and K genome of the section Arachis. The fact that these species share a same type of heterochromatin distribution pattern, different from that observed in the species so far examined of the genus Arachis, would support this proposal. From this ancestor, by chromosomal rearrangements, an entity with A chromosomes has been originated, from which all x = 10 species with A chromosomes (A genome species) derived, and by some cytogenetic phenomenon (probably disploidy), the unique species with x = 9 and A chromosomes (A. porphyrocalyx) derived (Fig. 4). Moreover, new molecular data (Silvestri; unpublished) show a minor genetic distance of A. porphyrocalyx from the A genome species, compared with any other species of the genus, including the rest of x = 9 species. In parallel, from that same first common ancestor, an evolutionary line of species without A chromosomes continued evolution, and from this, the x = 9 species with the G genome is derived by reducing the number of chromosomes (Fig. 4). Evidence in this direction results in several phylogenetic analyses (Bechara et al. 2010, Friend et al. 2010, Moretzshon et al. 2013), in which the x = 9 species appear as a sister clade to the species without A chromosomes of the section Arachis.

Figure 4.

Scheme showing the hypothesis of the multiple origin of x = 9 species with and without A chromosomes in the genus Arachis.