Leafcutter ants are considered the most derived of the tribe Attini and their origin is considered recent (Hölldobler & Wilson 1990; Schultz & Meier 1995; Villesen et al. 2002; Fernández-Marín et al. 2004; Schultz & Brady 2008; Mehdiabadi & Schultz 2010). These ants practice what Mehdiabadi & Schultz (2010) termed “leafcutter agriculture”. Leafcutter ants of the genera Atta and Acromyrmex have been referred to as the dominant herbivores of the Neotropics (Hölldobler & Wilson 1990), because they use plant, flower and seed fragments to cultivate a mutualistic fungus garden on which they feed (Della Lucia & Fowler 1993). These ants are significant agricultural pests of the Neotropical region because they cause great damage to pastures and cultivated agricultural crops (Della Lucia & Fowler 1993, Della Lucia & Souza 2011). In the natural ecosystems, leafcutter ants play an important role in seed dispersal (Teixeira et al. 2008), in modifying the structure of the vegetation surrounding the colonies (Della Lucia & Souza 2011), and in soil aeration which facilitates plant root penetration (Della Lucia 2003).

Bacci et al. (2009) conducted a molecular phylogenetic investigation of the genus Atta for 13 of the 15 described species, using DNA sequence information from the fragments of 4 genes (1 nuclear and 3 mitochondrial genes). The data supported monophyly of the genus and indicated its subdivision into 4 monophyletic groups: Atta sensu stricto, Archeatta, Neoatta and Epiatta, which was a major redefinition in the face of the various classifications of the genus proposed by different authors using morphological data.

Species are typically identified based on external morphological traits. However, morphological traits may or may not become altered over time by the selective pressure exerted by the environment to which each species is subjected, with consequences to the genotype. Thus, cytogenetic analysis is a useful tool which enables the identification and comparison of species (or sometimes populations) because chromosomal rearrangements can result in speciation (Kasahara 2009). Cytogenetics is, therefore, a suitable tool for phylogenetic, taxonomic and evolutionary studies and inferences (MacGregor 1993).

The chromosome numbers observed within the tribe Attini range from 2n = 8 in Mycocepurus goeldii (Barros et al. 2010) and Mycocepurus sp. (Murakami et al. 1998) up to 2n = 54 in Mycetarotes parallelus (Barros et al. 2011). Cytogenetic information can be used to compare populations, sympatric species and closely related ant groups (Mariano et al. 2012).

Atta and Acromyrmex are characterized by a conserved chromosome number within each genus of 2n = 22 chromosomes and 2n = 38 chromosomes, respectively, and by similar chromosome morphology within each genera (revised in Barros et al. 2011; Cristiano et al. 2013). Recent cytogenetic studies on Acromyrmex have reported some differences in the chromosome morphology and banding patterns in 6 species from Minas Gerais, Brazil and Barro Colorado, Panamá (Barros 2010), indicating that further cytogenetic studies in leafcutter ants are warranted. In the genus Acromyrmex, Acromyrmex striatus (2n = 22) differs the most from the other already known species with respect to both morphological traits and chromosome number. Its karyotype bears closer similarity to that of the species of the genus Atta (2n = 22) studied earlier, and associated with the molecular data it provided clear insights into a distinct phylogenetic position of this species among the leafcutter ants (Cristiano et al. 2013).

Cytogenetic data is now available for 5 taxa of Atta, all of which have 2n = 22 chromosomes (Table 1). Information on chromosome banding is available for a single Atta species, i.e., A. colombica, (Table 1) with heterochromatin in the interstitial region of 4 chromosomes added to the centromeric heterochromatic blocks on all the chromosomes (Murakami et al. 1998).

Table 1.

Atta species studied to date, localities, diploid chromosome number (2n) and chromosome banding data (Fluorochromes — fl and C-Banding — cb) studied in Brazil: Minas Gerais (mg), Rio Grande Do Sul (Rs); Panama: Barro Colorado and the respective subgenera according to the molecular phylogeny by Bacci et al. (2009)

Due to the lack of cytogenetic data on leafcutter ants, the aim of the present work was to broaden the studies, adding information on 3 species of Atta as well as to enable a better understanding of the karyotype evolution of this group and further comparative studies with other phylogenetically close groups, such as the genus Acromyrmex, including Ac. striatus or studies on Trachymyrmex, considered the sister group of the leafcutter ants. This is the first time that chromosome morphology of the genus Atta was classified by measurement, allowing an accurate comparison among similar karyotypes. The classification proposed by Imai (1991) is based on heterochromatin location and details in chromosome size mainly among similar karyotypes are not highlighted, especially when the species have low amount of heterochromatin.

Material and Methods

Cytogenetic studies were conducted on 3 species of the Atta genus, i.e., Atta sexdens rubropilosa (Forel, 1908), Atta laevigata (Smith, 1858) and Atta bisphaerica Forel, 1908, all collected in the state of Minas Gerais, Brazil (Table 1). A total of 26 individuals were analyzed for A. bisphaerica, 10 for Atta laevigata and 43 for A. sexdens rubropilosa. Mitotic metaphases were obtained according to Imai et al. (1988) by dissecting the cerebral ganglia of larvae after meconium elimination. Some slides were stained with 4% Giemsa and the metaphases were observed and photographed by a BX 60 microscope with a 100X lens attached to the Q-Color3 Olympus® image capture system. At least 6 individuals per species were subjected to the C-banding and fluorochrome techniques. The chromosomes were paired and ranked in decreasing order of size to determine the karyotype.

A total of 10 metaphases per species with similar degree of condensation, without overlappings and evident centromeres were measured and classified according to the specifications of Levan et al. (1964), based on the chromosome arm ratio (r) using the following features: long arm length (L), short arm length (S) and arm ratio between the long and short arms (r = L/S). The chromosomes were classified as: m = metacentric (r = 1-1.7), sm = submetacentric (r =1.7-3), st = subtelocentric (r =3-7) and a = acrocentric (r > 7). Karyotypes were organized using the Corel Photopaint X3® and Image Pro Plus® programs. The C-banding technique was performed according to Sumner (1972) with modifications proposed by Barros et al. (2013). The GC and AT rich regions were detected using the fluorochromes Chromomycin A₃ (CMA₃) and 4'6-diamidin-2-phenylindole (DAPI) according to Schweizer (1980).

Adult ant specimens were deposited as vouchers in the reference collection of the Laboratório de Mirmecologia, Centro de Pesquisas do Cacau (CPDC/Brazil) under the following numbers: #5712 (A. sexdens rubropilosa), #5715 (A. laevigata) and #5713 (A. bisphaerica).

Supplementary material is online at Florida Entomologist 97(4) (2014) online at  http://purl.fcla.edu/fcla/entomologist/browse.

Results

The 3 species of the Atta studied presented 2n = 22 chromosomes with a karyotypic formula was 2n = 18m + 2sm + 2st (Fig. 1). The morphometric data of the chromosome pairs 10 and 11 of Atta spp. are shown in Table 2.

Small and weak heterochromatic bands were observed in the centromeric region of the chromosomes in all the 3 species (Fig. 2). Besides the centromeric blocks, the presence of a small interstitial band was observed on the long arm of the fourth pair of metacentric chromosomes for A. sexdens rubropilosa (Fig. 3a —; inbox).

The CMA₃ fluorochrome staining showed interstitial marking on the long arm of the fourth pair of metacentric chromosomes for the 3 species, indicating that this region is rich in GC base pairs (Fig. 3; color version in Suppl. Fig. 3). CMA₃ positive regions in A. sexdens rubropilosa corresponded to a heterochromatic region.

Regions with differential staining with DAPI, indicative of regions rich in AT base pairs, were not observed.

Fig. 1.

Karyotypes of Atta spp.: (a) A. sexdens rubropilosa, (b) A. laevigata and (c) A. bisphaerica with 2n = 22 chromosomes and karyotypic formula 2n = 18m + 2sm + 2st. Arrows indicate the presence of a secondary constriction on the fourth pair of chromosomes. Bar: 5µm.

Discussion

The diploid number of 2n = 22 chromosomes reported for the Atta species in the present study was also observed in other studies (Table 1). The chromosome morphology was shown to be constant in the different Atta species, unlike the observations made for the species of Acromyrmex (Barros 2010; Cristiano et al. 2013). The genera Atta and Acromyrmex are considered phylogenetically close and recognized for the constancy of their karyotype, with respect to both chromosome number and similarity in morphology in most of the species studied (revised in Barros et al. 2011; Cristiano et al. 2013). This characteristic has already been observed in the different ant groups, like those of the subgenera Myrmothrix and Myrmamblys in genus Camponotus (Formicinae) (Mariano et al. 2003) and genus Pheidole (Myrmicinae) (revised in Lorite & Palomeque 2010). However, recent cytogenetic studies on the genus Acromyrmex have revealed differences in the morphology and chromosome banding patterns in 7 species studied in Brazil and Panamá (Barros 2010; Cristiano et al. 2013).

Table 2.

Morphological analyses of two chromosome pairs of Atta spp. and Acromyrmex striatus

In the present study, the C-banding technique indicated the existence of a small amount of heterochromatin on the chromosomes of the Atta species, with a distribution pattern in the centromeric region. Similar results regarding the heterochromatin distribution were reported by Murakami et al. (1998) for A. colombica, which has centromeric heterochromatic bands on all the chromosomes and additional interstitial bands on 2 pairs of chromosomes. However, metaphase was not illustrated in the study of Murakami et al. (1998). The genome of Atta cephalotes, included in the sensu stricto group has been sequenced and reveals that there is a low percentage of satellite DNA (Suen et al. 2011), the principal component of the constitutive heterochromatin. The C-banding technique confirmed the low amount of heterochromatin when using modern genome sequencing techniques.

Fig. 2.

Metaphases of Atta spp. submitted to the C-banding technique: (a) A. sexdens rubropilosa, (b) A. laevigata and (c) A. bisphaerica. The darker centromeric regions reveal the presence of heterochromatin. Bar: 5µm.

The interstitial heterochromatic band located on the fourth pair of chromosomes of A. sexdens rubropilosa corresponded to a region positively stained with CMA₃ fluorochromes, indicating that this heterochromatic block was rich in GC base pairs (Fig. 3; color version in Suppl. Fig. 3). This correspondence to the CMA₃ bands, however, was not observed in the centromeric heterochromatic bands of the chromosomes revealed by the Cbanding technique. The interstitial heterochromatic band of the fourth pair of chromosomes was observed only for A. sexdens rubropilosa, which was certainly due to the small band size in other species.

Fig. 3.

Metaphases of Atta stained with the fluorochromes CMA₃ (metaphases on the top) and DAPI fluorochromes (metaphases on the bottom) of: (a) Atta sexdens rubropilosa, (b) Atta laevigata and (c) Atta bisphaerica. The arrows indicate GC-rich blocks and complementary negative AT regions, respectively. The fourth metacentric chromosome pair is highlighted in Fig. 3a (in box) with GC-rich interstitial heterochromatin. Bar: 5µm.

The interstitial region positively stained with CMA₃ fluorochrome on the fourth pair of chromosomes was observed in all 3 Atta species; a secondary constriction was also visible, indicating that this GC rich heterochromatic region probably corresponds to the nucleolus organizer region (NOR). To date no published reports are available on the location of secondary constrictions in this genus of leafcutter ants. The GC-rich regions generally correspond to the NOR, which are well conserved during evolution. In the Tapinoma nigerrimum (Lorite et al. 1997) and Dinoponera lucida (Mariano et al. 2008) species the CMA₃ fluorochrome pattern coincided with the NOR site verified by the fluorescence in situ hybridization (FISH) technique, similar to most eukaryotes (Reed & Phillips 1995).

Recent cytogenetic studies on Ac. striatus revealed 2n = 22 chromosomes (Cristiano et al. 2013), the same chromosome number found in all the Atta species studied to date. However, the karyotype of this species did not have the same morphological traits of the chromosomes found in Atta species and morphological analyses of chromosome pairs 10 and 11 that differed between species are shown in Table 2. In Ac. striatus, the submetacentric pair is probably homeologous to the subtelocentric pair of the Atta species investigated in the present study. This difference suggests chromosome rearrangements, as for example, duplication and heterochromatin inversion during the evolution of this species, such that the positioning of a heterochromatic block on the telomere of the submetacentric pair of Ac. striatus would be possible. However, the opposite is also possible, in which the heterochromatic block could have been lost in an ancestor of the modern Atta species. Another pair of chromosomes that was also observed to be different in Ac. striatus was the smallest chromosome pair with their characteristic morphology. This small pair clearly has the 2 arms (long and short arm) which differ in size in all the Atta species already studied, although these 2 chromosome arms are similar in size in Ac. striatus. Both chromosomes have heterochromatin on the centromere of this homeologous chromosome pair, while duplication of a portion of the centromeric heterochromatin may have modified the chromosome morphology. However, variations in the euchromatic regions cannot be disregarded.

The heterochromatin distribution pattern seen in Atta species is similar to that of Ac. striatus (Cristiano et al. 2013) and is characterized by the presence of small and weak heterochromatic bands in the centromeric region of the chromosomes. However, Ac. striatus can be distinguished by the presence of additional pericentromeric bands and a telomeric marking on the short arm of the submetacentric pair. The differences noted in chromosome morphology and heterochromatin distribution patterns revealed by a comparison of the Atta species assessed in the present study with Ac. striatus show a divergence between these ant groups.

Karyotypes of the 3 leafcutter ant species of genus Atta, added to those of previously described karyotypes, were shown to be conserved in chromosome number, morphology and banding pattern, and constancy is inferred for chromosome number within the genus for the species studied to date. It included species in 3 of the 4 groups of the Atta genus based on the molecular data (Bacci et al. 2009) pertaining to A. laevigata and A. bisphaerica in the Epiatta group; A. colombica in the Atta sensu stricto group; A. sexdens rubropilosa and A. sexdens piriventris in the Neoatta group.

Data obtained from the chromosome banding techniques are scarce not only for the species of the genus Atta but for other species of the tribe Attini as well, (Barros et al. 2010; Barros et al. 2014) even though they could be very informative for our understanding of the evolution of the tribe.

Data in the present study revealed the importance of measuring the chromosomes to enable accurate interspecific comparison among the Atta spp. as well as the published information available on Ac. striatus which revealed interesting differences. The chromosome banding data indicated the importance of the techniques mentioned above for comparison among the species within the genera with a conserved chromosome number.