Classical Giemsa staining did not uncover structural variation between the karyotypes of A. planirostris and A. lituratus from Amazonas. Both species have the same diploid (2n = 30/31, XY₁Y₂) and fundamental numbers (FNa = 56), with 11 metacentric and three subtelocentric chromosome pairs (22m+6st+XX/XY₁Y₂). The X chromosome was a medium submetacentric, Y₁ had a dot-like morphology, and Y₂ was a small acrocentric (Fig. 1a, b).

Figure 1.

Karyotypes of A. planirostris (a, c, e, f, h) and A. lituratus (b, d, g). Conventional staining (a, b); C-banding patterns (c, d); Ag-NOR staining (left) and FISH with rDNA 18S (rigth; e), FISH using telomeric repeats as probes (f, g), FISH with probes from the open reading frame (ORF) II of LINE-1 from A. planirostris (h).

Despite having the same karyotype, slight differences of constitutive heterochromatin distribution were observed, especially for sex chromosomes, between A. planirostris and A. lituratus. C-banding revealed CH in the centromeric region of all autosomes of both species. Additionally, in A. planirostris small heterochromatic blocks were observed in the proximal region of long arms on two metacentric chromosomes (1^st and 2^ndpairs), as well as in the distal region of short arms on three subtelocentric pairs (5^th, 6 ^th and 7 ^th), which are adjacent to the location of active Ag-NORs. The Y₁ chromosome was euchromatic, and the Y₂ had centromeric heterochromatin and additional blocks on the long arms. Likewise, the X chromosome showed centromeric heterochromatin and blocks on the short arms. The long arm of the X chromosome however, was not particularly enriched with heterochromatin (Fig. 1c). A similar pattern of heterochromatin distribution on the autosomes was observed for A. lituratus, with large heterochromatic blocks on short arms of chromosome pairs 5^th-7^th (Fig. 1d). However, the 1^st and 2^nd pairs showed only centromeric blocks. The X chromosome showed centromeric heterochromatin, as well as CH blocks on the long arms. Finally, the patterns of CH distribution on Y₁ and Y₂ chromosomes were similar to those of A. planirostris.

C-banding did not disclose within-species variation in our Amazonian samples. The observed CH patterns are, however, distinct from those reported in non-Amazonian indivuduals, indicating the existence of interindividual variation in both Artibeus species (Souza and Araujo 1990, Rodrigues et al. 2003, Lemos-Pinto et al. 2012). For example, we did not detect heterochromatic blocks on the distal region of the 9^th pair in individuals of either species as previously described by Lemos-Pinto et al. (2012) for samples collected in Pernambuco state. Furthermore, our results indicate that heterochromatin distribution on X chromosomes can vary not only among species, but also within species (Fig. 2). In this regard, two patterns were previously reported for specimens of A. planirostris from Pernambuco, Northeastern Brazil: (i) heterochromatic sites at centromeres, long arm, and distally on the short arm (Souza and Araújo 1990); and (ii) heterochromatic sites at the centromere and long arm (Lemos-Pinto et al. 2012). Both results differ from the data presented here because the long arm of X chromosomes of Amazonian specimens lacked evident CH blocks. Similarly, specimens of A. lituratus from Pernambuco have two patterns: (i) centromeric, plus distal on the short arm, and long arm (Souza and Araújo 1990); and (ii) centromere and long arm (Lemos-Pinto et al. 2012). Additionally, in specimens collected in Pará state (Northern Brazil), the X heterochromatin was centromeric, distal on the short arm, and interstitial on the long arm (Rodrigues et al. 2003).

Figure 2.

Schematic representation of A. planirostris and A. lituratus sex chromosomes showing C-banding variation reported in different studies. Gray shading corresponds to heterochromatin and the euchromatic regions are depicted in white color.

Although the number of analyzed individuals (eight A. planirostris and three A. lituratus) from the Amazon is too low to make generalizations, the similar number and location of heterochromatic blocks between individuals from Pará and Amazonas (both Northern Brazil), might indicate that specimens from the same ecogeographic regions have similar CH patterns. Pará and Amazonas states are contiguous and covered mostly by Amazon rainforest, whereas Pernambuco is a coastal state, separated from the Amazon by dry forests, and transitional environments, which might serve as mild dispersion barriers. Therefore, additional studies including large sampling are required to test the hypothesis that CH variation occurs by differential turnover of repetitive DNA (derived either by their removal/amplification or by recombination), reinforced by geographical barriers through the distributional gradient of species.

Silver nitrate staining and 18S rDNA FISH detected NORs at multiple sites on chromosomes of both species, more specifically distally on the short arms of the 5^th, 6^th and 7^th pairs (Fig. 1e; data not shown for A. lituratus). The present Ag-NORs and 18S rDNA results agree with previously reported data (Morielle and Varella-Garcia 1988, Souza and Araújo 1990, Santos et al. 2002, Lemos-Pinto et al. 2012, Calixto et al. 2014). Many species in the subfamily Stenodermatinae have multiple NORs, which is considered a derivative condition. For example, this condition regards other Artibeus species (subtribe Artibeina H. Allen, 1898), Uroderma bilobatum W. Peters, 1866, U. magnirostrum Davis, 1968, Vampyriscus bidens (Dobson, 1878), and Vampyressa thyone O. Thomas, 1909 (subtribe VampyressinaBaker et al., 2016), as well as Centurio senex Gray, 1842 (subtribe Stenodermatina Gervais, 1856) (Baker et al. 1992, Santos et al. 2002, Gomes et al. 2016). However, the multiple NORs of the above mentioned groups are not necessarily located on homologous chromosomes (orthologous chromosome regions). Additionally, basal clades within Stenodermatinae (e.g., genus Sturnira Gray, 1842), and species in the same tribe as Artibeus (e.g., Platyrrhinus Saussure, 1860 and Mesophylla O. Thomas, 1901 species) do not have multiple NORs (Gomes et al. 2016). Therefore, we hypothesize that the presence of NORs on the three particular chromosome pairs of the analyzed species was a feature of the common ancestor of all Artibeus (Santos et al. 2002, Baker et al. 2016, Gomes et al. 2016). As another mammals, Artibeus NORs collocate (are adjacent) with heterochromatin and are likely associated with the amplification of heterochromatin in non centromeric regions.

In both species, in situ hybridizations detected (TTAGGG)_n telomeric sequences in all telomeres. Additionally, both species shared centromeric signals on three subtelocentric pairs (pairs 5^th, 6^th and 7^th; Fig. 1f, g). There are two potential explanations for the presence of telomeric sequences in interstitial position (ITS): (i) these sequences might be telomere motifs reallocated from the terminal region of a chromosome to another chromosome or chromosome position; and (ii) the ITS presence on the centromere derives from reorganization of repetitive sequences (satellite DNA) composing these regions, which could also indicate the presence of centromeric hotspots of recombination during Artibeus karyotype evolution (Nanda and Schmid 1994, Multani et al. 2001, Metcalfe et al. 2007, Faria et al. 2009, Kasahara 2009, Silva et al. 2016, Teixeira et al. 2016). The karyotypic evolution of Artibeus is considered extremely conservative, however the formation of the ancestral karyotype of the subfamily Stenodermatinae required extreme reshuffling (Baker and Bickham 1980, Pieczarka et al. 2013). Therefore, the ITS allocated on the pairs 5 ^th–7 ^th for both species might be remnants of chromosome rearrangements that have been amplified or lost differentially in different Stenodermatinae species. Calixto et al. (2014) have shown that many phyllostomids species present ITS, regardless of their trend of karyotypic evolution. For example, species with a conservative karyotypic evolution, such as Trachops cirrhosus (Spix, 1823) and Phyllostomus elongatus (É. Geoffroy St. -Hilaire, 1810) (both in the subfamily Phyllostominae Gray, 1825) present ITS, which suggest lineage-specific events of amplification of these sequences can occur independently. Furthermore, it is noteworthy that pairs 5^th–7^th have ITS, cetromeric and non-centromeric CH blocks, as well as the NORs in all Artibeus specimens analyzed, suggesting that differential dynamics of heterochromatin DNA in these particular chromosomes might have played a role in the establishment of their shared distinct architecture, when compared to other autosomes. Refined investigation of these chromosomes at the sequence level will help disclosing whether differential heterochromatin composition contributed to the establishment of centromeric ITS.

LINE-1 mapping on A. planirostris chromosomes revealed FISH signals near the centromere of most autosomes, except pairs 4, 7, 8, 13 and 14. (Fig. 1g). The centromeric FISH results were consistent in all analyzed individuals (n=4) and the centromeric pattern contrasts with the longitudinal distribution previously shown for most mammals, including other phyllostomid bats, Carollia brevicauda (Schinz, 1821), Lophostoma occidentalis (Davis & Carter, 1978), and Gardnerycteris crenulatum (É. Geoffroy St. -Hilaire, 1803) (Parish et al. 2002, Dobigny et al. 2006, Ferreri et al. 2011, Pieczarka et al. 2013, Sotero-Caio et al. 2015). Centromeric accumulation of retroelements in mammalian chromosomes is rare, but some cases have been described. For instance, Waters et al. (2004) found LINE-1-positive centromeres in the karyotypes of African mammals. Likewise, Sotero-Caio et al. (2015) showed centromeric LINE-1 accumulation in chromosomes of the phyllostomid bat Tonatia saurophila Koopman & Williams, 1951 (Phyllostominae). It was hypothesized that this unusual distribution might have contributed to the high degree of chromosomal reorganization in the genus Tonatia Gray, 1827. From our data, it is still premature to state that this retroelement or sequences derived from it are constitutional components of core centromeres. Similarly, because our probes comprised only a partial LINE-1 sequence, we cannot conclude that functional elements are contributing to the centromere dynamics of Artibeus. Despite the uncertainty on what factors were responsible for the massive “colonization” of LINE-1s at centromeres, processes such as gene conversion, which promote homogenization of centromeric sequences are expected to facilitate the maintenance of LINE-1 sequences in high copy numbers in this region (Shi et al. 2010).

We noticed inconsistent patterns when comparing the co-distribution of heterochromatin blocks and LINE-1 elements. Namely, in all analyzed individuals, interstitial CH blocks have LINE-1 signals in the second chromosome pair but not pair 1. Thus, non-centromeric heterochromatin formation on chromosomal arms of A. planirostris could be a result of amplification of different types of repeats (e.g. LINEs vs. satellite DNA) in specific chromosomes (Parish et al. 2002, Dobigny et al. 2006, Sumner 2008, Shi et al. 2010, Ferreri et al. 2011, Carbone et al. 2012).

The Y₁ and Y₂ sex chromosomes presented weak FISH signals, contrasting with the strong signal throughout the long arm of the X (Fig. 1h). LINE-1 accumulation on X chromosomes is a pattern observed in all mammal species, including other phyllostomid bats (Lyon 1998, Parish et al. 2002, Dobigny et al. 2006, Cantrell et al. 2008, Liu et al. 2011, Sotero-Caio et al. 2015). Parish et al. (2002) investigated the concentration of LINE-1 in C. brevicauda chromosomes, which also presents a multiple sex chromosome system. They found that the original X chromosome had higher levels of LINE-1 accumulation than the translocated autosome. The X-autosome translocation of Artibeus is different from that observed in C. brevicauda, with the small autosome component representing the short arm of A. planirostris X chromosome. In agreement with Parish et al. (2002) findings and Lyon hypothesis, we identified that Xq of A. planirostris had a significant accumulation of LINE-1, corresponding exactly to the original X chromosome. The LINE-1 accumulation on Y-chromosomes seems to be restricted to centromeric regions. In this case, we expect that similarly to other mammals, the original Y is mostly constituted of repeats other than retroelements, and that Y₂ pattern corresponds to that observed for other autosomes due to its autosomal origin.

Transposable element activity and accumulation have been linked to chromosomal rearrangements and can be directly or indirectly associated with speciation events (Lim and Simmons 1994, Dörner and Pääbo 1995, Gray 2000, Dobigny et al. 2004, Waters et al. 2004, Carbone et al. 2012, 2014). In addition, the dispersal dynamics of TEs are related to biological functions such as gene regulation, chromosomal rearrangements, X inactivation on females and horizontal transfer events among closely or distantly related species (Lyon 1998, Ostertag and Kazazian Jr 2001, Chow et al. 2010) . The distribution of repetitive elements in Artibeus might have played a significant role in shaping the chromosome architecture of the genus, and we are still unsure if this trend at centromeres can be observed in other species of Stenodermatinae. Because the LINE-1 accumulation patterns differ in the bat species analyzed to date (Parish et al. 2002, Sotero-Caio et al. 2015, present study), we hypothesize that these elements constitute potential contributors to the great karyotype reshuffling presented by some phyllostomid taxa since the divergence of their ancestral karyotype. Overall, our data suggest that different mechanisms might have contributed to the karyotype evolution of phyllostomid bats, explaining why only Artibeus and Tonatia species, but not C. brevicauda, L. occidentalis, and G. crenulatum differred in the patterns of LINE-1 distribution.