Samples, classical chromosomal analysis and banding

Five male and seven female Eneoptera surinamensis were collected in the Parque Estadual Edmundo Navarro de Andrade (Rio Claro, SP, Brazil) between May 2013 and March 2014 with the authorization of COTEC (process number 341/2013) and were maintained in captivity until ovipositioning occurred, whereas Gryllus assimilis animals were obtained from a pool of individuals that had been bred in the biotery of the Univ. Estadual Paulista—UNESP (Rio Claro, SP, Brazil). The embryo preparations for chromosome obtaining were prepared according to Webb et al. [41], with slight modifications of the fixation materials, as follows: After hypotonization was accomplished, the same volume of Carnoy’s modified solution (3:1, absolute ethanol:acetic acid) was applied for 15 minutes, then the embryos were transferred to fresh Carnoy’s solution and were stored in a -20°C freezer until use. At least 60 embryos of each species were used for cytological preparations. In addition, four adult male testes of each species were dissected and were fixed in Carnoy’s modified solution. Adult specimens (ten of each species) were stored in 100% ethanol for subsequent DNA extraction.

Karyological studies were performed using conventional staining with 5% Giemsa solution to confirm the previous karyotypic descriptions. The C-banding procedure was conducted according to Sumner [42], and fluorochrome staining (CMA₃/DA/DAPI) to identify G+C or A+T rich regions was performed as described by Schweizer et al. [43]. Female and male genomic DNA was extracted from femurs using the phenol/chloroform-based procedure described in Sambrook and Russel [44].

Probes for repetitive DNAs

The DNA probes for the 5S rRNA and H3 genes were obtained through polymerase chain reaction (PCR) amplification using genomic DNA from Abracris flavolineata as a template and the primers described by Cabral-de-Mello et al. [45] and Colgan et al. [46] for the 5S rRNA gene and the H3 gene, respectively. The sequences of the U snDNAs of Rhammatocerus brasiliensis were obtained using the primers described by Cabral-de-Mello et al. [47] for U1 snDNA and those of Bueno et al. [48] for U2 snDNA. These sequences were previously used as probes and are deposited in GenBank under accession numbers KC936996 (5S rDNA), KC896792 (H3 histone gene), KC896793 (U1 snDNA) and KC896794 (U2 snDNA). The 18S rDNA probe was obtained from a cloned fragment previously isolated from the genome of the Dichotomius semisquamosus beetle (GenBank accession number GQ443313 [45]).

The repetitive DNA-enriched samples were obtained based on the renaturation kinetics of C₀t-DNA (DNA enriched for highly and moderately repetitive DNA sequences), according to the protocol of Zwick et al. [49], with modifications [45]. It was denaturated 150 ng/μL of fragmented genomic DNA at 95°C. The reassociation time of 25 min was used for both species. The DNA was purified/extracted using a traditional phenol-chloroform procedure [44].

The telomeric probe was obtained through PCR using the self-complementary primers (TTAGG)₅ and (CCTAA)₅ according to Ijdo et al. [50]. Finally, the probes for the microsatellite motifs were directly labeled at the 5’ end with biotin-14 dATP (Sigma-Aldrich, St Louis, MO, USA) during their synthesis, including probes for mononucleotides (A)₃₀ and (C)₃₀, dinucleotides, (CA)₁₅, (CG)₁₅, (TA)₁₅ and (AG)₁₀, trinucleotides (CAA)₁₀, (CAC)₁₀, (TAA)₁₀, (GAA)₁₀, (CGG)₁₀, (CAG)₁₀, (TAC)₁₀ and (GAG)₁₀, tetranucleotides (GACA)₄ and (GATA)₈.

Fluorescence in situ hybridization (FISH)

The non-cloned 5S rDNA and U snDNAs sequences and telomeric probes were labeled through PCR using digoxigenin-11-dUTP (Roche, Mannheim, Germany). Plasmids containing the 18S rRNA gene or H3 histone gene and the C₀t-DNA fraction were labeled via nick translation using biotin-14-dATP (Invitrogen, San Diego, CA, USA). Single or two-color FISH of mitotic cells was performed according to Pinkel et al. [51], with modifications [45]. Fiber-FISH was conducted as described in de Barros et al. [52] and Camacho et al. [53] using suspensions of testis cells. Probes labeled with digoxigenin-11-dUTP were detected using rhodamine-conjugated anti-digoxigenin (Roche), and probes labeled with biotin-14-dATP were detected using Streptavidin Alexa Fluor 488-conjugated (Invitrogen).

The preparations were counterstained using 4’,6-diamidine-2’-phenylindole dihydrochloride (DAPI) and were mounted in Vectashield (Vector, Burlingame, CA, USA). The chromosomes and signals were observed using an Olympus microscope BX61 equipped with a fluorescence lamp and the appropriate filters. Grey-scale images were captured using a DP70 cooled digital camera and were processed using Adobe Photoshop CS2 software.

Genome size estimation

The flow cytometric (FCM) analyses were conducted as described by Lopes et al. [54], in the Laboratory of Cytogenetics and Cytometry, Department of General Biology, Universidade Federal de Viçosa (UFV). The nuclear DNA contents of adult G. assimilis and E. surinamensis males and females were determined using the C DNA content of a Scaptotrigona xantotricha female as an internal standard, which was confirmed by comparison with that of the international standard, Drosophila melanogaster.

To prepare the nuclear suspensions for FCM, brain ganglia were excised from the standard animal and the sample animals in physiological saline solution (0.155 mM NaCl). The samples consisted of three males and three females of each species, and each individual was manipulated and analyzed separately, representing three independent repetitions. The materials were simultaneously crushed 10–12 times in a tissue grinder using a pestle (Kontes Glass Company, NJ, USA) in 100 μL of OTTO-I lysis buffer [55] containing 0.1 M citric acid (Merck, NJ, USA), 0.5% Tween 20 (Merck) and 50 μg mL-1 RNase A (Sigma-Aldrich), pH = 2.3. The suspensions were adjusted to 1.0 mL using the same buffer, filtered through a 30-μm nylon mesh (Partec, Nuremberg, Germany) and centrifuged at 100 g in microcentrifuge tubes for 5 min. The pellets were then incubated for 10–15 min in 100 μL of OTTO-I lysis buffer and were stained using 1.5 mL of OTTO-I:OTTO-II (1:2) solution [56,57] supplemented with 75 μM propidium iodide (PI) (excitation/emission wavelengths: 480-575/550-740 nm, [58]) and 50 μg mL-1 RNase A (Sigma-Aldrich), pH = 7.8. The nuclear suspensions were filtered through a 20-μm nylon mesh filter (Partec) and were maintained in the dark for 30 min.

The suspensions were analyzed using a Partec PAS flow cytometer (Partec) equipped with a laser source (488 nm). PI fluorescence emitted by the nuclei was collected using an RG 610-nm band-pass filter and converted to 1024 channels. The equipment was calibrated for linearity and was aligned using microbeads and standard solutions according to the manufacturer’s recommendations. FlowMax software (Partec) was used for data analysis. The standard nuclear peak was set to channel 100 and more than 10,000 nuclei were analyzed. Three independent replications were conducted, and histograms with a coefficient of variation (CV) of greater than 5% were rejected.

Karyotypes and general organization of repetitive DNA

The two cricket species examined in this study have highly divergent karyotypes. Considering that fusions are more frequent in Orthopteran chromosomal evolutionary history (for example, see [8,9,59]), leading to a reduced diploid number, the karyotype of E. surinamensis could be considered more evolved relative to that of G. assimilis. Repetitive DNAs can be relevant markers for tracing the evolution of the karyotypes and genomes of eukaryotes, but there is little information concerning the sizes of cricket genomes or the organization of their repetitive DNAs and their possible role in chromosomal evolution [13,17,38,39].

The diploid number and the sex system of G. assimilis observed in this study is identical to that previously described [40] and resembles that of most of the other Gryllus species that have been analyzed [10,13,60,61,62]. However, reduction of the diploid number has been observed in this genus [10], which includes the polymorphic condition of G. assimilis [11]. The main difference observed here in relation to other Gryllus karyotypes is the chromosomal morphology and the chromosome bearing the evident secondary constriction (see above references), suggesting that although Gryllus exhibits great karyotypic stability, small structural rearrangements may have occurred during the evolution of the genus. The diploid number observed in E. surinamensis (2n = 9) is well established in this species, having been observed in certain populations [63], and considering that a higher number of diploid chromosomes is more common in Gryllidae crickets [9,12,64], the karyotype of this species appears to be highly rearranged. Autosomal and autosomal/sex chromosome reshuffling most likely caused the significant reduction in the diploid number and the origin of a multiple sex system (neo-X₁X₂Y) with large chromosomes. It is difficult to specify the evolutionary causes or the order of the specific chromosomal rearrangements that led to the origin of the extreme karyotype of E. surinamensis, due to its extensive reorganization and the scarce phylogenetic and chromosomal information available for crickets. However, Robertsonian translocations (Rb-translocation), tandem fusions and pericentric inversions are most likely involved. In all cases of Rb-translocations or tandem fusions, the telomeres appeared to be either lost during the chromosomal rearrangement or eliminated during chromosome differentiation and no internal telomere sites are observed, as indicated by our FISH-based mapping of the telomere motif, which revealed signals only in the actual telomeres. A similar condition was reported in a small number of studied grasshoppers with derived karyotypes in which centric fusion or Rb-translocation occurred [15,16]. However, due to the limited sensitivity of the classical FISH technique, the possibility of small internal telomeric sites cannot be completely ruled out. In Gryllus, if the karyotype experienced inversions, as justified by distinct chromosomal morphologies relative to those of other species, these rearrangements occurred without the involvement of telomeres or the telomeres were lost. An alternative explanation for this condition is that G. assimilis has a non-rearranged karyotype compared with those of other species of the genus, which should be evaluated using a better species sampling.

There is little information concerning heterochromatin distribution in crickets, but in the genus Gryllus, the occurrence of mainly centromeric and terminal heterochromatic blocks has been observed in certain species, i.e., G. bimaculatus, G. argentinus and Gryllus sp. [13,60,62], as observed in this study in G. assimilis. In contrast, in E. surinamensis, the C-positive blocks were most organized in the centromeric region of the autosomes, as well as in other chromosomal regions, with extensive spreading in the neo-Y chromosome. These patterns, including those observed in Cycloptiloides americanus [17], suggest the intense reorganization of the C-positive blocks in the rearranged karyotypes. These data indicate that the heterochromatin chromosomal dynamic is more intense in crickets than in other Orthoptera, such as grasshoppers, in which pericentromeric C-positive blocks are most commonly observed [65,66], although other species should be studied. The data obtained through fluorochrome staining suggests the occurrence of repetitive DNA families with distinct richness in A+T or G+C base pairs, which in E. surinamensis are G+C-rich, whereas in G. assimilis, they are mainly neutral for A+T or G+C, suggesting the restructuring of repetitive DNA families in their two karyotypes. This data was confirmed in E. surinamensis through genomic analysis that revealed the occurrence of mainly G+C-rich satDNAs (Palacios-Gimenez et al., unpublished data). In another Gyllus species, i.e., G. bimaculatus, some terminal heterochromatic blocks were enriched for two A+T-rich satDNAs families, one of which is also present in other Gryllus species, including G. rubens and Gryllus sp [13]. The occurrence of A+T-rich repetitive families in the genome of G. assimilis could not be ruled out, and DAPI⁺ blocks could not be observed in this organism due to the limits of resolution.

The chromosomal organization of repetitive DNAs obtained through C-banding was corroborated by C₀t-DNA mapping that revealed a similar pattern, as well as additional regions enriched in repetitive sequences, such as the centromere of the neo-X₁ chromosome of E. surinamensis and some faintly labelled dispersed areas. In other Orthopteran species, the C₀t-DNA fraction is also enriched in C-positive autosomal blocks, but divergent patterns have been observed in the sex chromosomes, including the derivate neo-sex chromosomes, as observed in E. surinamensis. For example, enrichment of repetitive DNAs in the neo-Y element of the grasshopper Ronderosia bergi [16] and the X₂ element of the cricket C. americanus [17] have been reported, suggesting a role of this genomic fraction in sex chromosome differentiation, causing chromatin to change into heterochromatin. The accumulation of repetitive DNAs is a common feature during the differentiation of the Y or W sex chromosomes, and it has been documented in various animal and plant species, such as Rumex species [27,67], Silene latifolia [30], lizard species [68], Schistosoma mansoni [69] and Drosophila miranda [70,71]. However, in some grasshopper species, the C₀t-DNA fraction is restricted to the centromeres of the sex chromosomes [15].

The FISH mapping results regarding the microsatellites suggest that a major strong differentiation in the chromosomal organization of the karyotypes of the two species occurred. Although the repeats are mainly dispersed in both euchromatin and heterochromatin, clustering was observed, with the repeats forming a band-like pattern, particularly those in the heterochromatic regions of the pair 1 and sex chromosomes of E. surinamensis. This result suggests that the reduction in the diploid number was followed by the “compartmentalization” of the microsatellites in the genome, with these sequences being involved in the heterochromatin, including those of the sex chromosomes. The accumulation of some of the microsatellite arrays in the neo-Y chromosomes of E. surinamensis was expected, considering that they are non-recombining elements, as suggested by the lack of contact of these three chromosomes during meiosis [8], and this data is consistent with that observed in other species [72,73,74,75]. As previously observed in the grasshopper R. bergi, the microsatellites of E. surinamensis are clearly involved in the differentiation between sex chromosomes [16]. Although poorly studied in Orthoptera, the dispersal or specific chromosomal distribution of the microsatellite arrays in band-like patterns was recently reported in other species, such as the cricket C. americanus [17] and the grasshoppers Abracris flavolineata [76], R. bergi [16], Locusta migratoria and Eyprepocnemis plorans [77]. However, next generation sequencing analysis revealed uneven and nonrandom localization of microsatellites in L. migratoria and E. plorans, with the dinucleotide motifs predominantly associated with other repetitive DNAs, such as the histone gene spacer, rDNA intergenic spacers (IGSs) and transposable elements [77]. A similar pattern was observed in E. surinamensis, in which some of the microsatellite arrays that are distributed in a band-like pattern are co-localized with some of the multigene families and the anonymous C₀t-DNA fraction.

The increased size of the genome of E. surinamensis compared with that of G. assimilis is notable, and considering the little information available concerning the size of the genomes of crickets [38,39], this is apparently a derivate characteristic. The increased size of this genome could be attributed to the amplification, spreading and accumulation of repetitive DNAs in specific chromosomal regions, such as observed for C₀t-DNA and some satellite repeats (Palacios-Gimenez et al., unpublished data), besides some microsatellites. Although, there is no evidence that the reduction in the diploid number contributed to this process. This process is more evident in the derivate sex chromosomes, most likely due to the non-recombinant state of the neo-Y chromosome and the restricted recombination of the neo-X₁ and neo-X₂ chromosomes in the female germline. Additionally, the structural mutation rates of the neo-Y chromosomes of some species are more rapid than those of other genomic regions, suggesting the reduced efficacy of natural selection in this chromosome [18,78,79,80], which is likely to occur in the genome of E. surinamensis.

The multigene families

No variation in the location and number of clusters of 18S rDNA and U2 snDNA was observed, both of which are conserved on the pair 1. In grasshoppers [81,82] and other insects, such as Lepidoptera [83], Coleoptera [84] and Heteroptera [85], extremely variable patterns of the major rDNA was observed contrasting with this study. Remarkably, the pattern of U2 snDNA on the pair 1 is highly conserved in some grasshopper species [15,48] and in the cricket C. americanus (Mogoplistidae) [17]. Conservation of the number of U2 snDNA clusters is observed also in other groups, such as Gymnotus species [86], in contrast to the case in some of the other fish species of the Batrachoididae family, in which the U2 snDNA signals appear to be widely scattered [87]. The 18S rDNA and U2 snDNA signals were located in the same sites and exhibited the same association in the two species, which was supported by the fiber-FISH results, which may indicate an ancestral characteristic in Gryllidae that was maintained despite the exceptional chromosomal divergence. However, other species in the family should be studied to confirm this hypothesis. The association or interspersion of distinct multigene families has been reported in distinct groups, but the significance of these patterns is not clear and they have no apparent selective advantage [45,82,84,88,89,90,91].

The presence of 5S rDNA and U1 snDNA in the small chromosomes of G. assimilis in contrast to their presence in the pair 1 of E. surinamensis may be the consequence of large-scale chromosomal rearrangements, as mentioned above, which translocated these genes to the same chromosome. However, the localization of 5S rDNA in the neo-Y chromosome of E. surinamensis suggests both amplification and transposition or even further chromosomal rearrangement involving sex chromosomes and ancestral autosomes bearing this gene, as has been suggested for other Orthopteran species that have neo-sex chromosomes [15,16,17]. The presence of this gene in the neo-Y chromosomes and absence in the neo-X chromosomes reinforce the differentiation between the neo-sex chromosomes of E. surinamensis that was established after the chromosomal rearrangements occurred.

Contrary to what was observed for rDNAs and U snDNAs, significant dispersion of the H3 histone gene occurred in G. assimilis and E. surinamensis, with differentially spread patterns observed in the two species, with these DNA elements clearly in blocks in G. assimilis and faint signals observed in E. surinamensis. These results suggest the independent amplification/dispersion and dynamism of the H3 histone genes at the chromosomal level. Multiple H3 histone sites, in blocks such as observed in G. assimilis, have rarely been reported in grasshoppers, such as Abracris flavolineata [48], Dichromatos lilloanus [15] and Ronderosia bergi [16]. Although, this pattern is less common, being commonly observed an interstitial cluster in the pair 8 a highly conserved character [92]. Our data suggest that the organization and evolution of the H3 histone genes may be more dynamic than was previously reported and that this dispersal in G. assimilis and E. surinamensis could be due to other mechanisms, such as association with other repetitive DNAs (transposons or satDNAs), ectopic recombination or extrachromosomal circular DNA (eccDNA), as has been proposed for rDNAs [19,81,93]. An alternative hypothesis for the dispersal of H3 histone genes is the similarity of the sequences of these genes with those of unknown repetitive DNAs. Remarkably, a main cluster of H3 histone genes occurs in the pair 1, as is the case for the other four multigene families investigated in this study. This arrangement of certain multigene families in one unique bivalent chromosome has not yet been described and could be result of the large-scale rearrangements observed in the E. surinamensis karyotype.

In conclusion, our study provided an opportunity to explore the evolutionary dynamics of repetitive DNAs in two non-model species and its possible involvement in karyotypic organization and genome increasing. As we suggested some of the differences between the two evaluated karyotypes were most likely a consequence of the evolution and distribution of repetitive DNA that were accumulated in specific chromosomal regions. Whereas certain characteristics that these species shared could indicate their ancestral relationship. Moreover, the distinct patterns of the classes of repetitive sequences that were mapped in this study suggests that the differential amplification and accumulation of this class of DNA occurred mainly in the neo-Y chromosomes, highlighting the differentiation of the neo-sex chromosomes of E. surinamensis.