Insights into the origin of the high variability of multivalent-meiotic associations in holocentric chromosomes of Tityus (Archaeotityus) scorpions

Scorpions represent an intriguing group of animals characterized by a high incidence of heterozygous chromosomal rearrangements. In this work, we examined six species of Tityus (Archaeotityus) from Brazilian fauna with a particular focus on elucidating the rearrangements responsible for the intraspecific variability of diploid number and the presence of long chromosomal chains in meiosis. To access any interpopulation diversity, we also studied individuals from four species representing distinct localities. Most species demonstrated intraspecific polymorphism in diploid number (2n = 19 and 2n = 20 in T. clathratus, T. mattogrossensis, and T. pusillus, 2n = 16, 2n = 17 and 2n = 18 in T. paraguayensis, and 2n = 16 and 2n = 24 in T. silvestris) and multi-chromosomal associations during meiosis I, which differed even among individuals with the same chromosome number. In some species, the heterozygous rearrangements were not fixed, resulting such as in Tityus clathatrus, in 11 different chromosomal configurations recognized within a same population. Based on meiotic chromosome behaviour, we suggested that independent rearrangements (fusion/fission and reciprocal translocations), occurring in different combinations, originated the multi-chromosomal chains. To evaluate the effects of these chromosome chains on meiotic segregation, we applied the chi-square test in metaphase II cells. The non-significant occurrence of aneuploid nuclei indicated that non-disjunction was negligible in specimens bearing heterozygous rearrangements. Finally, based on our analysis of many chromosome characteristics, e.g., holocentricity, achiasmate meiosis, endopolyploidy, ability to segregate heterosynaptic or unsynapsed chromosomes, (TTAGG)n sequence located in terminal regions of rearranged chromosomes, we suggest that the maintenance of multi-chromosomal associations may be evolutionarily advantageous for these species.

Introduction Karyotype comparisons between many species of plants and animals have revealed large-scale chromosomal differences [1][2][3][4][5][6][7]. Additionally, new technologies, such as genome scanning have shown that submicroscopic chromosomal variants are more common than previously thought [8][9][10]. These microscopic and submicroscopic chromosomal variants play important roles in species evolution, promoting changes in gene expression and influencing aspects of local adaptation [11][12][13][14][15][16][17]. Rearrangements that give rise to chromosome variants frequently reduce the fertility of hybrids due to problems related to chromosome segregation in meiosis and the generation of unbalanced gametes via duplication or deletion. However, unbalanced gametes are suppressed when chromosomally rearranged regions are protected from recombination [14,17].
Karyotype changes originate from deletions, duplications, inversions or translocations occur in both holocentric and monocentric chromosomes [15,18]. In monocentric chromosomes, rearrangements sometimes produce acentric or dicentric fragments, which are frequently lost during cell division. In contrast, in holocentric, all chromosome segments arising from rearrangements might segregate normally to the poles due to the attachment of microtubules along the entire chromosome length [19][20][21].
Chromosomal heterozygosity originating from translocation, fusion or fission leads to the formation of trivalent, quadrivalent, or even multivalent chromosomal associations during meiosis I. Multivalent associations of more than four chromosomes are rare or underestimated, although they have been observed in certain plants [22][23][24], invertebrates [25][26][27][28][29][30] and vertebrates [31][32][33][34][35][36]. These multi-chromosomal associations form via rearrangements involving both autosomes and sex chromosomes or autosomes alone. Generally, multivalent associations encompassing sex chromosomes appear as a linear chain, and those formed only by autosomes have a ring configuration [28,33,36,37]. In some cases, numbers of chromosomes involved in multivalent associations vary within a species, as observed in the cockroach Periplaneta americana, which presents 2n = 33 chromosomes organized in rings, representing multivalent associations of four, six or eight chromosomes [25,38]. The huntsman spider, Delena cancerides, has 2n = 22-43 chromosomes in linear chains comprised of three to 19 chromosomes [28]. Intraindividual variations have been rarely documented, but examples include one population of the monocot plant Allium roylei with 2n = 16 chromosomes and cells containing one trivalent as well as other cells containing complex multivalent associations comprising all chromosomes [22].
In buthid scorpions, heterozygous reciprocal translocations or fission/fusion rearrangements are hypothesized to be responsible for the origin of multi-chromosomal associations during meiosis I [27,29,[39][40][41][42]48]. The mechanisms responsible for correct chromosome pairing and segregation of long chromosomal chains remain poorly characterized. However, permissive meiosis, which facilitates the correct alignment and alternate segregation of chromosomes involved in complex chains, is necessary to maintain fertility and produce balanced gametes [28,33]. Moreover, only a small number of species have routinely demonstrated multi-chromosomal association in meiosis; thus, there appears to be a genetic ability to balanced chromosomal chain segregation [28,33].
In this work, we examined six species of Tityus (Archaeotityus) from Brazilian fauna using standard and molecular cytogenetic techniques with a particular focus on elucidating the rearrangements responsible for the intraspecific variability of diploid number and the presence of long chromosomal chains in meiosis. To access any interpopulation diversity, we also studied individuals from four species representing distinct localities, as shown in

Material and methods
We investigated a sample of 137 scorpions, all of which belonged to the subgenera Tityus (Archaeotityus), specifically T. clathratus, T. maranhensis, T. mattogrossensis, T. paraguayensis, T. pusillus and T. silvestris. Data describing the numbers of analyzed individuals for each species and the collection localities are listed in Table 1. Collecting permits from all localities were High variability of multivalent-meiotic associations in holocentric chromosomes of Tityus scorpions granted by the Instituto Brasileiro de Meio Ambiente e Recursos Naturais Renováveis (IBAMA) and the Instituto Chico Mendes de Conservação da Biodiversidade (ICMBIO) (23561-1; 25471-1; 40014-2; 41383-2; 48224-1). Scorpion species within the subgenus Archaeotityus present a rather similar morphology, with variegated pigmentation and a very rhomboidal subaculear tooth [59]. However, the group has not undergone a robust taxonomic revision, and the interspecific morphological differences are narrow, thus hampering species recognition. The scorpions were identified using relevant taxonomic literature [55][56][57][58][59][60][61]  Chromosomal preparations to study mitotic and meiotic chromosomes were obtained according to the technique described by Schneider et al. [39] and standard-stained with 3% Giemsa solution. Active nucleolar organizer regions (Ag-NORs) were detected by silver impregnation [62]. To locate major ribosomal cistrons and telomeric sites, the preparations were subjected to fluorescent in situ hybridization (FISH) following the procedure of Pinkel et al. [63] with modifications as described. The 28S rDNA probes were obtained by PCR using the genomic DNA of T. silvestris and the primers 28S-F 5' GACCCGTCTTGAAACACGGA and 28S-R 5' TCGGAAGGAACCAGCTACTA as described by Nunn et al. [64]. Telomeric probes were generated by PCR without DNA template using only the primers Tel-F 5' TAGGTTAGG TTAGGTTAGG and Tel-R 5' AACCTAACCTAACCTAACC. In both cases, probes were labeled with biotin-16-dUTP (1mM) by PCR and detected with anti-biotin conjugated to Alexa Fluor 488 (200μg/mL). For FISH, slides were incubated in 1% pepsin solution and RNAse/2xSSC (40 μg/mL) for 1h at 37˚C. The chromosomes were fixed in 1% formaldehyde (1xPBS/50mM MgCl 2 ) for 10 min, washed in 1xPBS for 5 min and dehydrated using an ethanol series (70%, 90% and 100%) for 5 min per step. Chromosomal DNA was denatured in 70% formamide/ 2xSSC for 3 min and 30 s at 70˚C and dehydrated in an ice-cold ethanol series. The hybridization mixture (50% deionized formamide, 20% dextran sulfate, 10% 20xSSC, 20% DNA probe) was denatured in a thermocycler for 10 min at 95˚C and applied to the slides, which were maintained in a humid chamber overnight at 37˚C. Post-hybridization washes were performed twice in 15% formamide/2xSSC for 5 min each at 42˚C and in 0.5% Tween for 5 min at room temperature. The chromosome spreads were counterstained with DAPI/Vectashield. The images of the mitotic and meiotic cells were captured in a Zeiss Imager A2 microscope coupled to a digital camera using Axio Vision software.
For the five cytogenetically analyzed species, T. clathratus, T. mattogrossensis, T. paraguayensis, T. pusillus and T. silvestris, all cells on each slide were photographed. The number of cells per individual and the chromosome configuration observed in postpachytene nuclei are described in Tables 2 and 3. Cells with a variable number of chromosomes and those in which chromosome configurations were not determined due to the complexity of the chromosome chain were categorized as "variable number" (VN). Diploid set length (DSL) was determined for the following: two males of T. clathratus, CHNUFPI 1853 (18 cells) and CHNUFPI 1863 (24 cells), which showed differences in the number of bivalents but equal numbers of chromosomes involved in multivalent associations; and three males of T. mattogrossensis, UFMT 1362 (27 cells), UFMT 1357 (14 cells) and UFMT 1375 (25 cells), which exhibited chains of four Table 2 Table 3.

Conventional Giemsa staining
The chromosomes of all species studied in this work (Table 1) were holocentric, exhibited synaptic and achiasmatic behavior during male meiosis I, and had multivalent associations in postpachytene cells. Intraspecific variabilities in diploid number and number of chromosomes involved in the chains were observed in the analyzed species (Tables 2 and 3). Tityus maranhensis showed chromosomal features similar to those previously reported by Mattos et al. [29]; thus, herein only the results obtained using molecular cytogenetic techniques are described.
Mitotic metaphase cells of five Tityus (Archaeotityus) species showed the following diploid numbers: 2n = 19 and 2n = 20 in T. clathratus, T. mattogrossensis, and T. pusillus (Fig 2A-2H, 2M-2O); 2n = 16, 2n = 17 and 2n = 18 in T. paraguayensis (Fig 2I-2L); and 2n = 16 and 2n = 24 in T. silvestris (Fig 2P-2T). Diploid number variability occurred within populations of T. clathratus from Amajari, T. mattogrossensis from Cuiabá, and T. paraguayensis from Campo Grande and between populations of T. silvestris. The chromosomes of all species gradually decreased in size (Fig 2). However, differences in chromosome size were verified in specimens of T. clathratus with 2n = 19, which presented one chromosome that was slightly larger than the other elements in the diploid set (Fig 2A); in two individuals of T. mattogrossensis from Cuiabá with 2n = 20, a single small-sized chromosome was identified (Fig 2H). Three individuals of T. silvestris (2n = 24) demonstrated terminal constrictions in two chromosomes ( Fig 2S). Unsynapsed interstitial and/or terminal chromosomal segments were observed in early pachytene cells of T. clathratus (2n = 19-20), T. mattogrossensis (2n = [19][20] and T. paraguayensis (2n = 17); these regions were generally continuous with completely paired segments. Total synapsis of the chromosomes was verified in late pachytene cells in which multivalent associations were also visualized (not showed). Some late pachytene cells of all species showed continuous and entirely synapsed chromosomes. The numbers and percentages of postpachytene cells with different chromosomal configurations were compiled in Tables 2 and 3. Among the five Tityus species, multi-chromosomal associations appeared closed in most postpachytene cells and open or completely disorganized in later substages of meiosis I (Fig 3).
The postpachytene cells of T. mattogrossensis with 2n = 19 presented three configurations: 10II, 6II+III+IV and 8II+III (Fig 3M-3O). Specimens with 2n = 20 exhibited 10II and 8II+IV ( Fig 3P). Additionally, among the specimens with 8II+IV, the following differences regarding the sizes of chromosomes involved in the quadrivalent association were observed: 1) chain with chromosomes of similar size in males from Alto Araguaia ( Fig 3P); 2) chain with two large and two medium-sized chromosomes in the specimen from Chapada dos Guimarães ( Fig 3Q); and 3) chain with two large, one medium and one small-sized chromosomes in the male from Januária (Fig 3R). To verify whether differences in the sizes of these chromosomes were significant, DSL values were determined. The comparison of total DSL values was not significant, indicating no differences among individuals in terms of total chromosome length (bivalents and chromosomes in the chain). However, when we compared only the length of quadrivalents, the population from Januária differed from other two (Fig 4), revealing that chromosomes of distinct size were rearranged (scree plot on the x-axis 82.8%).
In T. paraguayensis with 2n = 16 and 2n = 18, postpachytene nuclei exhibited eight and nine bivalents, respectively (Fig 5A and 5B). In some cells, the bivalents revealed conspicuous regions with a low degree of condensation and the presence of gaps (Fig 5C). In specimens with 2n = 17, the configuration most commonly observed was 5II+VII (Fig 5D and 5E). In anaphase I, the total disassembly of the bivalents and multi-chromosomal associations was verified ( Fig 5F).

Ag-NORs and in situ hybridization
Mitotic metaphase cells of T. clathratus, T. mattogrossensis, T. paraguayensis, T. pusillus and T. silvestris revealed Ag-NORs localized on the subterminal or terminal regions of two chromosomes (Fig 2B, 2D, 2G, 2L, 2O, 2Q and 2T). In T. clathratus, T. mattogrossensis and T. paraguayensis, the NORs occurred on small-sized chromosomes compared with on medium-sized chromosomes in T. pusillus and T. silvestris. Additionally, association of the NORs with prominent terminal constriction was visible in T. silvestris (Fig 2T). The number of Ag-NORs was confirmed by FISH using a 28S rDNA probe (Fig 7). However, the analysis of postpachytene cells revealed that the ribosomal cistrons were not localized in well-paired meiotic elements resembling bivalents in all species. Tityus clathratus (10II) showed ribosomal cistrons in the terminal region of one bivalent (Fig 7A). Specimens of T. maranhensis (2n = 20) presented intraspecific variation in the number of bivalents (9II and 10II), but in both cases, only one bivalent was the carrier of the ribosomal cistrons (Fig 7B). The same pattern of 28S rDNA localization was observed in postpachytene nuclei of T. mattogrossensis with 10II and 8II+IV (Fig 7C and 7D), T. paraguayensis with 8II, 9II and 5II+VII (Fig 7E and 7F), and T. pusillus with 10II and 7II+IV (Fig 7G and 7H). However, T. silvestris with 2II+XII revealed ribosomal genes localized in two chromosomes involved in the multivalent association (Fig 7I). Mitotic metaphase cells demonstrated ribosomal sites in the terminal regions of two chromosomes. Frequently, there were differences in the sizes of ribosomal cistrons between homologues. In T. silvestris (2n = 24), the 28S rDNA sites were coincident with the constriction present in one chromosome pair. Metaphase II cells exhibited ribosomal sites localized in only one chromosome, indicating correct and balanced chromosomal segregation.
Among the six Tityus species, FISH with the (TTAGG)n probe revealed typical telomeric signals in all chromosomes or at least at one chromosome end (Fig 8). Additionally, there was no evidence of positive signals in the interstitial regions of the chromosomes in the analyzed species. Early prophase I nuclei contained brightly labeled regions in one nuclear polo, indicating the clustering of telomeres to form a bouquet. Early prophase I cells exhibited the brightest telomeric signals due to the low degree of chromosome condensation. Postpachytene nuclei also exhibited bright labeling only in the terminal regions of the bivalents and/or chromosomes of the chains (Fig 8A, 8B, 8D and 8F).

Discussion
In this work, we cytogenetically characterized six out of the eight Tityus (Archaeotityus) species that occur in Brazil, emphasizing diploid number variability and chromosome behavior during meiosis. These findings represent approximately 20% of the known described species belonging to the subgenus Archaeotityus. Representatives of this subgenus are included in the most diverse scorpion genus, Tityus, with about 210 described species, all of which are exclusive to the Neotropical region [66]. Although there is no comprehensive phylogenetic hypothesis for this genus, in a recent study of the southernmost American buthids, Ojanguren-Affilastro et al. [67] recovered Tityus as a monophyletic clade and also the four studied species of the subgenus Archaeotityus. Based on the six species described cytogenetically (T. clathratus, T. maranhensis,T. mattogrossensis, T. paraguayensis, T. pusillus and T. silvestris), Archaeotityus scorpions have diploid numbers ranging from 2n = 16 to 2n = 24, with a predominance of 2n = 20 ([29, 68-69], this work). The 2n = 24 reported herein for one population of T. silvestris is the highest chromosome number for this group.
The extraordinary diploid number diversity recorded for Tityus and Buthidae species, in general, is coupled with genome organization. The first sequenced scorpion genome (Mesobuthus, Buthidae) suggested the presence of gene duplication events resulting in more than 32,000 protein-coding genes, thus representing more than any sequenced arthropods [70]. Based on genomic and karyotype data, we hypothesize that some chromosomal features may facilitate gene duplication in buthids, i.e., the presence of polyploid cells, holocentric chromosomes, absence of meiotic recombination, and ability to segregate multi-chromosomal associations. Prophase I cells with an increased numberof chromosomes (bivalents and chains) have been observed in T. bahiensis [50,71,72], T. paraguayensis [29], and T. pusillus (present work, Fig 5L and 5M). This type of cell appears to recur in buthids and likely arose via endopolyploidy (see [29]). In this manner, some duplicated chromosome or chromosomal segments may have been incorporated into the diploid set, giving rise to changes in diploid number or structural chromosome organization. Holocentricity permits the correct segregation of any chromosomal segment, while achiasmatic chromosomal behavior and the balanced segregation of chromosomal chains avoid the formation of gametes with potentially disadvantageous deletions. These factors may be responsible for chromosome variability, which appears to occur freely in buthids, as well as the expansion of gene families and neofunctionalization, as described by Cao et al. [70] and Sharma et al. [73]. Alternatively, as proposed for angiosperms [74][75][76], polyploidy/paleopolyploidy followed by repeated disploidy events also lead to the high variability of diploid numbers during meiosis I. In scorpions, however, the whole-genome duplication has been debated (for reviews, see [70,77]).
Intraspecific variability in diploid number has been recorded in 40% of the 18 karyotyped Tityus species [42,69]. Similar to the species investigated herein, previous studies revealed variations at the inter-and intra-population levels and attributed differences in chromosome number to fission/fusion rearrangements [29,39,41,42,48]. In specimens of T. clathratus with 2n = 19, we observed one chromosome that was slightly larger than the others in the diploid set, corroborating the chromosome fusion from 2n = 20. This same rearrangement should occur in specimens of T. mattogrossensis and T. pusillus with 2n = 19. However, in these species, the karyotypes with 2n = 19 did not originate only via simple fusion rearrangement because meiotic cells showed multi-chromosomal associations with more than three chromosomes (see Tables 2 and 3). In a simple fusion, the expected result is a trivalent association during meiosis I, as observed in one male specimen of T. mattogrossensis (UFMT 1369), in T. bahiensis with 2n = 5, T. curupi with 2n = 31, and T. obscurus with 2n = 13 and 2n = 11 [39,41,42]. In T. paraguayensis, it is premature to speculate whether the karyotype with 2n = 16 originated from 2n = 18 by fusion or if 2n = 18 arose from 2n = 16 by fission. Nevertheless, karyotype 2n = 17 is clearly not a hybrid between individuals with 2n = 16 and 2n = 18 because postpachytene cells exhibited a multivalent association with seven rather than three chromosomes. However, the large difference in diploid number between two distant populations (1.040 Km) of T. silvestris, 2n = 16 in the state of Mato Grosso and 2n = 24 in the state of Amazonas, must be evaluated because these individuals may correspond to distinct taxonomic entities with a low degree of morphological differentiation.
In addition to diploid number variability, all species characterized in this work showed multi-chromosomal associations during meiosis, indicating that at least some individuals are carriers of heterozygous rearrangements. These multivalent associations differed even among specimens with the same chromosome number, such as T. clathratus, 2n = 19 -3II+XIII, 4II +XI; T. mattogrossensis, 2n = 19 -6II+III+IV, 8II+III; and T. pusillus, 2n = 19 -7II+V, 8II+III. Additionally, multivalent associations appeared in individuals lacking odd diploid numbers, such as T. clathratus, 2n = 20 -5II+X; T. mattogrossensis, 2n = 20 -8II+IV; T. pusillus, 2n = 20 -7II+VI; and T. silvestris, 2n = 16 -2II+XII. Based on these results, we suggested that independent chromosomal rearrangements (fusion/fission and reciprocal translocations) occurring in different combinations are present in specimens with the same diploid number. This phenomenon is particularly apparent in specimens of T. mattogrossensis with 8II+IV, which showed significant interpopulational differences regarding the rearranged chromosomes that constituted the quadrivalent association. Furthermore, certain heterozygous rearrangements likely become established in populations, as observed for T. silvestris from Mato Grosso, in which all examined specimens exhibited 2II+XII. According to Faria and Navarro [17], chromosome rearrangement fixation depends on many factors, such as selection, meiotic drive, and small and semi-isolated demes. In Tityus species, all these factors may be present, but given the recurrence of chromosome changes, it is unlikely that rearrangements arose and were fixed only by genetic drift.
The intra-individual variability of the chromosome chains observed in T. clathratus, T. paraguayensis and T. pusillus may have several explanations: 1) a variable degree of heterosynapsis between non-homologous chromosome regions; 2) the early dissociation of the chromosomes constituting the multivalent chains (for details, see 29 and 78); and 3) the persistence of unsynapsed chromosome regions, which are visualized in chains as gaps or open configurations. The two first factors lead to variations in the number of bivalents and/or chromosomes in chains, such as T. clathratus with 3II+XI, 4II+X, 4II+XI or 4II+X, 4II+XI or 5II+IX, 5II+X, 5II+XI. These events have also been described for other analyzed buthids based on light and electron microscopy observations [29,39,78] and appear to be recurrent in other organisms bearing heterozygous rearrangements such as grasshopper, lizard and rodent [79][80][81]. To confirm whether any other mechanisms of chromosomal change are responsible for intraindividual variation in the number of bivalents without alterations in the number of chromosomes in chains, we determined the DSL for two specimens of T. clathratus with 3II+X, 4II+X and 5II+X (CHNUFPI 1853) and with 2II+XI, 3II+XI, 4II+XI (CHNUFPI 1863). In both specimens, the comparison of DSL values was not statistically significant, indicating that intra-individual variation in postpachytene cells occurred due to a variety of heterosynaptic interactions between chromosomes.
As mentioned above, the presence of unsynapsed chromosomes segments may also be reflected in the intra-individual variability of chromosome configurations visualized at meiosis I, such as in T. paraguayensis with 5II+VII and 5II+VIII, and T. pusillus with 7II+IV and 7II +V. In these two species, when a gap appears on one chromosome, the chain appears to be composed of the largest number of elements. In an elegant study in mice bearing multiple Robertsonian translocations, Manterola et al. [82] demonstrated that unsynapsed chromosomal regions persist until late prophase I, escaping the pachytene checkpoint that ensures correct synapsis. According to the authors, this permissive pachytene checkpoint contributes to the maintenance and spreading of chromosomal translocations intopopulations in an evolutionary context.
To evaluate the effects of multi-chromosomal association on meiotic segregation, we applied the chi-square test in metaphase II cells of individuals with odd diploid numbers. The non-significant occurrence of aneuploid metaphase II nuclei indicated that non-disjunction was negligible in specimens bearing heterozygous rearrangements. As multi-chromosomal chains are widespread among Tityus species, we already made assumptions regarding the correct number of chromosomes in metaphase II as a mechanism to avoid unbalanced gametes, which potentially reduce the reproductive fitness of heterozygotes. Although segregation patterns have only been analyzed for certain specimens, these results provide evidence regarding the role of selection in the maintenance of chromosomal rearrangements in scorpions.
Studies examining specific chromosomal regions in scorpions have increased significantly in recent years [29, 39-42, 48, 83-87]. Data regarding the localization of NORs and rDNA has revealed remarkable stability, particularly in Tityus species, in which more than 80% of the species presented a similar pattern, i.e., NORs/rDNA in the terminal regions of two chromosomes. Furthermore, in the six species analyzed here, FISH revealed that the 28S rDNA genes were always located in bivalents with a similar size, despite the extensive variability of postpachytene configurations, with the exception of T. silvestris from Mato Grosso, which exhibited rDNA sites in two chromosomes of the chain. Heckmann et al. [88] discussed the functional interrelationship between holocentricity and terminal NOR position. After revising the positions of NORs in many holocentric species and examining the structures of holocentric chromosomes in Luzula elegans, the authors suggested that rDNA sites in distal centromere-free chromosome regions ensure chromosome stability, avoiding breaks. Scorpiones are particularly interesting subjects in which this prediction may be tested because it exhibits both holocentric (Buthidae) and monocentric chromosomes (Bothriuridae, Hormuridae, Scorpionidae and Scorpiopidae).

Conclusions
In summary, we demonstrated that karyotype variability in Tityus (Archaeotityus) species originated not only via chromosome fission/fusion but also via reciprocal translocation, which gave rise to multi-chromosomal associations during meiosis. Moreover, in the absence of abnormal metaphase II cells, we validated the balanced segregation of chromosomes in individuals bearing heterozygous rearrangements. Finally, based on our analysis of many chromosome characteristics, e.g., holocentricity, achiasmate meiosis, endopolyploidy, ability to segregate heterosynaptic or unsynapsed chromosomes, (TTAGG)n sequence located in terminal regions of rearranged chromosomes, we suggest that the maintenance of multi-chromosomal associations may be evolutionarily advantageous for these species.