Emerging patterns of genome organization in Notopteridae species (Teleostei, Osteoglossiformes) as revealed by Zoo-FISH and Comparative Genomic Hybridization (CGH)

Notopteridae (Teleostei, Osteoglossiformes) represents an old fish lineage with ten currently recognized species distributed in African and Southeastern Asian rivers. Their karyotype structures and diploid numbers remained conserved over long evolutionary periods, since African and Asian lineages diverged approximately 120 Mya. However, a significant genetic diversity was already identified for these species using molecular data. Thus, why the evolutionary relationships within Notopteridae are so diverse at the genomic level but so conserved in terms of their karyotypes? In an attempt to develop a more comprehensive picture of the karyotype and genome evolution in Notopteridae, we performed comparative genomic hybridization (CGH) and cross-species (Zoo-FISH) whole chromosome painting experiments to explore chromosome-scale intergenomic divergence among seven notopterid species, collected in different African and Southeast Asian river basins. CGH demonstrated an advanced stage of sequence divergence among the species and Zoo-FISH experiments showed diffuse and limited homology on inter-generic level, showing a temporal reduction of evolutionarily conserved syntenic regions. The sharing of a conserved chromosomal region revealed by Zoo-FISH in these species provides perspectives that several other homologous syntenic regions have remained conserved among their genomes despite long temporal isolation. In summary, Notopteridae is an interesting model for tracking the chromosome evolution as it is (i) ancestral vertebrate group with Gondwanan distribution and (ii) an example of animal group exhibiting karyotype stasis. The present study brings new insights into degree of genome divergence vs. conservation at chromosomal and sub-chromosomal level in representative sampling of this group.


Results
Whole chromosome painting of XN-1 probe. As a control experiment, we first applied the XN-1 probe (prepared from the first chromosome pair of the X. nigri complement) back against the X. nigri metaphase plates and the first chromosome pair was completely painted, as expected; with prominent hybridization signals in both terminal regions. Additionally, faint hybridization blocks in the centromeric region of three other chromosomal pairs were also observed (Fig. 1a).
The hybridization of XN-1 probe in the remaining set of six Notopteridae species showed a bright hybridization signals in the centromeric regions of a medium-sized acrocentric pair in all of them. However, despite the impossibility to accurately identify this chromosome pair, it seems to represent orthologous chromosomes taking into account their similar size and morphology (Fig. 1b).

Comparative genomic hybridization (CGH).
In the set of interspecific CGH experiments, the comparative hybridization of the probes prepared from the whole genomic DNAs (gDNAs) produced only a limited number of overlapping signals (Figs 2-4). More specifically, while the gDNA probes hybridizing back against their own chromosome complements highlighted many heterochromatic blocks abundantly present in the centromeric and terminal chromosomal regions, the probes derived from the gDNA of species of other genera usually produced only weak hybridization patterns, with few consistent signals accumulated in the terminal portions of some chromosomes. Some of these signals were considerably stronger and could be related to major rDNA sites. On the other hand, when performing the intrageneric experiments in Chitala species, chromosomes were almost equally stained with both genomic probes (again with a stronger binding preferentially in terminal or pericentromeric heterochromatic regions), suggesting significant sequence homology. Additionally, several exclusive genome-specific signals were also detected (Figs 2-4).

Discussion
It is well known that orthologous genome regions from the scale of DNA sequences to entire chromosomes are sentenced to diverge over time as a byproduct of independent evolutionary histories and accompanying molecular mechanisms acting on genome. In this manner, groups with known temporal divergence offer excellent perspectives for tracking the pathways of chromosome diversification over defined evolutionary periods. Sets of phylogenetically shared chromosomal features can indicate similar levels and patterns of chromosomal evolution in a clade. Indeed, peculiar orthologous chromosome regions can favor similar and nonrandom rearrangements acting in small or large scale within a given taxonomic group 33,34 . Among vertebrates, fishes provide very attractive models for investigation of karyotype and genome evolution in lineages with very recent divergence 35 , or that experienced separated evolutionary pathways over a long divergence time 9,12 . According to molecular data, African and Asian Notopteridae lineages diverged during the Cretaceous period ~120 Mya 17 . Still, during the Cretaceous period, more specifically close to its end, the divergence between the African lineages (Papyrocranus and Xenomystus) is hypothesized to occur and, subsequently, the split of the Asian lineages (Notopterus and Chitala) is estimated to take place around 50 Mya (i.e. in the Tertiary period) 36 . Remarkably, even with respect to such divergent evolutionary times, the karyotype macrostructure of these fishes underwent low macrostructural changes, where only P. afer and C. lopis (2n = 50 and, 2n = 38, respectively) show some variation in their 2n. Except for P. afer, the karyotypes of all remaining species studied to date are composed exclusively of 42 acrocentric chromosomes 9 .
The conservative karyotype structure and 2n among Notopteridae species are maintained over long evolutionary time evolutionary time, apparently with only slight disturbances of collinearity (Fig. 5). This conserved cytogenetic trend contrasts with the more dynamic evolution showed by other osteoglossiform lineages. In fact, Mormyridae, a sister group of Notopteridae 17 whose species diverged more than 100 Mya 36 have experienced a noticeable karyotype diversification, mainly modified by pericentric inversions (NF = 52-68) (reviewed in 12,13 ).
Karyotype and chromosome diversifications may accompany speciation [37][38][39][40] . During the phyletic diversification, the rate of karyotype evolution can be highly variable in phylogenetically related fish groups 41 . In fact, despite the long divergence time, several fish groups have shown no karyotype differentiation or only some minor interspecific differences. For example, in the cohort Clupeocephala, the largest clade of Teleostei with a long evolutionary history (~153 Mya), this condition is characterized by an extensive sharing of a basal karyotype with 2n = 48 acrocentric chromosomes 42,43 . Besides the macrostructural karyotype similarities, cytogenomic analyses have supported a conservative scenario regarding the genomic organization 44,45 . Although some fish groups exhibit structural karyotype diversifications, in the case of Cichlidae, for instance, chromosomal homologies have been identified by BAC-FISH, revealing large syntenic chromosome segments being maintained conserved during evolution 46 .
The high macrostructural karyotype conservation over time represents an evolutionary process defined earlier as karyotype stasis 38 and exemplified in several fish families (e.g. 33 ). Despite that, this process has not been demonstrated for all fish clades yet, and its probable causes have been delineated for some groups 47 , such as its recurrence and phylogenetic extension 34,41 . Although karyotype stasis can be associated with a recent phyletic radiation, cases where the absence of marked chromosomal changes resists to large divergence episodes suggest that the karyotype stasis can be modulated mainly by "extrinsic" and "intrinsic" causes 47 . Extrinsic causes for such low evolutionary dynamics are generally related to population structure, where: (i) the limited occurrence of biogeographic barriers 48 , (ii) range of larval pelagic phase in marine environments or active adult migration 49,50 , (iii) level of parental care, occupation of exclusive habitats and population size, can be listed 47 . On the other hand, "intrinsic" factors are usually associated with chromosome organization, leading to variable tempo of evolutionary changes in some lineages. In fact, specific chromosome characteristics related to the organization and evolutionary dynamics of particular DNA sequences can promote less stable evolutionary environment for chromosomes in a particular group of species. Among these characteristics, the amount and distribution of heterochromatin 44,45 , specific repetitive elements 51,52 , or sequences/regions that contribute to distortions during asymmetric meiotic segregation 34,41 can be listed.
Heterochromatin is a repository of very complex sets of repetitive DNA sequences 53 , including mobile elements, notoriously involved in chromosome changes 54 , and duplications, that can propitiate substrates for illegitimate recombination, resulting in chromosome rearrangements [55][56][57] . Divergences in tandem repeats occur in different taxonomic levels, even in populations, and can precede the evolution of species 58 . In some characiform groups with very high karyotype diversification, the heterochromatic regions are extremely polymorphic among populations (e.g., 59 ). In contrast, reduced and homogeneous distribution of heterochromatin have been suggested as one of the probable causes of evolutionary inertia in certain chromosome sets 45 . In this sense, the high homology of orthologous heterochromatic regions in several Notopteridae species can indicate an accessory role of the repetitive elements in the karyotype divergence and it can serve as a possible explanation for the observed long-term evolutionary maintenance of karyotype macrostructure in this family.
A well-resolved phylogeny for knifefishes allowed investigations into the historical chromosome changes that occurred in this fish group. The cytogenetic divergence level was calibrated by comparative analysis of homologous chromosomal traits within and among Notopteridae clades. The prediction of growing karyotype identity is followed by all clades (Fig. 5) except for the clade formed by the African lineages Xenomystus and Papyrocranus. The cytogenetic traits analyzed indicate elevated karyotype conservatism among Asian genera/species (90%), but very low similarity between African lineages represented by X. nigri and P. afer (50%). Comparisons between African and Asian lineages (70%) are generally biased towards marked diversification observed in P. afer. Remarkably, this species is unique with the possession of 2n > 42, biarmed chromosomes and additionally a co-localized arrangement of 18S and 5S rDNA tandem arrays, also presented in C. chitala 9 , indicating a peculiar pattern of karyotype evolution. Since the evolutionary split between African Notopteridae lineages (~55 Mya), the level of chromosome differentiation among the extant genera was lower for all Notopteridae lineages. The clade Notopterus/Chitala with a more recent origin (~50 Mya), presents considerably less cytogenetical divergence (90%). Molina et al. 41 analyzed rates of chromosome evolution in Percomorpha groups, which comprise more than 25% of all living vertebrates, and they demonstrated considerable variation in rates of karyotype change. In this metadata study, some families showed very low rates of karyotype differentiation, reaching the patterns of chromosome stasis.
In accordance with the phylogenetic hypothesis 17,36 the cytogenetic similarity between all Chitala species (diverged ~35 Mya) was lower than between C. blanci and C. ornata, which are phylogenetically closer and demonstrated identical cytogenetic characteristics. Surprisingly, the species divergence lasting approximately 30 Mya was not sufficient for fixation of any distinct cytogenetic traits, indicating a very slow temporal divergence, like other cases of karyotype stasis in fishes 47 . In Chitala species, the extensively shared cytogenetic features are supported by CGH experiments as these compared taxa only slightly differed in overall hybridization patterns, thus pointing to a high degree of sequence homology (Figs 2 and 3). This condition is coincident with other vertebrate groups, whose evolutionary stasis is perceived across widely phylogenetically distributed clades by whole chromosomes that have remained mostly intact during 100 million years, like birds 51,60-62 , lizards 63 and mammals 64,65 species. In fishes, evidence of the high level of cytogenetic conservatism has been described among closely-related cichlid species 66 .
The set of our CGH experiments aimed to compare genomes among the different notopterid genera (Xenomystus, Papyrocranus,Notopterus, and Chitala), suggest an advanced stage of sequence divergence, except fot the bright signals, highly likely corresponding to NOR sites (as might be compared with previous rDNA FISH analysis 9 ). Such result is typical for distantly related or substantially diverged genomes, see 67,68 ). The decrease of shared sequences is hardly surprising considering the ancient time of divergence between the clades. The evolutionary genetic differentiation among Notopteridae genera is also perceived in mitochondrial and nuclear genomes using mtDNA markers and new generation sequencing technology by DArTseq markers 9,17 . The progressive temporal reduction of chromosome homology in Notopterids seems to occur slowly and suggests the cumulative action of several factors associated with chromosome evolution, such as intrachromosomal rearrangements. The internal reorganization in chromosomes is likely to be related with less identified rearrangements in fish karyotypes, such as paracentric inversions or bursts of amplification of repetitive sequences, mutations, with marked role in the karyotype evolution of several animal groups (e.g., 69 ). The similarities between the karyotype structures in Notopteridae species raise the question: are the chromosome pairs homeologous among these species? This question is pertinent to confirm chromosome markers in groups with much diversified karyotype evolution 70 , or to establish the conservation level of syntenic groups   in those with much conserved karyotype macrostructure 21 . Thus, inter-specific cross-hybridization experiments using the painting probe XNI-1 derived from the first chromosome pair of X. nigri karyotype were performed to help to clarify this question. In X. nigri, the first pair was painted, with conspicuous bright signals detected in both terminal regions (heterochromatin sites) of the chromosome after FISH. Besides, bright signals were also identified in few additional specific regions of other chromosomes of the complement. The XN-1 probe further showed diffuse and limited homology in inter-generic cross-hybridization, showing a temporal reduction of syntenic sequences. In fact, among the notopterid species of distinct genera it highlights a homeologous region, possibly heterochromatic 9 , in the centromeric region of a medium-sized acrocentric pair for all the six species. Given the larger temporal divergence among Notopteridae lineages, these shared hybridization signals are indicative of an ancestrally conserved synteny. The presence of these preserved syntenic blocks in these chromosomes open perspectives that several other homologous syntenic regions have remained conserved during the course of genome differentiation of the examined species despite the spatio-temporal isolation. Considering birds, a group with recognized diversity, comparative mapping of orthologous genes in the Z chromosome of species belonging to different orders, confirmed the maintenance of this syntenic group. However, the linear gene order has been changed, indicating that intra-chromosomal rearrangements (mainly pericentric and paracentric inversions) occurred several times during avian evolution 71,72 .
In summary, the maintenance of the gross structure of karyotypes among the Notopteridae lineages shows that similar evolutionary processes occurred within these lineages. The conserved pattern in the clades is only disrupted by divergent numerical and structural chromosome rearrangements occurred in P. afer. The extant representatives from Xenomystus, Notopterus, and Chitala genera preserved a considerable level of cytogenetic identity. In general, the temporal decrease of homology among their chromosomes suggests the involvement of intrachromosomal rearrangements that likely operate to gradually reduce the degree of collinearity and conserved synteny. The tempo of chromosome evolution in this family, except by an episodic divergence in African lineages, appears to be constant over time. To sum up, our novel cytogenetic data on sub-chromosomal level corroborate the high extent of karyotype stasis in the Notopteridae family -an important teleost group for tracking the chromosome evolution, worthy further investigation with finer-scale genomic methods.

Mitotic chromosome preparations. Seven Notopteridae species were collected in different African and
Southeast Asian River basins, as indicated in Fig. 6 and Table 1. The fishes were captured with cast-nets, placed in sealed plastic bags containing oxygen and clean water, and transported to the laboratory. The specimens were deposited in the fish collection of the Museu de Zoologia da USP, Brazil (MZUSP, vouchers 20557, 20558 and 119845). The experiments followed ethical and anesthesia conducts and were approved by the Ethics Committee on Animal Experimentation of the Universidade Federal de São Carlos (Process number CEUA 1926260315). Mitotic metaphases were prepared directly from the anterior portion of the kidney after in vivo colchicine treatment of the specimens following the protocol described in 73 . Chromosome microdissection, probe preparation and fluorescence in situ hybridization (FIsH) used for WCp. For cross-species painting, we selected the first chromosome pair from the X. nigri complement, as it is unambiguously the largest element in the karyotype, allowing us to identify precisely both homologues after Giemsa staining. Twenty copies of this chromosome were isolated by glass-needle based microdissection, and amplified using the procedure described in Yang et al. 74 . The whole chromosome-derived probe (hereafter designated as XNI-1) was labelled with Spectrum-Orange-dUTP (Vysis, Downers Grove, IL, USA) through 30 cycles of secondary DOP PCR, using 1 μl of the primary amplification product 74 . The final probe cocktail was composed of 100 ng/μg of the XNI-1 probe and 60 µg of C 0 t-1 DNA (i.e. fraction of genomic DNA enriched for highly and moderately repetitive sequences) isolated from the X. nigri male total genomic DNA (for details, see 75 ), in order to outcompete the hybridization of highly-repeated DNA sequences.  Hybridization procedure followed the protocol described in 21 and was performed for 6 days (144 h) at 37 °C in a moist chamber. After washing procedures, the chromosomes were counterstained with DAPI (1.2 µg/ml) and mounted in antifade solution (Vector, Burlingame, CA, USA). preparation of probes for CGH. The genomic DNAs from male and female specimens of all species listed in Table 1 were extracted from liver tissues by the standard phenol-chloroform method 76 . Three different experimental designs were used for this study, as outlined in Fig. 7. In the first set of experiments, the gDNA of X. nigri and P. afer were compared with the gDNA of N. notopterus against metaphase chromosomes of the N. notopterus. In the second set of experiments, the gDNA of X. nigri was compared with the gDNA of P. afer against metaphase chromosomes of the X. nigri. And in the third set of experiments, the gDNA of all species were compared with the gDNA of C. lopis, in separated CGH experiments, against metaphase chromosomes of the C. lopis. For these purposes, gDNAs of all species were labelled either with digoxigenin-11-dUTP using DIG-nick-translation Mix (Roche, Mannheim, Germany) or biotin-16-dUTP using BIO-nick-translation Mix (Roche). For blocking the repetitive sequences in all experiments, we used C 0 t-1 DNA prepared according to Zwick et al. 75 . The chosen ratio of probe vs. C 0 t-1 DNA amount was based on the experiments performed in our previous studies in fishes 23,24 . We have also performed the same set of experiments without Cot1-DNA and, although we have obtained the same results, the background was higher (data not shown). Hence, the final probe cocktail for each slide was composed by 500 ng of gDNA of one species + 500 ng of gDNA corresponding to one of the comparative species + 15 μg of derived C 0 t-1 DNA of each species. The probe were ethanol-precipitated and the dry pellets were suspended in hybridization buffer containing 50% formamide, 2 × SSC, 10% SDS, 10% dextran sulfate and Denhardt's buffer, pH 7.0.
FIsH used for CGH. For the CGH experiments, we used the methodology described in 77 , with several modifications. Briefly, a thermal aging of slides was performed prior to hybridization, at 37 °C for 2 h. Next, a treatment with RNase A (100 µg/ml, 90 min, 37 °C) took place, followed by the pepsin digestion (50 µg/ml in 10 mM HCl, 3 min, 37 °C). Denaturation of chromosomes was done in 75% formamide/2 × SSC at 74 °C for 3 min, and slides were then immediately dehydrated in 70% (cold), 85%, and 100% (RT) ethanol. The probe cocktail was denatured at 86 °C for 6 min, chilled on ice for 10 min and then applied to each slide. The hybridization was performed at 37 °C for 3 days in a dark humid chamber. Subsequently, non-specific hybridization was removed by stringent washing: once or twice in 50% formamide/2 × SSC (44 °C, 10 min each) and three times in 1 × SSC (44 °C, 7 min each). To block non-specific binding sites for antibodies, slides were incubated with 3% non-fat dried milk (NFDM) at 37 °C and subsequently hybridization signals were detected using Anti-Digoxigenin-Rhodamin (Roche) and Avidin-FITC (Sigma). Finally, the preparations were mounted in antifade containing 1.5 µg/ml DAPI (Vector).
Microscopic analysis and image processing. At least 30 metaphase spreads per individual were analyzed to confirm the 2n, karyotype structure and FISH results. Images were captured using an Olympus BX50 microscope (Olympus Corporation, Ishikawa, Japan) with CoolSNAP and the images processed using Image Pro Plus 4.1 software (Media Cybernetics, Silver Spring, MD, USA). Chromosomes were classified as metacentric (m), subtelocentric (st) or acrocentric (a), according to their arm ratios 78 . estimating levels of cytogenetic similarity among Notopteridae clades. The estimates of cytogenetic evolutionary diversification among the Notopteridae species was obtained from a data matrix with 13 specific chromosome features, including macrostructural traits (2n; presence of biarmed elements), organization of repetitive sequences in the chromosomes [e.g., rDNAs sites; (TTAGGG)n sites], and CGH homology (intraclade genomic similarity). The groups/species relationships (Notopteridae; African Notopteridae species; African/ Asian Notopteridae species; Asian Notopteridae species) followed phylogenetic hypothesis based on DNA molecular markers available for family 17,36 . Cytogenetic similarity indexes were calculated for the two Notopteridae clades.

Data Availability
All data generated or analyzed during this study are included in this published article.