The eight species of the family Haemulidae used in the cytogenetic analysis (Figure 1) were collected using fishing nets from the Rio Grande do Norte state coast in northeastern Brazil, the Fernando de Noronha Archipelago in the western Atlantic region, and the Caribbean, along the coast of Key Largo, Florida (United States). Information regarding the species and samples is shown in Table 1.

The samples were collected with permission from the Chico Mendes Institute for Biodiversity Conservation (ICMBio/SISBio) (License# 02001.001902/06-82). All the experimental procedures were approved by the Committee of Ethics for the use of Animals of the Federal University of Rio Grande do Norte (# 044/2015).

The individuals were subjected to mitotic stimulation for 18–24 h (Molina et al., 2010). The chromosome preparations were obtained from the cells of the cephalic kidney, in accordance with the method described by Gold et al. (1990). The nucleolar organizer regions (NORs) and heterochromatin were analyzed by the Ag-NOR technique (Howell and Black, 1980) and C-banding (Sumner, 1972), respectively. Additionally, the chromosome spreads were staining with CMA₃/DAPI base-specific fluorochromes (Schweizer, 1980).

The 5S and 18S rDNA probes were isolated from the genome of Rachycentron canadum (Euteleostei, Perciformes) and amplified using primers A 5′-TAC GCC CGA TCT CGT CG ATC-3′ and B 5′-CGA GCT GGT ATG GCC GTA AGC-3′ (Pendás et al., 1994) and NS1 5′-GTA GTC ATA TGC TTG TCTC-3′ and NS8 5′-TCC GCA GGT TCA CCT ACG GA-3′ (White et al., 1990), respectively. The 5S and 18S rDNA probes, which were obtained by the PCR amplification of nuclear DNA, contained 200 bp and 1400 bp of the rRNA genes, respectively. The mapping of 5S and 18S rDNA sequences was performed by in situ hybridization (Pinkel et al., 1986). The 5S rDNA probes were labeled by nick translation with biotin-14-dATP (Roche, Mannheim, Germany), and the 18S rDNA probes were labeled with digoxigenin-11-dUTP (Roche, Mannheim, Germany), according to the manufacturer’s specifications.

About thirty metaphases of each individual were analyzed using optical microscopy under a 1000× magnification, and photographed using an Olympus^TM BX51 epifluorescence photomicroscope, coupled with an Olympus DP-72 digital capture system using the cellSens^® Olympus^TM software. For the preparation of karyotypes, the chromosomes were classified in terms of their arms ratios (Levan et al., 1964) and organized in the descending order of size.

The survey of karyotype data for members from the Eupercaria series (sensu Hughes et al., 2018) was performed by running searches in several scientific web portals.

To ensure adequate phylogenetic representation, karyotype data (2n and NF – chromosome arms) of 671 species (∼10% of total species) distributed into 62 families (38.5% of the families) and 15 orders (88% of the orders) of Eupercaria (data summarized as Supplementary Material) were analyzed. The karyotypes of species within the Eupercaria series were categorized based on 2n and NF values greater than, equal to, or smaller than the basal karyotype of 2n = 48 acrocentric chromosomes.

NF (chromosome arm number) was determined considering metacentric/submetacentric chromosomes having two arms and subtelocentric/acrocentric chromosomes having one single arm. In species with sex chromosome systems, the karyotypes of the homogametic sex were considered as the standard for the species.

The cytogenetic data set for members from the Haemulidae family reveals a remarkable karyotype conservatism. The karyotype stasis in the grunts was characterized by the extensive sharing of a standard karyotype composed of 2n = 48 acrocentric chromosomes, with a small variation in size between the largest and smallest chromosomes of the karyotype, distribution of heterochromatin in the centromeric regions of chromosomes, and distribution of Ag-NOR sites recurrently on the same chromosome pair.

In a broader context, stasis has been found to be maintained even in case of comparisons of diverse chromosome regions (Motta-Neto et al., 2012), at a wide phylogenetic spectrum, wide periods of divergence (>25 M.a), and varied biogeographic scales (Caribbean and western Atlantic).

Heterogeneous heterochromatins, when are present in variable amounts among chromosomes could suggest diverse evolutionary origins and dynamics in the karyotype complement. In contrast, the heterochromatin content in the haemulid species showed a regular and reduced distribution in the centromeric regions of the chromosomes, with only the NORs showing a high GC content.

Comparative analyses of replication bands in chromosomes from the haemulid species have shown regions of repetitive DNA with synchronous replication patterns, suggesting that it was likely that they shared functional and compositional similarities (Motta-Neto et al., 2012). In fact, cross-amplification tests for microsatellite sequences have shown a high level of success in Haemulon spp. (Quintero et al., 2018), providing support for the evolutionary conservation of repetitive DNA.

Analysis of repetitive DNA classes has served as a very useful approach to assess karyotype changes among groups of fishes. The physical mapping of rDNA in the chromosomes allows the identification of cryptic rearrangements in the chromosomes (e.g., Borges et al., 2019) and cytotaxonomic markers (Nakayama et al., 2012; Gornung, 2013; Almeida et al., 2017). However, the interspecific recurrence of rDNA distribution patterns on the chromosomes from the haemulid species restricts their use for this purpose. In fact, in this family, the Ag-NORs/18S rDNA sites are simple, and in most of the species, they were positioned in the short arm of the smallest chromosome pair (24th pair). In some species, they may also occur in other pairs of chromosomes (pair 10 in H. melanurum, pair 5 in B. chrysargyreum, and pair 18 in C. nobilis), but the repertoire of changes is quite restricted.

The 5S rDNA sites were shown to be more evolutionarily diverse among the haemulid species. They were localized in one or two chromosome pairs; however, they also exhibited a considerable regularity of number and position in homeologous chromosomes, usually in the pairs 5 (long arm) and 10 (short arm). Often, these sites were non-syntenic with 18S rDNA; however, co-localized 5S/18S rDNA arrays were found in H. melanurum (pair 10) and P. moricandi (pair 24) (Figure 4).

The mapping of the 5S rDNA sites revealed five basic patterns of the organization of rDNA regions in Haemulidae (Figure 4). The first, which was the most frequent (35% of spp.), consisted of karyotypes with a single site in the pericentromeric position of chromosomes in pair 10 and was present in Anisotremus surinamensis, Haemulon flavolineatum, H. parra, Haemulon squamipinna, B. chrysargyreum, and H. melanurum, as well as in species of more basal genera like Conodon nobilis. The second (21% of spp.) showed the presence of sites in the terminal arm of chromosomes in pair 5 and was present in Haemulon aurolineatum, Haemulon steindachneri, Haemulopsis corvinaeformis, and A. virginicus. The third consisted of sites on chromosomes in pairs 5 and 10 (28% of the spp.) and was present in Haemulon plumierii, Haemulon bonariense (Nirchio and Oliveira, 2014), possibly in H. aurolineatum (Nirchio et al., 2007), and in C. nobilis.

The fourth pattern involved the presence of a syntenic 5S/18S rDNA array in the chromosomes of pair 10 (7% of the species) and was exclusive to H. melanurum; the fifth pattern, which showed a 5S/18S rDNA array in the chromosomes of pair 24 (7% occurrence), was present in P. moricandi (Figure 4).

Chromosomal stasis mainly extends to the conservation of the order of genes (Ellegren, 2010). This condition in Haemulidae has been ascertained through similarities of the replication band patterns in the chromosomes (Motta-Neto et al., 2012). Additionally, the number and similarity of the localization of 5S and 18S rDNA sites have aided the identification of homeolog chromosomes, which have been widely identified in several fish groups (Vicente et al., 2001; Centofante et al., 2002; Costa et al., 2016).

Changes in the 5S rDNA sites in the grunt karyotypes were restricted to only two chromosome pairs (pairs 5 and 10). Similarly, 18S rDNA sites almost invariably occur in pair 24, which is the smallest chromosome pair, evidencing the occurrence of syntenic groups and the low evolutionary dynamics of these repetitive sequences. Thus, the phylogenetic diversification of rDNA sites mainly reveals “variations on the same theme,” indicating a slow and recurrent “birth-death” process of these sequences, with sporadic co-localized 5S/18S rDNA arrangements resulting from preferential transpositions between rDNA regions.

Environmental factors and the life history of the species likely contribute to the karyotype patterns of marine fish (Molina and Galetti, 2004; Sena and Molina, 2007).

The roles played by ecological or physical barriers on gene flow and the origin of genetic diversity are well known. The action of physical barriers is evidenced by greater karyotype diversity in freshwater fish species, while in case of marine species, a greater degree of conservation is normally observed (Brum and Galetti, 1997).

The species of Haemulidae have large populations distributed over wide geographic areas, where oceanographic variables (temperature, salinity, phytoplankton biomass, zooplankton biomass, and station depth) play an important role in the distribution patterns of ichthyoplankton (Souza and Mafalda-Junior, 2019).

The western Atlantic haemulid species have broad geographic distribution areas, which range from 800 to 7,000 km² (Figure 4), depicting a geographical context very different from that normally found in case of freshwater fish species. This environmental scenario contributes sufficiently to gene flow, reducing the possibility of the formation of isolated groups, in which, by chance, chromosome rearrangements could be more frequently fixed. Analysis of the extensive geographic distribution, with a marked spatial overlap of sibling species, has revealed that besides vicariant events, ecological speciation also participates in the evolutionary diversification of this family of fishes (Rocha et al., 2008). In fact, in this family, recurrent sympatric speciation processes occurring in short periods of time (∼5 M.a), especially in species of the genus Haemulon, have been observed (Tavera et al., 2012; Tavera and Wainwright, 2019); recent events of genetic introgression have also been observed in some cases (Bernal et al., 2017).

In ecological speciation processes, with the fragmentation of ecological niches, in general, the split of the gene pool occurs without remarkable genetic changes or high selection pressures (Rocha and Bowen, 2008), thereby producing less propitious conditions to fix conspicuous chromosome rearrangements.

An extensive sharing of karyotypes in an evolutionary lineage could be caused from a recent diversification burst, during which genetic or chromosome changes have not yet been achieved. However, although some Haemulon species have been shown to demonstrate recent diversification (Rocha et al., 2008; Tavera et al., 2018), merely this condition seems insufficient to explain the karyotype stasis in members from this family. In fact, shared cytogenetic patterns are maintained among genera with long periods of divergence, which may reach up to 20 M.a (Tavera et al., 2018). Other factors, such as large populations, environmental continuities, and large distribution areas, seem to contribute to the panmixia of haemulid populations.

A test of the role of biogeographic barriers in the population structure of members from Haemulidae is offered by the plume of the Amazon and Orinoco rivers, which divides the Brazilian and Caribbean Provinces and plays an important role in the diversification of the faunas of these regions (Rocha, 2003; Bowen and Karl, 2008).

Cytogenetic analyses of populations of fishes from the genera Haemulon and Anisotremus located in the north and south of the Amazonas/Orinoco barrier have revealed a dichotomous pattern of karyotype divergence. Divergences of the rDNA sites were perceptible when the populations of H. aurolineatum from the Caribbean (Nirchio et al., 2007) and those from the Brazilian coast were compared (Motta-Neto et al., 2011b). On the other hand, populations of A. virginicus from the Caribbean and Brazilian coastal areas and those of B. chrysargyreum from the Fernando de Noronha Archipelago (Western Atlantic) region and the Caribbean did not present a detectable karyotype differentiation.

The interpopulational karyotype patterns underscore the chromosome conservatism in this group, even under distinct biogeographical pressures strengthened by action of the Amazonas/Orinoco barrier.

The dimension of karyotype stasis has been noted in several groups of Percomorpha (=Percomorphaceae) (Galetti et al., 2000; Molina, 2007), the most species-rich clade of modern fishes (Hughes et al., 2018). This clade has been reported to have a controversial taxonomic cohesion; however, the use of comparative genomics databases has promoted a remarkable phylogenetic redefinition, leading to the identification of its major lineages (Near et al., 2012; Betancur-R et al., 2017; Hughes et al., 2018). Its enormous diversity and increasingly robust phylogenetic relationships have made this group an excellent model to investigate karyotype evolution.

Among the 9 Percomorpha series, Eupercaria, the most diverse series, with more than 163 families and 6,000 species (sensu Hughes et al., 2018), encompasses orders with marked karyotype conservatism. In fact, members of the clade Percomorpha have been shown to present karyotypes with 2n = 48 acrocentric chromosomes both in basal and recent groups, clearly indicating that this pattern is symplesiomorphic (Galetti et al., 2000; Motta-Neto et al., 2012; Calado et al., 2014; Molina et al., 2014; Costa et al., 2016). However, thus far, its phylogenetic extent has only been inferred superficially.

Presently, the karyotype conservation estimated in members of the Eupercaria series, allowing a robust view of their karyotype evolution. In ten of the fifteen orders, most species show karyotypes of 2n = 48. Analysis of groups with greater diversity and cytogenetic information representativeness (Figure 5) revealed a high frequency of this diploid number (74–100%). Several groups, such as Chaetodontiformes (80% of spp.), Serranidae, which belongs to the large order Perciformes, or incertae sedis families such as Haemulidae, Sciaenidae, and Lutjanidae (86–100% of spp.), also presented high structural conservatism (NF = 48). The NF maintenance trend has been shown to continue when orders for which lower levels of cytogenetic information are available were considered (Gerreiformes, Ephippiformes, Lobotiformes).

Among the Eupercaria series groups, Labriformes represents one of the most morphologically and ecologically diversified clades in size, shape, and color (Westneat and Alfaro, 2005). Previous cytogenetic studies in this group have showed intermediate rates of chromosomal changes (Molina et al., 2014). Our data reinforce this condition, indicating considerable diversification in the NF (79% of species), despite a more conservative 2n variation (61% of species), with regards to a basal karyotype with 2n = 48a. On the other hand, the order Tetraodontiformes showed the most accentuated diversification of 2n (present only in 5.6% of species) and NF values (4.3% of the species), among the groups. This order, which comprises 438 spp. (Fricke et al., 2019), represents one of the main branches of the teleost diversification; its members present notable morphological and ecological characteristics, and variations in genome size (Brainerd et al., 2001; Santini and Tyler, 2003; Jaillon et al., 2004), which seem to have influenced their rather singular karyotype trends (Sá-Gabriel and Molina, 2005).

Evolutionary stasis is a broad process that involves the absence of conspicuous modifications in different dimensions throughout the evolutionary history of the organisms; its causes have not always been well understood. In some cases, this process may be transient, and has been shown to vary with regards to the occurrence of disruptive factors (Burt, 2001). In some cases, it has been shown to be characterized by the maintenance of a trait that has persisted for millions of years as a possible result of stabilizing selection (Wake et al., 1983). In adaptive terms, evolutionary stasis may reflect developmental homeostasis or static permanence at adaptive peaks (Lerner, 1954; Charlesworth et al., 1982; Kirkpatrick, 1982; Mayr, 1982; Templeton, 1982).

This process could still be related to parameters such as time, inertia, or selection. An explanation of temporal causes would involve the extensive sharing of a karyotype pattern in a group that recently underwent radiation and still did not have enough time to fix karyotype changes (Sola et al., 1981). Another possibility stems from the evolutionary inertia of karyotypes, caused by the absence of environmental destabilizing triggers (extrinsic to chromosomes) or cytogenetic factors (intrinsic to chromosomes). The extrinsic factors favoring gene flow include the absence of geographic (or ecological) barriers, dispersive capacity, population size, and the occurrence of historically stable environments. An additional possibility, but one that is less likely, is that stasis could be caused by the action of stabilizing selection, promoting active restriction to changes in chromosomes. Alternatively, it could be the product of orthoselective processes, where different karyotypes present the tendency to fix certain types of chromosome rearrangements along their evolutionary trajectories, promoting the formation of symmetrical karyotypes, with chromosomes showing approximately the same sizes and morphologies (White, 1973; King, 1981).

Using the available cytogenetic data set from the Eupercaria series, it was possible to analyze the validity of some explanations for the slow karyotype diversification. This clade encompasses evolutionary groups with extremely high diversity (e.g., with regards to natural history, ecological patterns, biogeography, habitat variety) (Nelson et al., 2016) and with very long divergence (70−38 M.a) (Yang et al., 2018); this makes it unlikely that karyotype stasis represents an adaptive condition stricto sensu or is a result of a short divergence time. On the other hand, given the evidence of extensive chromosome synteny, one cannot rule out the possibility that chromosome particularities of a group or related groups, such as low and homogeneous heterochromatin contents, contribute to the promotion and establishment of rearrangements in members from the Eupercaria series.

The vast majority of Eupercaria species are marine organisms (Betancur-R et al., 2017), which are present in large populations, and in general, have extensive distribution limits (Nelson et al., 2016). Under these conditions, a set of species of these groups show remarkable karyotype conservatism (Molina, 2007; Motta-Neto et al., 2011a, b, 2012; Molina et al., 2014) suggesting a role of the environment in the slow process of chromosome diversification.

These conditions presented in members from the Eupercaria series strengthen the hypothesis that environmental and intrinsic factors influencing the chromosomes of a group of fish can act synergistically, underscoring the chromosome conservatism in this group. The period of greatest divergence for members from the Eupercaria series (>70 M.a), which coincides with that for groups that share basal karyotypes, still clarifies that processes of karyotype stasis in fish can be maintained for extremely long periods of time.

Evolutionary karyotype patterns of large vertebrate groups are estimated based on the phylogenetic spectrum, temporal patterns, and cytogenetic approaches. Cytogenetic data from a representative number of haemulid species, which were used here as an evolutionary model, and the meta-analysis of the karyotype patterns of members of the Eupercaria contributed for a better understanding of the process of karyotype stasis in fishes. Haemulid karyotypes are remarkably conserved with regards to the macro- and microstructure of the chromosomes, suggesting the maintenance of extensive synteny throughout their evolutionary history. The ancient periods of evolutionary divergence among its lineages do not support the hypothesis that karyotype stasis is a byproduct of recent processes of speciation. On the other hand, these data suggest that the causes of this evolutionary process are complex, resulting from the synergistic action of the intrinsic characteristics of the chromosomes, together with their biological characteristics and the geographical context in which the species are inserted. The sharing of a generally homogeneous and reduced content of heterochromatin in members of the Haemulidae family seems to minimize the destabilizing role promoted by a high and heterogeneous content of repetitive DNAs and a repository of extensive sequence repertoires (e.g., transposable elements, microsatellites, multigene families) capable of generating evolutionary instability in the chromosomes. In addition, life history and biological characteristics, such as large populations, which are found in members in the Eupercaria series, could restrict the fixation of heterozygous variants. In general, cytogenetic analysis of the members from the Eupercaria revealed a broad scenario of karyotype stasis, except for the members from Tetraodontiformes, an order with diversified karyotype evolutionary trends. In summary, karyotype stasis is not an evolutionary process with a single cause, but is rather, a product of the conjunctural and synergistic effects of different factors. These data create a counterpoint for understanding the extreme diversification observed in other groups that do not meet the criteria for determining karyotype stasis.