Parodontidae is a small family of Neotropical fish with 32 valid species (Fricke et al., 2021), grouped into three genera: Parodon Valenciennes, 1849, Saccodon Kner, 1863 and Apareiodon Eigenmann, 1916 (Pavanelli, 2003). They have a wide geographical distribution in South America and Panama rivers, except for some coastal basins in the Atlantic and Patagonia (Pavanelli, Britski, 2003).

The La Plata basin covers an area of 3,2.10⁶ Km² over five South American countries (Berbery, Barros, 2002). Paraná, Uruguay, La Plata, and Paraguay rivers constitute the main tributaries of the La Plata basin (Berbery, Barros, 2002; Júlio Jr. et al., 2009). The Sete Quedas waterfalls (a 114-meter high natural geographic barrier currently submerged due to the Itaipu dam built about 150 km downstream to waterfalls) represented a significant geographic barrier for fish species dispersal in the Paraná River (Agostinho, Zalewski, 1996; Abell et al., 2008). Based on the geographic barrier provided by the Sete Quedas waterfalls, the La Plata basin ichthyofauna showed gene flow restriction between these areas, isolating species from the Lower and Upper Paraná River systems (Abell et al., 2008; Júlio Jr. et al., 2009). The Itaipu dam is currently considered a division point between the Lower and Upper Paraná systems (Júlio Jr. et al., 2009). In this region, a fish pass system, called Canal da Piracema, was built to connect downstream and upstream rivers in the Itaipu dam (Makrakis et al., 2007). The Canal da Piracema could allow long-distance migratory species to find suitable spawning and nursery areas in the tributaries of Itaipu Reservoir and in the floodplain located upstream (Agostinho et al., 1993; Gomes, Agostinho, 1997). Many endemic species in the Lower Paraná colonized and dispersed to the Upper Paraná after constructing the Itaipu dam and the Sete Quedas waterfalls submersion (Júlio Jr. et al., 2009).

Parodon nasus Kner, 1859 inhabits lotic and lentic waters, has migratory reproductive behavior in shoals and spawn with external fertilization (Godoy, 1975). The Cuiabá River (Upper Paraguay River drainage in the Lower Paraná River system) is the type locality for P. nasus (Kner, 1859). According to Britski et al. (1999), P. nasus occurrence would be restricted to rivers belonging to the Paraguay River basin. In opposite, Parodon tortuosus Eigenmann & Norris, 1900, nowadays a junior synonym of P. nasus (Pavanelli 1999, 2003), had its distribution described for the Upper Paraná River system (Britski, 1972). Analyzing representatives from the Lower and Upper Paraná River, Pavanelli (1999, 2003) did not find robust morphological characters to corroborate P. tortuosus, extending the occurrence of P. nasus for all rivers that constitute the La Plata Basin.

Parodon nasus and its junior synonym P. tortuosos were also cytogenetically investigated (Moreira-Filho et al., 1984, 1985; Jesus, Moreira-Filho, 2000; Vicente et al., 2001; Centofante et al., 2002; Bellafronte et al., 2005, 2011; Schemberger et al., 2011, 2014, 2016; Ziemniczak et al., 2014). Despite some slight chromosome differences, representatives of P. nasus from the Lower and Upper Paraná River systems demonstrate a shared karyotype (Bellafronte et al., 2005).

In addition to chromosomal and morphological data, the molecular markers are helpful in taxonomy and systematics as they consider nuclear or organellar sequence evolutionary history (Padial et al., 2010; Schilick-Steiner et al., 2010; Travenzoli et al., 2015). These methodologies directly seek variation in the DNA level, allowing phylogenetic and gene flow analysis (Travenzoli et al., 2015; Nascimento et al., 2018; Santos et al., 2019; Traldi et al., 2020). Thus, the concept of integrative taxonomy has been used, which uses different sources of information for the delimitation of species in supposedly homogeneous taxa (Padial et al., 2010; Ruane, 2015; Gąsiorek et al., 2017; Grković et al., 2017; Traldi et al., 2020). In this context, the study compared karyotypes and genetic parameters of P. nasus representatives from four rivers distributed between the Lower and Upper Paraná River system to assess the role of the Sete Quedas barrier on their distribution, population structuring, and speciation.

Biological samples. Thirty-five Parodon nasus specimens (Fig. 1A) from four rivers in the La Plata basin were collected: Cuiabá River - Lower Paraná River; and Mogi-Guaçu River, Passa Cinco River, and Paiol Grande stream - Upper Paraná River (Tab. 1 and Fig. 1B). Specimens were deposited as vouchers in the ichthyological collection at Núcleo de Pesquisas em Limnologia, Ictiologia e Aquicultura (NUP), Universidade Estadual de Maringá, Maringá and Museu Nacional, Rio de Janeiro (MNRJ), Brazil (Tab. 1).

FIGURE 1

TABLE 1

River	Hydrographic system	Location	Specimens	Voucher
Cuiabá River	Lower Paraná River	56°09’59”W 15°34’41”S	8	MNRJ 29787
Mogi-Guacu River	Upper Paraná River	47º22’03”W 21°55’34”S	10	MNRJ 28199
Passa Cinco River	Upper Paraná River	47°46’54”W 22°22’19”S	4	NUP 23215
Paiol Grande stream	Upper Paraná River	45°41’00”W 22°40’34”S	6	MNRJ 28131

Cytogenetics analyzes. Mitotic chromosomes were obtained from the animals’ kidneys, according to Bertollo et al. (2015). Heterochromatic regions were identified by the C-banding procedure, according to Sumner (1972), with the final stage of propidium iodide staining (Lui et al., 2012). Fluorescence in situ hybridization (FISH) was performed following Pinkel et al. (1986), and three sequences of repetitive DNAs were physically mapped: 18S and 5S rDNAs, and satellite DNA pPh2004. 18S probe was obtained according to Hatanaka, Galetti Jr. (2004) and labeled with Biotin-16-dUTP (Biotin-Nick Translation Mix; Roche Applied Science). The sequence of the 5S rDNA was obtained and labeled with digoxigenin-11-dUTP (DIG-11-dUTP; Jena Bioscience) by PCR using the genomic DNA of P. nasus and the primers 5SA and 5SB (Martins and Galetti, 1999). The pPh2004 probe was obtained according to Vicente et al. (2003) and labeled with digoxigenin-11-dUTP (Dig Nick Translation Mix; Roche Applied Science). FISH was performed under the following conditions: 300 ng of each probe, 50% formamide, 10% dextran sulfate, 2xSSC, 37 ºC for 16h. Signal detection was carried out using Streptavidin Alexa Flour 488 (Molecular Probes) and anti-digoxigenin-rhodamine (Roche Applied Science). The chromosomes were stained with DAPI (0.2 µL/mL) present in Vectashield mounting medium (Vector) and analyzed in an epifluorescence microscope (Leica DM 2000) coupled to a DFC3000 G CCD camera (Leica). The chromosomes were classified following the arms ratio rule (Levan et al., 1964) and arranged into karyotypes.

Molecular analyzes. Genomic DNAs of nineteen individuals of P. nasus from the four different rivers of the La Plata basin were extracted from liver or fin samples following the CTAB (cetyltrimethylammonium bromide) method of Murray, Thompson (1980). DNAs were used to amplify the barcode region of the gene Cytochrome C Oxidase subunit I (COI) by PCR using the primers Fish F1 and Fish R1 (Ward et al., 2005). Reaction mix contained: 1x Taq Reaction buffer (200 mM Tris pH 8.4, 500 mM KCl), 1 mM MgCl₂, 0.2 mM dNTPs, 0.4 µM of each primer, 1 U Taq DNA polymerase (Invitrogen), and 40 ng of DNA. The following reaction program was used: initial denaturation for 10 minutes at 94 ºC, 35 cycles of 94 ºC for 1 min, 54.5 ºC for 45 sec and 72 ºC for 90 sec, and a final extension at 72 ºC for 10 min. PCR products were purified with the Illustra GFX PCR DNA and Gel Band Purification (GE Healthcare) and sequenced in an ABI-prism 3500 Genetic Analyzer (Applied Biosystems).

Sequences were checked and corrected in the software Geneious v 7.1.9 (Kearse et al., 2012) and stored as voucher in GenBank. The obtained sequences were aligned using the algorithm Clustal W, integrated into Geneious, and used in the populational analysis. The sequences were arranged into groups corresponding to specimens in sampled areas. The genetics distances were calculated in MEGA v 7.0 (Kumar et al., 2016), under the Kimura-2-parameters evolution model and 1000 repetitions, considering the four rivers and the Paraná basin structure. The Mantel test was performed in the software Alleles in Space (Miller, 2005). The nucleotide (π) and haplotype (H) diversities were estimated in the software DnaSP v5 (Librado, Rozas, 2009). Population structuring and analysis of molecular variance (AMOVA) (Excoffier et al., 1992) were performed by the Arlequin 3.5.2.2 software (Excoffier, Lischer, 2010). Structural analysis was performed by assignments of each individual to the respective populations using Bayesian Analysis of Population Structure - BAPS 6 (Corander et al., 2004, 2008). The haplotype network was generated in PopArt 1.7 software (Leigh, Bryant, 2015) through the median-joining criterion (Bandelt et al., 1999).

For the phylogenetic analyses and delimitation of species, one Leporinus piau Fowler, 1941 COI sample (HM405030.1) was used to root the trees. Leporinus piau COI sequence was aligned with the P. nasus sequences with Clustal W. The final matrix was submitted to jModeltest2 (Posada, 2008) to select the best-fit model of nucleotide evolution to be used in downstream analysis. A Bayesian inference tree was generated in MrBayes 3.2 program (Huelsenbeck, Ronquist, 2001), applying 100.000.000 interactions of Markov Chain Monte Carlo (MCMC), sampling trees every 20.000 generations.

For the species delimitation, two methods were used: (a) General mixed Yule coalescent (GMYC) (Pons et al., 2006; Fujisawa, Barraclough, 2013); (b) Automatic Barcode Gap Discovery (ABGD; Puillandre et al., 2012). In the GMYC method, an ultrametric gene tree was inferred in Beast v. 2.6.1 (Bouckaert et al., 2019) under the Yule model prior and Strict Clock. Markov chains included 10.000.000 generations, storing trees every 1.000 generations to obtain 10.001 trees. Tracer v. 1.6 (Rambaut et al., 2014) was used to examine the average standard deviation of split frequencies and the convergence of MCMC searches, considering > 200 as an appropriate and effective sample size value. The first 1,000 trees were discarded as burn‐in, and the 9,001 trees were summarized in the Maximum Clade Credibility tree (MCC) from the posterior distribution in TreeAnnotator v. 2.6 (Bouckaert et al., 2019). The tree was imported into the R software (R Development Core Team, 2013), and delimitation of species was made with the package SPLITS (Species’ Limits by Threshold Statistics - http://r-forge.r-project.org/projects/splits/) using the single threshold method. For the ABGD analysis, the alignment generated for the sequences was used as an input file in the ABGD web server (https://bioinfo.mnhn.fr/abi/public/abgd/abgdweb.html), the Kimura (K80) TS/TV was selected for distance mode, and the others parameters were kept default.

Cytogenetics analysis. All P. nasus analyzed presented 2n = 54 chromosomes, karyotype formulae composed of 48m/sm + 6st and FN = 108. Heteromorphic sex chromosomes were not detected. Heterochromatic blocks were distributed in the centromeric or terminal regions along the chromosomes of specimens analyzed (Figs. 2A-D). A single cluster of 18S rDNA at the q terminal region in the subtelocentric pair 25 was detected (Figs. 3A-D), while the 5S rDNA cluster was observed at the p arm terminal region in pair 25 in the four populations (Figs. 3A-D). Individuals from the Cuiabá River also presented an additional 5S rDNA site at the proximal region in the metacentric pair 17 (Figs. 3A). pPh2004 DNA satellite sites were located in 4 chromosomal pairs: at the terminal regions of the long arms in chromosome pairs 6, 13, 26, and 27 for the four analyzed populations (Figs. 3A-D - boxes).

FIGURE 2

FIGURE 3

Molecular analysis. Nineteen partial COI sequences were obtained for P. nasus from representatives from four rivers (accession numbers in GenBank: OL584306 - OL584324). All sequences presented a high quality and showed no evidence of indels, deletions and stop codons. The sequences were aligned, and a 631 bp matrix was generated. A total of 7 haplotypes were observed, with nucleotide diversity (π) of 0.00410 and haplotypic (H) of 0.819. The haplotypes were distributed as follow: Haplotype H1 comprises all individuals belonging to the Cuiabá River; haplotype H2 was shared between individuals from the Mogi-Guaçu River and Paiol Grande stream; haplotype H3 was shared by the populations of the Mogi-Guaçu and Passa Cinco rivers, and Paiol Grande stream; haplotype H4 and H7 were exclusive to the Passa Cinco River; and the haplotypes H5 and H6 were exclusive to the population of the Mogi-Guaçu River (Fig. 4A).

FIGURE 4

The intraspecific genetic distance for each population ranged from 0 to 0.19%, while the interspecific genetic distance ranged from 0.15 to 0.88% (Tab. 2). When considering the division of the La Plata Basin in Lower and Upper Paraná River, the intraspecific genetic distance in the Lower Paraná was 0% and in the Upper Paraná was 0.16%. In addition, the interspecific genetic distance between Lower and Upper Paraná River samples was 0.79%. The Mantel test demonstrated a robust positive correlation between genetic distance and geographic distance (r = 0.92; p < 0.001; Fig. S1 ).

TABLE 2

	Cuiabá River	Mogi-Guaçu River	Passa Cinco River	Paiol Grande stream
Cuiabá River	0.0000 (0.0000)
Mogi-Guaçu River	0.0083 (0.0033)	0.0019 (0.0011)
Passa Cinco River	0.0088 (0.0035)	0,0017 (0.0007)	0.0016 (0.0011)
Paiol Grande stream	0.0070 (0.0032)	0.0015 (0.0009)	0.0017 (0.0010)	0.0010 (0.0009)

For population structure analysis, specimens were gruped in: Lower Paraná (Cuiabá River) and Upper Paraná (Mogi-Guaçu River + Passa Cinco River + Paiol Grande stream). The BAPs result demonstrated the probability of formation of these two clusters (k = 2) of 0.99292 (Fig. 4B). The AMOVA analysis showed variance values among groups of 83.57% and within populations of 13.14%, with an FST value = 0.8686 (Tab. 3). The pairwise FSTs values between each population ranged from -0.00515 to 0.93182 and showed the population structure between individuals from the Cuiabá River and those from the other localities (Tab. 4).

TABLE 3

Source of variation	Degrees of freedom	Sum of squares	Variance components	Percentage of variation
Among groups	1	16.692	2.16220Va	83.57
Among populations within groups	2	1.471	0.08523Vb	3.29
Within populations	15	5.100	0.34000Vc	13.14
Total	18	23.263	2.58743
Fixation Indices FSC: 0.20043 FST: 0.86860 (p<0.001) FCT: 0.83566

TABLE 4

	Cuiabá River	Mogi-Guaçu River	Passa Cinco River	Paiol Grande stream
Cuiabá River	-	0.00000+-0.0000	0.00901+-0.0091	0.00901+-0.0091
Mogi-Guaçu River	0.88462	-	0.85586+-0.0390	0.62162+-0.0317
Passa Cinco River	0.92147	-0.00515	-	0.07207+-0.0227
Paiol Grande stream	0.93182	0.06250	0.28760	-

The best-fit model of nucleotide evolution inferred using the corrected Akaike information criterion (AICc) for the phylogenetic and species delimitation analyzes was HKY+G (-LNL = 1402.4146). The Bayesian tree showed the individuals belonging to the Cuiabá River in a clade. In contrast, individuals belonging to the other rivers (Upper Paraná) were distributed in two clades (Fig. 4C). The two species delimitation methods showed identical results. The GMYC analysis suggested the number of two species (outgroup included, confidence interval 2-9 species; the maximum likelihood of null model: 119.1766; the maximum likelihood of GMYC model: 128.8128; likelihood ratio: 19.27235; LR test: 6.532229e-05***; threshold time: -0.0061). The second method, ABGD, resulted in 10 partitions: one partition presented eight species, while the other nine found two species. For partition 5, two species (outgroup included) were observed (prior maximal distance P = 0.007743).

Numerous valid species in Parodontidae occur in La Plata basin, but most are restricted to the Lower or Upper Paraná River system (Pavanelli, 2003). Parodon nasus and Apareiodon affinis Steindachner, 1879 are the single valid taxa with broad distribution in Lower and Upper Paraná River systems (Pavanelli, 2003). Due to gene flow restriction between Lower and Upper Paraná River systems, differences in morphological characters, chromosomes, and gene sequences have been reported in P. nasus and A. affinis (Moreira-Filho et al., 1980; Jorge, Moreira-Filho, 2000, 2004; Pavanelli, 2003; Bellafronte et al., 2005, 2013; Nascimento et al., 2018). Some taxa assignments within Parodontidae are controversial despite chromosomal differences because family members lack reliable diagnostic morphological traits to support accurate phylogenetic analysis (Pavanelli, 2003).

Sequence analysis of the COI can help identify species (Hebert et al., 2003). Ward (2009) showed that DNA Barcoding is a valuable tool for identifying fish species and proposed the value of 2% of the genetic distance between fish species as a threshold for their separation, a value confirmed by Pereira et al. (2013) for Neotropical fish. Still, according to Pereira et al. (2011, 2013), this value should only be used as a starting point for investigating divergence between specimens. Other characteristics of the species/group studied, such as their evolutionary history, should be considered before defining a limit for species (Pereira et al., 2011). In Parodontidae, integrative studies of DNA Barcoding and chromosomal analysis proved to be effective in delimiting species, even though morphological characters have been overlapped in the taxa (Bellafronte et al., 2013; Nascimento et al., 2018; Santos et al., 2019; Traldi et al. 2020). In analyzed P. nasus samples, the Sete Quedas barriers did not generate speciation between representatives from Lower and Upper Paraná River systems since the data demonstrated genetic and chromosomal features compatible with a single taxon and population structuring.

The K2P genetic distance among P. nasus individuals from Upper and Lower Paraná was below the threshold value of 2%, indicating a single species. In addition, the observed interspecific distance values are less than 10X those found for intraspecific distances. The barcoding gap (the difference between the greatest intraspecific distance and the smallest interspecific distance) greater than 10X is used to separate cryptic species, regardless of the K2P distances less than 2% (Hebert et al., 2004). In Parodontidae, Parodon nasus X P. morerai Ingenito & Buckup, 2005, Apareiodon piracicabae Eigenmann, 1907 X A. vittatus Garavello, 1977, and A. machrisi Travassos, 1957 X A. cavalcantePavanelli & Britski, 2003 demonstrated K2P distances less than 2% (Bellafronte et al., 2013; Traldi et al., 2020). However, the cytogenetic and morphological characteristics validate these species, possibly indicating recent speciation events between them (Bellafronte et al., 2013; Traldi et al., 2020). The two species delimitation methods and the Bayesian phylogenetic analysis also demonstrated that the P. nasus analyzed samples represent a single species.

All P. nasus sampled in this study shared the 2n, karyotype formulae and FN, demonstrating extensive conservation in their karyotype structure between Lower and Upper Paraná River systems. Previous cytogenetic studies showed a similar karyotypic organization among P. nasus populations (Moreira-Filho et al., 1984, 1985; Jesus, Moreira-Filho, 2000; Vicente et al., 2001; Centofante et al., 2002; Bellafronte et al., 2005; Ziemniczak et al., 2014). The C-banding is an excellent chromosomal marker for identifying sex chromosomes in Parodontidae species since the heterochromatin tends to be located in a sizeable interstitial portion of the W chromosome (Schemberger et al., 2011, 2019). In opposite, autosomes show just small pericentromeric and terminal heterochromatin regions in Parodontidae karyotypes (Bellafronte et al., 2011; Ziemniczak et al., 2014). Parodon nasus show a proto sex chromosomes pair (without a sex-specific heteromorphic chromosome region) corresponding to chromosome 13 in the karyotype (Schemberger et al., 2011), observed in all analyzed samples in this study. Parodon nasus chromosomes show the most prominent heterochromatin regions in the pPh2004 satellite DNA sites and 45S rDNA cluster, besides that pericentromeric and terminal region of the chromosomes (Centofante et al., 2002; Bellafronte et al., 2005; Schemberger et al., 2011). Despite variations usually found in C-bands, the chromosomal marker was homogeneous among all P. nasus specimens analyzed.

In situ localization of the repetitive DNAs has been used to compare Parodontidae karyotypes (Vicente et al., 2001; Centofante et al., 2002; Vicari et al., 2006; Bellafronte et al., 2011; Schemberger et al., 2011, 2014, 2016; Ziemniczak et al., 2014; Nascimento et al., 2018). Usually, the mapping of repetitive DNAs shows chromosomal sites diversification between closely related species but tends to be conservative when comparing Parodontidae populations (Bellafronte et al., 2005, 2009, 2011; Rosa et al., 2006; Schemberger et al., 2011). Slight differences in rDNA sites location were demonstrated among P. nasus sampled in distinct areas (Vicente et al., 2001; Bellafronte et al., 2005; Paula et al., 2017). This difference is attributed to additional 5S or 45S rDNA sites due to repetitive DNA units transposition and could not represent a substantial genetic difference in the genomic comparison. The mapping of 18S and 5S rDNAs showed a syntenic location for these genes in chromosome 25, and specimens from the Cuiabá River showed an extra site of 5S rDNA in the metacentric pair 17, probably resulting from a transposition event in the karyotype. Despite that, the syntenic condition of rDNAs is considered an apomorphic feature for P. nasus (Bellafronte et al., 2011), reinforcing their low levels of karyotypic differentiation.

Repetitive DNAs usually generate chromosomal remodeling between populations or closely related species (Cioffi et al., 2009; Parise-Maltempi et al., 2013; Poltronieri et al., 2013; Blanco et al., 2017; Crepaldi, Parise-Maltempi, 2020). The distribution and expansion of pPh2004 satellite DNA had a significant role in chromosomal diversification in Parodontidae (Bellafronte et al., 2011; Schemberger et al., 2011; Ziemniczak et al., 2014). Despite the high pPh2004 satellite DNA sites dispersion among Parodontidae karyotypes, P. nasus from Upper and Lower Paraná River systems have a homogeneous condition in the four samples analyzed. Unlike P. nasus, the populations of A. affinis distributed along the La Plata basin studied by Nascimento et al. (2018) demonstrated a karyotypic and molecular variety among the regions of Lower and Upper Paraná River, despite the absence of morphological differences, suggesting that this group may be a complex of species. Besides karyotype formulas, pPh2004 satellite DNA sites demonstrate extensive chromosomal remodeling among A. affinis sampled in Lower and Upper Paraná River regions (Nascimento et al., 2018).

Geographical barriers isolate fish populations, causing intraspecific polymorphisms and genetic divergences among populations over time (Sekine et al., 2002). The population differentiation is expected between Lower and Upper Paraná River representatives caused by the geographical barrier historical of the Sete Quedas waterfalls and currently by the Itaipu dam. Parodon nasus from Lower and Upper Paraná River represent two highly structured populations, as demonstrated by the values of pairwise FST and the analyzes of AMOVA and BAPs. The Mantel test show a significant correlation between geographic distance and genetic distance in P. nasus. In an isolation-by-distance (IBD) model, genetic distance should be positively correlated with geographical distance (Felsenstein, 1976). It is expected that more abundant and less distant populations have greater gene flow and, therefore, be more genetically similar and have greater genetic diversity relative to smaller, rare and more distant populations (Eckert et al., 2008). Thus, the high genetic distance found in P. nasus from the Cuiabá River compared to other samples could be reinforced by the long-distance isolation between Lower and Upper Paraná River. Besides that, the haplotype network showed that specimens from the Cuiabá River have a unique haplotype, not shared with specimens from Upper Paraná, and presented an exclusive second pair of chromosomes carrying 5S rDNA cistrons. The haplotype found here for the Cuiabá River is the same observed by Bellafronte et al. (2013), reinforcing the low diversity of this population and the lack of haplotype sharing with Upper Paraná.

The population structure between specimens from Lower and Upper Paraná River was accessed and demonstrated for several groups of fish, including species with migratory behavior, through different molecular markers (Sekine et al., 2002; Zawadzki et al., 2008; Pereira et al., 2009; Piálek et al., 2012; Nascimento et al., 2018; Prado et al., 2018). However, it is known that the flooding of Sete Quedas waterfalls, for the construction of the Itaipu hydroelectric plant, allowed the dispersion of species from Lower to Upper Paraná (Agostinho, Júlio Jr., 2002; Júlio Jr. et al., 2009). Apareiodon affinis was the second most collected species in the Canal da Piracema, occurring mainly in unstructured littoral and lentic waters, but it was classified as a sedentary species with no parental care (SNPC) as a reproductive strategy (Makrakis et al., 2007). Parodon nasus was not collected in the Canal da Piracema (Makrakis et al., 2007). Still, it was only mentioned in the region of influence of the Itaipu reservoir after its formation (Cecílio et al., 1997), which may indicate that specimens from the Lower Paraná may have migrated to the Upper Paraná region in the interval between the floods of the Sete Quedas waterfalls and the construction of the Itaipu dam. In the São Francisco River basin, P. nasus was observed after the transposition of the Piumhi River (Bellafronte et al., 2010). However, it is not clear whether this dispersion was due to the change in the river course or through the flooded areas of the swamp and the effect of this dispersion (Bellafronte et al., 2010).

The Sete Quedas waterfalls (currently Itaipu dam) represent an important geographical barrier restricting gene flow in fish populations. This isolation effect has promoted chromosomal e genetic diversification in some Parodontidae species from Lower and Upper Paraná River systems. In P. nasus analysis, the similarity of the chromosomal features allied to genetic population parameters indicates a single species broadly distributed in La Plata basin and high population structuration between Lower and Upper Paraná River regions.