Population genetics of three threatened catfish species in heterogeneous environments of the Cauca River, Colombia

Neotropical catfishes Ageneiosus pardalis, Pimelodus grosskopfii and Sorubim cuspicaudus are migratory fishes of commercial importance that exhibit decreasing populations due to overfishing and other anthropic interventions. This study used species-specific microsatellite loci to test the hypothesis that threatened fish populations show genetic vulnerability signs and are genetically structured in the middle and lower sections of the Cauca River. The studied species exhibit genetic diversity levels higher than the average values reported for Neotropical Siluriformes; however, they seem to have suffered recent bottlenecks and they present significant endogamy levels that are higher for the critically endangered catfish P. grosskopfii. Furthermore, both Ageneiosus pardalis and S. cuspicaudus are each formed by one genetic group, while Pimelodus grosskopfii comprises two coexisting genetic groups. The information obtained in this study is useful for the decision making in management plans that are appropriate for the sustainability of these three species populations within the proposal for the expansion of the hydroelectric development and other anthropic activities.

Contrary to the majority of population genetics studies in Amazonas and Paraná basins, very few have been performed in Colombian basins, most of which are found in grey literature (see Mancera-Rodríguez et al., 2013;. For the remaining freshwater catfishes in northwestern South America, there is still unawareness of genetic diversity, demography and population structure which limits the implementation of effective rules for their management and protection. The latter is important since populations of these species have been affected by overfishing and other indirect effects of anthropogenic activities like mining, dam construction, gravel sand extraction, basins contamination and continuous water extraction for agricultural and livestock purposes, which have led to the decay and decrease of the populations of said species in all Colombian basins (Galvis, Mojica, 2007;Usma-Oviedo et al., 2009;Barletta et al., 2010;Jiménez-Segura et al., 2016).
The appointed species are a good model for population genetics analysis since they are exposed to anthropic interventions such as loss of habitats, deforestation, river-course fragmentation, introduction of exotic fish species, contamination and overfishing (Barletta et al., 2010;FAO, 2015;Jiménez-Segura et al., 2016;AUNAP, 2020). The decreasing population density of the three species led to their classification as vulnerable in the Colombian red list of freshwater fishes (Mojica et al., 2012b). Additionally, due to this decline in populations, P. grosskopfii was included as a critically endangered threatened species in the red list of threatened species of the International Union for Conservation of Nature -IUCN . Thus, the a priori expectation was that the species showed high values of inbreeding and recent bottlenecks, as they are exposed to different anthropic interventions that might affect the density of their populations. 4/18 ni.bio.br | scielo.br/ni Moreover, the studied area comprises the middle and lower sections of the Cauca River, which includes near 500 km of the main channel of the river, 110 km of torrents and 40.000 ha of swamps. This area also includes the Cauca River canyon, the steepest margin of the Antioqueño Plateau in the northern portion of the Central Cordillera (Restrepo-Moreno et al., 2009), which has been considered a geographic barrier for many fish species (Dahl, 1971). Additionally, this area presents differences in habitats, water speeds, temperatures and it constitutes the influence zone of the Ituango hydroelectric project. In this context, this study also tested the hypothesis that populations of these species, collected before the beginning of the construction of the hydroelectric, were genetically structured. The a priori expectation was that the environmental heterogeneity of the zone caused genetic structuration. The answer to these queries generates information of interest to propose rules of management and conservation, suitable for the preservation of natural populations of these fish species.

MATERIAL AND METHODS
A total of 492 preserved in alcohol muscle tissues of A. pardalis (193), P. grosskopfii (170), and S. cuspicaudus (129) were analyzed. Said tissues were provided by Integral S.A through the scientific cooperation agreement CT-2013-002443, framed in the environmental license of Ministerio de Ambiente, Vivienda y Desarrollo Territorial # 0155 of January 30 th , 2009. Samples were collected between 2011-2014 in the middle (sections S1-S4) and lower (sections S5-S8) sections of the Cauca River, located upstream (S1) and downstream (S2-S8) of the construction zone of Ituango hydroelectric project ( Fig. 1; Tab. 1). The middle section of the Cauca River encompasses the Cauca River canyon, characterized by a steep topography with the presence of rapid zones (S1-S3) followed by a vast alluvial plain surrounded by mountains, with swamp ecosystems connected with the main river channel (S4). The section S5 includes the mouth of the Nechí River, a hydrologically monomodal river prone to flooding and scouring, and the remaining lower section (S6-S8) shows soft topography with flat and wavy surfaces wherein several swamps connected to rivers through canals are formed, conforming complex swampy systems (Mejía-Rivera et al., 2007;Betancur-Vargas et al., 2009).
To determine the number of possible genetic groups for A. pardalis, P. grosskopfii, and S. cuspicaudus in the middle and lower basins of the Cauca River, it was used different complementary approaches. The Bayesian clustering method with STRUCTURE v2.3.4 software (Pritchard et al., 2000) used the following conditions: 100000 MCM, 10000 burn-in, admixture model, correlated allele frequency, and LOCPRIOR option. Simulations were repeated 20 times per each number of evaluated probable populations: 1-6 for A. pardalis, 1-8 for S. cuspicaudus, and 1-9 for P. grosskopfii. The most probable number of population was calculated using the ΔK method (Evanno et al., 2005)

RESULTS
The sampling performed by Integral S.A. between 2011-2014 revealed that A. pardalis, P. grosskopfii, and S. cuspicaudus were differentially distributed in the middle and lower sections of the Cauca River. Namely, and despite having the same migration range, P. grosskopfii was found in all sampled sites, including upstream of the Cauca River canyon, whereas S. cuspicaudus was found in the downstream area of Cauca River canyon and the construction zone of the dam (S3-S8). In contrast, short-distance migration range A. pardalis was only captured in the lower basin of the river (S5-S8; Tab. 1). 7/18 ni.bio.br | scielo.br/ni Genetic diversity and population demographics. Pimelodus grosskopfii showed similar or nearly similar genetic diversity values in all evaluated sections (Na=10.308-13.308; He=0.849-0.866; Ho=0.607-0.682) and in the genetic groups proposed by the genetic structure analysis described below (Na=12.846-16.615; He=0.858-0.866; Ho=0.614-0.664; Tab. 2), further to statistical significance in heterozygosity deficit and inbreeding coefficients (F IS =0.108-0.218; P<0.000; Tab. 2).
In accordance with the other two species, A. pardalis showed similar genetic diversity values (Na=10.850-12.550; He=0.832-0.839; Ho=0.791-0.822; Tab. 2); statistical significance in heterozygosity deficits and coefficients of inbreeding in all the evaluated samples (F IS =0.026-0.086; P<0.007) although in lesser magnitude than in P. grosskopfii (Tab. 2  The analysis for detecting bottleneck showed significant values for all the species in the evaluated sections under the IAM, while no significance was found under the SMM and the TPM showed significance in all or almost all of the evaluated sections in A. pardalis, P. grosskopfii (except S8), and S. cuspicaudus (except S4, S6, and S7/8). Additionally, the three species showed values of the M ratio of Garza, Williamson, (2001) lower than 0.680 (P. grosskopfii: 0.197-0.227; S. cuspicaudus: 0.229-0.258; A. pardalis: 0.242-0.245; Tab. 2), which has been proposed as limit value for inferring that the population has suffered a recent decline in population (Garza, Williamson, 2001).

DISCUSSION
This study employed species-specific microsatellite loci to acknowledge the level of genetic diversity, demographic status and genetic structure degree of three species of Colombian endemic Neotropical catfishes that coexist in the middle and lower sections of the Cauca River basin. The a priori expectation was that the species showed high values of inbreeding and recent bottlenecks as they are subject of different anthropic interventions that might affect the density of their populations. Furthermore, it was expected a population genetic structure concordant with some heterogeneous features of the basin.
In S. cuspicaudus, genotypic frequencies were distributed according to Hardy-Weinberg equilibrium and the inbreeding values were not significant save for sections S5 and S8 in which values lower than 10% suggest a careful management to prevent adverse effects regarding fitness (Frankham et al., 2014). On the contrary, and despite their high genetic diversity, A. pardalis and P. grosskopfii exhibited a heterozygosity deficit that cannot be explained by the Wahlund effect in the only genetic group formed by A. pardalis nor in each of the two genetic groups of P. grosskopfii that co-migrate in the studied area, wherein the heterozygosity deficit stayed significant.
Unlike P. grosskopfii, which has external reproduction (Loir et al., 1989), the assortative mating can explain the heterozygosity deficit in A. pardalis since it has internal insemination and elaborated courtships (Galvis et al., 1997) and the females store spermatozoa inside their ovaries (Contreras et al., 2012). However, to verify this hypothesis, it remains necessary to further research the reproductive behavior of these species.
Alternatively, the heterozygosity deficit could be explained by inbreeding as A. pardalis (2.600-8.600%) and P. grosskopfii (10.800-21.800%) showed significant F IS values. This effect may result from different anthropic activities like overfishing due to diminishments in captures since 1994 (Jiménez-Segura,Villa-Navarro, 2011) and their presence in the first 10 species captured for the Magdalena-Cauca basin (AUNAP, 2020). Other factor that could negatively contribute to the studied catfish species is the contamination in basins caused by mining as in the studied zone there is gold, silver, and platinum extraction (UPME, 2019), and there is evidence of mercury bioaccumulation, a residual contaminant of this activity, in several species that cohabit in the zone (Marrugo-Negrete et al., 2008;Álvarez et al., 2012).
All those factors above mentioned may explain the recent decline in populations suggested by bottlenecks in the three studied species, supporting the hypothesis of genetic vulnerability of fishes due to anthropic interventions. A non-excluding alternative considers the climate factor as an additional influence of the demographic aspects of the populations of these species. Extreme changes in the Cauca River flow as consequences of the climatic phenomenon El niño-Southern Oscillation (ENSO) presented in the region in the prior years, wherein since 2007 there has been extreme NIÑA (2007-2008and 2010-2011) and NIÑO (2009-2010, Hoyos et al., 2013 phenomena, could have caused bottlenecks as described in the same region for Curimata mivartii (Landínez-11/18 ni.bio.br | scielo.br/ni García, Marquez, 2018). This climate factor could contribute to bottlenecks phenomena associated to overfishing, which has been described for other Neotropical catfishes as P. corruscans, P. reticulatum, and B. flavicans (Coronel et al., 2004;Prado et al., 2018).
Genetic structure analysis. This study did not find significant differences among individuals of middle and lower sections of the Cauca River, which indicates that the environmental heterogeneity of the basin and, in particular, the steep slope of the Cauca River canyon is not a barrier for the gene flow of P. grosskopfii. However, such slope seems to limit of dispersion of S. cuspicaudus and A. pardalis in the middle section of the Cauca River since no individuals of these species were captured in the higher sections of the studied area in the sampled period (2011)(2012)(2013)(2014). This result is not surprising for a short migrator like A. pardalis; nonetheless, the spatial distribution differed in the middle range migrators S. cuspicaudus and P. grosskopfii, suggesting that besides from their migration range, other factors as body conformation, habitat preferences and behavior, can play an important role in the observed spatial distribution, which is an important approach for future researches. A strong limitation to spatial distribution by the Cauca River canyon is also observed in Pseudopimelodus atricaudus, a new species described in the lower section of the Cauca River that diverged about 16 mya from its congener P. magnus distributed upstream the Cauca River canyon Restrepo-Gómez et al., 2020).
In the spatial distribution analyzed for each species, the Bayesian analysis of STRUCTURE, the DAPC, the AMOVA and the pairwise distances showed absence of geographic population structure in the studied species. These results concord with the ones described for other Neotropical catfishes that perform medium and large distance migrations as P. maculatus (Almeida et al., 2001(Almeida et al., , 2003Ribolli et al., 2012), B. flavicans (Coronel et al., 2004), B. rousseauxii (Batista, Alves-Gomes, 2006, P. corruscans and P. reticulatum (Vaini et al., 2016;Prado et al., 2018). Nevertheless, our results with A. pardalis and S. cuspicaudus contrast with the results for P. corruscans in the basin of rivers Paraguay and Paraná wherein the authors hypothesize a possible homing behavior for the species Dantas et al., 2013).
Distinct from A. pardalis and S. cuspicaudus, each formed by one genetic group, P. grosskopfii represented two coexisting genetic groups in the evaluated sections of the Cauca River. This result in P. grosskopfii could reflect a possible spatial reproductive isolation, considering the preference for reproductive sites (homing) described for other species of Pimelodidae family as B. rousseauxii (Batista, Alves-Gomes, 2006), P. corruscans Dantas et al., 2013), and P. reticulatum (Abreu et al., 2009). Another non-excluding alternative is that P. grosskopfii experiences temporary reproductive isolation as the Cauca River presents two hydrological periods of shallow and deep waters that could benefit two reproductive periods of the fish species that habit this basin (Jiménez-Segura et al., 2010;López-Casas et al., 2016). This explanation is supported in the observation of a greater number of individuals of P. grosskopfii participating in the migration during the first period compared to the second hydrological period of the Magdalena River (López-Casas et al., 2016). Despite the latter, to the date, there is no detailed information as per the reproductive periods and behaviors of P. grosskopfii that allows to elucidate the origin of the genetic groups.
Under this context, it is recommended that the information obtained in this study 12/18 ni.bio.br | scielo.br/ni regarding the structure and diversity of three coexisting species of catfishes in the middle and lower sections of the Cauca River is considered within the expansion proposals of the hydroelectric development for Colombia (Angarita et al., 2018) and of other anthropic activities. The higher values of the inbreeding coefficient were found for P. grosskopfii, included in 2016 as a critical danger species in the red list of the International Union for Conservation of Nature -IUCN , for which it is necessary to enforce the proposed measures for the recovery and conservation of natural populations of the species. Despite the inbreeding values found for S. cuspicaudus and A. pardalis are lower than 10%, which is the suggested limit to avoid substantial losses in the reproductive features (Soulé, 1980), it is advisable to continue monitoring these populations to evaluate the changes of the species over time since it has been indicated that inbreeding values higher than cero (F IS >0) can cause negative effects in the fitness of natural populations (Frankham et al., 2014).
In conclusion, employing species-specific microsatellite loci, this study determined the diversity and possible genetic structure in the middle and lower sections of the Cauca River for the Neotropical catfishes A. pardalis, P. grosskopfii and S. cuspicaudus. The genetic diversity levels of the three species (expected heterozygosity and number of alleles/locus) are high; however, all the studied species seem to have suffered recent bottlenecks and exhibit significant inbreeding values, which are greater for P. grosskopfii, currently in a higher risk category. Furthermore, P. grosskopfii shows the coexistence of two genetic groups while A. pardalis and S. cuspicaudus are mainly represented by a genetic group. Contrary to the expectations, none of the evaluated species showed population structure related to some heterogeneous features of the basin. It was also evidenced that the cannon of the Cauca River in this zone does not represent a barrier to the gene flow of P. grosskopfii but it seems to limit the dispersion of A. pardalis and S. cuspicaudus.