Genetic diversity in Cichla piquiti and cross-amplification for Cichla kelberi in the Serra da Mesa reservoir, Goiás, Brazil

. Species Cichla piquiti and Cichla kelberi are found in the Serra da Mesa reservoir, Goiás and are sedentary with diurnal habits. This study aimed to evaluate the magnitude and distribution of genetic variability in subpopulations of C. piquiti with specific microsatellite loci and to test transferability in other microsatellite markers for C. kelberi . We analyzed 99 individuals of C. piquiti from seven points to evaluate genetic diversity and structure with 10 microsatellite loci. Transferability of 75 loci was tested in C. kelberi to increase microsatellite markers available. Genetic structure was assessed with Bayesian clustering. Global F ST for C. piquiti was weak (0.056), but F IS (0.598) and F IT (0.621) were significantly high, indicating that the mating system has a strong influence on the organization of genetic variability with most mating among related. Two genetic groups were evidenced with most individuals allocated to a single group. Transferability of microsatellite loci for C. kelberi had low polymorphism. The level of genetic diversity was low, increasing inbreeding and suggesting that few individuals of C. piquiti colonized the reservoir during its installation due founder effect. Other factors as reproductive behavior and overfishing can act to decrease genetic diversity. Therefore, we reinforce the need for genetic monitoring to avoid loss of genetic diversity that can be intensified both construction of hydropower plants and ecological and reproductive aspects in some fish species.


Introduction
The richest fish biodiversity in the world is located in Brazil (estimated at 4,756) of which 3,512 are freshwater fish, representing 10% of the diversity of vertebrates in Brazil (Burgess, 2004;Froese & Pauly, 2022). However, more than 1,170 species are at risk of extinction, and of those, 26% are freshwater fish from the Actinopterygii class.
Thus, predicting and possibly mitigating the consequences of anthropogenic changes on aquatic systems requires knowledge of biological, ecological, and genetic dynamics of these species. Mainly regarding the maintenance of the diversity of the species genetic for the long-term preservation of these aquatic populations (Hutchings, 2000;Santamaria & Mendez, 2012;Herden et al., 2020).
The genus Cichla Block and Schneider, 1801 represents one of the main groups of piscivorous fish of the Cichlidae family in South America (Goldstein, 1973;Willis et al., 2015). These species are popularly known in Brazil as peacock basses (Tucunaré) and come from the Amazon and the Tocantins-Araguaia River basins and are widely introduced in the basins of the Paraná, Paraguai, Paraíba do Sul and Paraguaçú rivers (Gasques, Fabrin, Gonçalves, Prioli, & Prioli, 2014). Unlike most Amazonian fish, peacock basses are sedentary and diurnal, and they chase prey until successful capture. They show high dispersion across the national territory, i.e., tropical and subtropical regions, making it interesting for sport fishing (attracted by their aggressiveness), and extensive and semiextensive fish farming (Gomeiro, Villares-Junior, & Naous, 2009). Several cases of Cichla species introduced in the southeastern region have been described in the literature, based on the evaluation of mtDNA regions, showing concern about the impact they may cause for the native ichthyological community (Santos, Salgueiro, Franco, Marques, & Nóbrega, 2016;Diamante et al., 2017).
Due to the aggressive and territorial behaviors of Cichla species, mainly in the reproductive season, natural populations of lakes and rivers may present genetic differentiation, despite connections between them Acta Scientiarum. Biological Sciences, v. 44, e58294, 2022 water volume, with 54.4 billion m3. The reservoir comprises mainly the municipalities of Minaçu, Campinorte, Uruaçu, Barro Alto and Niquelândia.
Sampling was performed (collection permit SISBIO -16594) during August 2009 and December 2010, totaling 99 individuals of C. piquiti and 21 individuals of C. kelberi. The study area was divided into seven collection points (Table 1) located in four tributary rivers in the Serra da Mesa reservoir: Bagagem, Maranhão, and Traíras Rivers, in the municipalities of Niquelânida, Colinas do Sul, and Uruaçu ( Figure 1). The specimen collection was performed using nets with 10 different mesh sizes : 2, 4, 3, 4, 5, 6, 7, 8, 10, 12, and 16 cm. The nets were placed both during the day and at night and reviewed every 12 hours. ] Figure 1. Sampling sites of Cichla piquiti and genetic clusters in Serra da Mesa reservoir, Goiás. Seven points were selected for collecting 99 individuals of this species. Genetic clusters with Bayesian analysis are visualized in each circle, in which the color represent assignment to genetic cluster (K1 is orange and K2 is blue).

DNA extraction and amplification
Muscle tissue samples of C. piquiti and C. kelberi were collected at the bottom of the dorsal fin and kept in 96% ethanol for the conservation of the material and DNA extraction followed the protocol by Taggart, Hynes, Prodöuhl, and Ferguson (1992).
For genetic diversity of C. piquiti, we initially tested ten microsatellite markers developed specifically for the species, but just six loci were polymorphic for 99 individuals of C. piquiti (Table 2) (Carvalho et al., 2009).
The DNA of all individuals was amplified via PCR from the protocol with a final volume of 15 μL, using 5 ng of DNA template, 10X buffer (10 mM Tris-HCl, pH 8.3), 50 mM magnesium chloride (KCl), 250 μM of each dNTP, 250 μg BSA, 0.9 μM of each, and 1 U Taq DNA polymerase (Phoneutria, BR). The amplification reactions were performed in a Veriti 96-Well Thermal Cycler thermocycler (Applied Biosystems, CA). Genotyping for both species was performed using vertical electrophoresis on 6% polyacrylamide gel with a 10 bp DNA ladder marker (Invitrogen TM ), and the detection of amplified products using silver nitrate (Creste, Tulmann, & Figueira, 2001). Table 2. Genetic diversity of the six microsatellite loci (Carvalho et al., 2009) analyzed in Cichla piquiti from the Serra da Mesa reservoir. Annealing temperature (TA), number of alleles per population (A), (AR) allelic richness, expected heterozygosity (HE), observed heterozygosity (HO), probability of identity (I) and paternity exclusion (Q).

Genetic characterization and structure
The following genetic parameters were estimated to characterize the genetic variability of C. piquiti: number of alleles (A), allelic richness (A R ), observed (H O ) and expected (H E ) heterozygosity, and fixation index (f). These genetic estimates were evaluated for each locus and per population. We performed randomizationbased tests with Bonferroni correction to verify deviations in the Hardy-Weinberg equilibrium (Goudet, Raymond, Meeus, & Rousser, 1996). All of these analyses were performed using the FSTAT 2.9.3.2 software (Goudet, 2002).
Moreover, the power of individual discrimination of loci was estimated based on the probabilities of genetic identity (I), which is the probability of two random individuals of a population having identical genotypes (Paetkau, Calvert, Stirling, & Strobeck, 1995) and paternity exclusion (Q), which is the probability of excluding false paternity (Weir, 1996). These estimates were obtained for each polymorphic locus using the IDENTITY 4.0 software (Wagner & Sefc, 1999).
The magnitude and distribution of population genetic variability was evaluated just in C. piquiti using the F IT , F ST , and F IS coefficients (Wright, 1951) obtained by the analysis of variance of allele frequencies (Weir & Cockerham, 1984), carried out in the FSTAT 2.9.3.2 software (Goudet et al., 1996). We used a Bayesian approach to infer existing genetic clusters, considering the genetic structure of the subpopulations sampled within the Serra da Mesa reservoir. For this, we used the STRUCTURE 2.3.4 software, assuming the assumptions of correlated allele frequencies and admixture (Pritchard, Stephens, & Donnelly, 2000). The number of K varied from 1 to 7, with 10,000 burn-in iterations and 100,000 steps in the Markov chain Monte Carlo (MCMC), with ten replications for each K. The choice of the best K was carried out on the STRUCTURE HARVESTER platform using the method by Evanno, Regnaut, and Goudet (2005). The coancestry coefficients for each group and individuals were calculated in the CLUMPP 1.1.2 software (Jakobsson & Rosenberg, 2007), and the cluster visualization was generated in the DISTRUCT 1.1 software (Rosenberg, 2004).
Spatial patterns of genetic variability were investigated using the simple Mantel test, correlating the geographic and genetic distance matrices. The linearized pairwise F ST matrix (F ST /(1-F ST ), calculated in the Arlequin 3.1 software, and the geographic distance (log) were used for this analysis (Rousset, 1997). This analysis was performed using the 'mantel' function of the 'vegan' package (Oksanen et al., 2019)
The PCR protocol and verification of the amplified product used vertical electrophoresis on 6% polyacrylamide following the above conditions described for C. piquiti. The annealing temperatures of the 75 loci evaluated varied between 44 and 62ºC. The genetic characterization of the transferred loci occurred by analyzing the genetic parameters of expected and observed heterozygosity, and by analysis of genetic discrimination of loci.

Results
The genetic diversity found for C. piquiti was low because among the 10 microsatellite loci evaluated, four of them were monomorphic (Tuc04, Tuc11, Tuc12, and Tuc18). Thus, we removed these four monomorphic loci from the analysis, once they were not good evaluators of the local population genetic diversity of C. piquiti. The average observed and expected heterozygosity within the six remaining polymorphic loci was 0.073 and 0.192, respectively (Table 2). We found 24 alleles ranging from 1.3 to 3 in the evaluated populations (Table).
Six subpopulations had significant deviations between observed and expected heterozygosity (Table 1), leading to high inbreeding values. This occurred in the population of Point 6 that was more isolated and next to Traíras river, where f value was equal to 1, that is, reproduction occurs exclusively among related individuals in this population. The characterization of the loci is shown in Table 2, where the loci evaluated in the populations of the Serra da Mesa reservoir exhibit low genetic variability.
The analysis of genetic structure showed values of F ST = 0.056, F IS = 0.598, and F IT = 0.621, all significant with p < 0.05. This result indicates that the reproduction mechanisms of C. piquiti have strong influence on the organization of genetic variation. The high F IT value suggests that mating among C. piquiti individuals in the Serra da Mesa reservoir does not occur in a panmictic way. The F ST value, despite low, was significant, suggesting a slight genetic structure among the subpopulations within the lake.
The Bayesian analysis identified two clusters with better K = 2 (K1 is orange, and K2 is blue; Figure 1. However, when observing the coancestry coefficient values (Q), we noticed that all populations are grouped in cluster 2 (blue), while the population of Point 3 obtained higher levels of admixture (Q < 0.8, see Table 3). The subpopulation of Point 3 obtained higher values of genetic diversity for both expected heterozygosity and allelic richness. The low genetic structure found with F ST does not reflect the influence of geographic space since the Mantel test between genetic and geographical distances was not significant (-0.384; p > 0.05). The cross-amplification tests were not enough to identify suitable loci that permit to characterize genetic diversity of C. kelberi. The transferability of microsatellites for this species had low amplification success (10%) because eight of the 75 loci tested (Tuc 4, Tuc 10, Tuc 11, Tuc 16, Hmal_59, Pre 26, Ast 4, and Pcos 14; see Table 4 and Figure 2) were transferred to C. kelberi genome, using 21 individuals collected in the Serra da Mesa reservoir. Of these eight loci, only the Tuc 16 was polymorphic, with three alleles in 21 individuals. The Hmal 59 locus had two alleles, yet all individuals were heterozygous for the same genotype. Consequently, we did not find microsatellite loci set suitable for population genetic analyses with C. kelberi.

Discussion
The subpopulations of C. piquiti from the Serra da Mesa reservoir showed low genetic variability with high intrapopulation inbreeding (F IS = 0.598). The low values of expected and observed heterozygosity reflect the low genetic diversity in populations of artificial reservoirs. This is confirmed by the monomorphism in four microsatellite loci for C. piquiti, that the loci Tuc 4, Tuc 11 and Tuc 12 had just two alleles in populations of Amazonian basin with low observed and expected heterozygosity (Carvalho et al., 2009).
The genetic differentiation among populations was low, yet the weak genetic structure found, considering that the individuals were sampled within the same reservoir, indicates that the mating system and territorial behavior of the species can influence the organization of population genetic variability in the Serra da Mesa reservoir. As reported by Gouskov, Reyes, Wirthner-Bitterlin, & Vorburger (2016) for the species Squalius Acta Scientiarum. Biological Sciences, v. 44, e58294, 2022 cephalus Linnaeus the construction of hydroelectric plants affects the connectivity between different populations and consequently genetic structure.
Another major issue is the behavior of freshwater fish species in reservoirs, where C. piquiti is known to be a voracious, territorial, and sedentary predator species, which increases the mating chances among related individuals and leads to an interpopulation structuring (Gomeiro et al., 2009). However, interpopulation genetic differentiation of C. piquiti and C. kelberi, evaluated between native and introduced river populations, indicates high pairwise F ST values, suggesting that different tributaries become independent genetic stocks (Carvalho, Oliveira, Sampaio, & Beheregaray, 2014). In this same study, genetic diversity was low for native and introduced populations in both species, indicating that bottleneck events act on the genetic structure of species.
The low genetic diversity of C. piquiti in the Serra da Mesa reservoir can be related with the colonization of the reservoir with only a few individuals of C. piquiti, contributing to increasing the high total inbreeding (F IT = 0.621). The high inbreeding rates observed in the subpopulations of C. piquiti may be related to the founder effect, an ecological and genetic process resulting from the recent formation of the Serra da Mesa reservoir (< 30 years old), even though C. piquiti appears naturally in the Tocantins-Araguaia River basin. This process recruits few founders of a new population, causing the fixation of some alleles, and consequently, the low genetic variability caused by inbreeding (Freitas & Galetti Jr, 2005;Aho, Rönn, & Piironen, 2006), reducing the population genetic differentiation. The drastic change in local landscapes affects all types of organisms, causing the loss of countless species. The few surviving specimens will recolonize the environment, yet causing an increase in inbreeding (Aho, et al., 2006) and increases the risk of local extinction (Wright, Tregenza, & Hosken, 2008), although the specie is not classified as threatened. Founder effect in populations of reservoir reaches not just fish species but zooplankton community due to rapid growth in modified environment (Haileselasie, Mergeay, Vanoverbeke, Orsini, & De Meester, 2017).
The Bayesian analysis identified two groups (K= 2), but we emphasize that most populations are found to group 2 (in blue) and only one population had sufficient levels of admixture (Q < 0.8). This result is related to the low genetic differentiation among the populations evaluated (F ST = 0.056) and confirms the existence of total inbreeding since all populations have a similar arrangement of allele frequencies. Genetic differentiation is not sufficient to separate the groups, and it is not affected by the spatial component, because there was no correlation between genetic and geographical distances.
The absence of spatial heterogeneity was a significant result because the habits of the species could influence the organization of genetic variability across geographic space. This can be justified by the collections in fairly close points (maximum distance among populations was 100 km), in which the geographical distance does not constitute a barrier to interpopulation gene flow. Moreover, intrapopulation and total inbreeding showed that the reproductive behavior of this species influences the organization of genetic variability because of the higher chance of genetic similarity over the analyzed geographical extent, due to the high degree of kinship.
The evolutionary dynamics of these subpopulations within the Serra da Mesa reservoir led to low levels of individual genetic discrimination and genetic linkage from the battery of loci in C. piquiti. The probability of paternity exclusion (Q) is dependent on the number and frequency distribution of alleles in the population, thus reduced levels of polymorphism or allele fixation for some markers cause low values of this parameter. In contrast, the probability of combined identity (I) must be practically zero to demonstrate that the microsatellite is an excellent marker for discriminating individuals (Collevatti, Brondani, & Grattapaglia, 1999). However, the set of loci analyzed showed high I value (0.010) and low Q value (0.436). In addition, fish dispersal strategies and life-history traits contribute to the effects of genetic variation (Pilger, Gido, Propst, Whitney, & Turner, 2017) and for C. piquiti, there is naturally low gene flow, due to its territorial and sedentary behavior (Gasques et al., 2014).
The species C. kelberi appears in the same geographic area of C. kelberi in the Serra da Mesa reservoir. Comparative genetic studies would be important to describe genetic diversity between two species. However, the transferability of microsatellites from different fish species to C. kelberi was not satisfactory because of the low success in amplifying the evaluated loci and low polymorphism. Transferability tests for C. kelberi were performed by Carvalho et al. (2009), but were not polymorphic in individuals of Serra da Mesa. In addition, optimized microsatellite loci of others species of fish in this study were not satisfactory to characterize the population diversity of this species in the Serra da Mesa reservoir. Even analyzing few individuals, this result demonstrates the need to develop specific microsatellite primers for C. kelberi since the transferred loci are not sufficient to evaluate the genetic variability of this species.
The transferability of microsatellite loci among evolutionarily close species is quite common, seeking to use universal microsatellites (Barbará et al., 2007;Vieira et al., 2016) in genetic studies of population, that have shown positive results with fish, reducing expenses related to the development of species-specific loci (Barbosa, Corrêa, Galzerani, Galleti, & Hatanaka, 2006. Positive cross-amplification among taxonomically related groups is possible due to the conservation of microsatellite flanking regions. Rico, Rico, and Hewitt (1996) had success in cross-amplification in different fish species belonging to the taxonomic groups Agnatha, Chondrichthyes, and Osteichthyes, where they found conservation of microsatellite flanking regions since the ancestral group with 470 million years. In our study, the conservation of microsatellite flanking regions may explain the transferability at a distant taxonomic level. However, the low polymorphism for the evaluated regions can be justified mainly by the analysis of individuals from the Serra da Mesa reservoir, providing low genetic diversity.
Assessing the genetic structure and the low genetic diversity in C. piquiti and high monomorphism in the transferred loci in C. kelberi highlights the existence of mechanisms leading to the homozygosity of individuals for these loci, which requires further studies since this is a native species of Serra da Mesa flooding region (Gasques et al., 2014). Besides, information regarding the genetic variability of these species, degree of inbreeding, or genetic linkage is essential for establishing management actions for fishing and even control of the populations of C. kelberi and C. piquiti. Besides, these and other fish species experience negative impacts from this enterprise such as changes from lotic to lentic environments, changes in limnological characteristics, and fragmentation of the environment, which constitutes a physical barrier to ecological and genetic processes (Oliveira, Castro, & Baptista, 2008;Almeida-Ferreira, Oliveira, Prioli, & Prioli, 2011).
Fish species encounter loss of historical genetic diversity. Thus, the formation of artificial lakes without an adequate design can aggravate this ichthyofauna peculiarity (Osborne, Perkin, Gido, & Turner, 2015). This study shows the need to evaluate the structure of fish assemblages in environments modified by hydropower plants both before and after its establishment, because they become barriers that can lead to a decrease in genetic diversity, especially for species with low movement habits (Gouskov et al., 2016).

Conclusion
Our study contributes to understanding the genetic effects on fish populations from hydropower plants, and artificial lakes. The C. piquiti sedentary behavior was not sufficient for genetic differentiation between populations, probably because first individuals were from same population, leading to an increase in inbreeding and low genetic differentiation. The microsatellite loci were transferred to C. kelberi, but there was low polymorphism for the analyzed individuals. The high rates of inbreeding observed, are probably related to the mating preference among related individuals, the founder effect and territorial behaviors, which may have been accentuated during barriers imposed by the lake formation.