Temporal changes in gene expression and genotype frequency of the ornithine decarboxylase gene in native silverside Basilichthys microlepidotus: Impact of wastewater reduction due to implementation of public policies

Abstract Human activity has caused a deterioration in the health and population size of riverine species; thus, public policies have been implemented to mitigate the anthropogenic impacts of water use, watercourse transformation, and pollution. We studied the Maipo River Basin, one of the most polluted with untreated wastewater in Chile, for a period of 12 years (2007‐2019). Since the implementation of new public policies, including the operation of a wastewater collector (2012), the Maipo River Basin is currently much less polluted by untreated water than before. To analyze the impact of wastewater reduction in this river basin, we studied the native silverside (Basilichthys microlepidotus), which inhabits both polluted and unpolluted areas of the river basin. Previous studies reported the overexpression of the ornithine decarboxylase (odc) gene, heterozygote deficit, and high frequency of a homozygote odc genotype in silverside populations that inhabit wastewater‐polluted sites, suggesting a phenotypic change and genotypic selection in response to pollution. Here, a population affected and another population unaffected by wastewater were studied before and after implementing the wastewater collector. The physicochemical data of water samples, changes in odc expression and microsatellite variability, and odc genotype frequencies were analyzed. The results showed physicochemical changes in the affected site before and after the operation of the wastewater collector. The microsatellite loci showed no changes in either population. The odc expression in the affected site was higher before the operation of the wastewater collector. Significant changes in the genotype frequencies of the odc gene before and after the wastewater collector operation were detected only at the affected site, wherein the homozygous dominant genotype decreased from >59% to <25%. Our results suggest that public policies aimed at mitigating aquatic pollution can indirectly affect both gene expression and genotype frequencies of important functional genes.

Although most studies have focused on the evidence of impacts caused by human activity (e.g., Jones & Stanley, 2016), there are also studies reporting the positive effects of changes in environmental policies. One well-known positive example is the policy to remove reservoirs, allowing the reestablishment of natural watercourses and free movement of minerals and nutrients from the mountains to the sea (Bednarek, 2001). Furthermore, policies related to the reestablishment of a population (Hitt, Eyler, & Wofford, 2012) and recolonization of the fish species upstream of the reservoir have been reported (Catalano & Bozek, 2007). In the case of pollution, returning the systems to their original state is a major challenge for environmental policies. In some rivers, an increase in species richness has been described (Brown et al., 2011;Northington & Hershey, 2006); however, in some cases, the expected reduction in the pollution level was negligible (Schiff & Macbroom, 2011). To our knowledge, no studies have been performed to determine the relationship between pollution reduction and allele or genotype frequency changes in freshwater fish populations, which can be further associated with phenotype and gene expression that ultimately contribute to an organism's fitness (Agashe, Martinez-Gomez, Drummond, & Marx, 2013;Goudet & Keller, 2002).
During the last decade, we monitored the Maipo River Basin, which is one of the most polluted river basins in Chile. According to the 2017 census, 7 million people (approximately 40% of the Chilean population) live close to this basin. Because of the organic matter present in untreated sewage generated by the growing number of human population and industrial facilities, the Maipo River has experienced water quality deterioration and eutrophication (Pardo, Vila, & Bustamante, 2008;Vega-Retter, Muñoz-Rojas, Vila, Copaja, & Veliz, 2014). Furthermore, the Maipo River Basin has the largest number of factories in the country and mining operations in the Andes Mountains, which affect the river's flora and fauna.
A previous study on fish diversity in this river basin has revealed a significant reduction in species richness and abundance over the last 30 years (Muñoz, 2007). The Mapocho River, a tributary of the Maipo River that crosses the city of Santiago, has been receiving untreated wastewater since the city was established. At the end of the 19th century, the pollution in the Mapocho River increased because of human migration from the countryside to the city and incipient industrialization (Castillo-Fernández, 2018 (Veliz et al., 2012). It is a microphagous species that feeds on insect larvae, small invertebrates, filamentous algae, and detritus (Bahamondes, Soto, & Vila, 1979;Duarte, Feito, Jara, Moreno, & Orellana, 1971). Basilichthys microlepidotus has a 1-year generation time, with a reproductive period ranging from August to January (Comte & Vila, 1992). This species is considered vulnerable based on its conservation status (Ministerio de Medio Ambiente Chile, 2018). The B. microlepidotus populations inhabiting polluted and unpolluted sites in the Maipo River Basin have been described previously . In particular, an RNA-seq analysis revealed that the expression of three cancer-related genes in fish inhabiting polluted areas was higher than those inhabiting unpolluted areas (Vega-Retter et al., 2018). One of these genes, ornithine decarboxylase (odc), which is related to tumor development and progression, showed synonymous mutations and evidence of change in genotype frequencies, such as heterozygote deficit and increased frequency of a homozygote genotype, in polluted areas. Further,  reported that pollution did not affect its neutral genetic diversity (microsatellite loci), while Vega-Retter, Vila, and Véliz (2015) hypothesized that pollution is potentially implicated in genotype selection. It is important to note that these results were based on the samples analyzed before the construction of the wastewater collector (Vega-Retter et al., 2018). Therefore, this system is a suitable model to determine the effects of pollution reduction on genotype selection.
Thus, owing to the significance of the odc gene in species health and substantial changes in genotype frequencies, we hypothesized that odc expression and genotype frequency in silverside may be affected by the operation of the wastewater collector. In this study, In addition, the variability of eight microsatellite loci used as the control loci in the same sites was used to describe possible changes in neutral loci variability over time.

| Physicochemical characterization of water from sampling sites
To assess the pollution level at each sampling site, we determined the physical and chemical characteristics of the water samples collected before (2007 and 2011) and after (2016 and 2017) policy changes in the Maipo River Basin. Water physicochemical parameters, including electrical conductivity (EC), pH, and total dissolved solids (TDS), were measured in situ three to six times at each site on each sampling date using a multiparameter device (Hanna Instruments).
To determine the dissolved oxygen (DO) concentration, one to six water samples collected from each site in each sampling year were taken in 200-mL polycarbonate bottles (Nalgene), fixed with manganese sulfate and alkaline iodide, and analyzed using the Winkler method, according to Strickland and Parsons (1968

| Change in odc expression over time
To quantify the expression of the odc gene, the RNA-sequencing The raw data from the samples were assembled de novo using the Mira Assembler (Cheveruex, Wetter, & Suhai, 2009 Raw reads with a mean Phred score of <Q30, ambiguous bases >10%, and reads < 50 bp were removed using the software programs Trim Galore (https://www.bioin forma tics.babra ham.ac.uk/ projects/trim_galore/), Cutadapt (Martin, 2011), and prinseq-lite.
pl script (http://prins eq.sourc eforge.net/manual.html). The de novo assembly of the reads was performed using the Bridger software (Chang et al., 2015) with the default parameter settings: k-mer length = 25, seed k-mer minimum coverage = 2, and seed k-mer minimum entropy = 1.5. In this assembly, the detection of isoforms Subsecretaria de Pesca). Genomic DNA was isolated using the salt extraction method (Aljanabi & Martinez, 1997) and then used for the amplification of the odc gene and microsatellite loci.

| Amplification of microsatellite loci
Eight microsatellite loci were used as a control neutral loci, including four (Odon02, Odon07, Odon09, and Odon39) described by Beheregaray and Sunnucks (2000) for Odontesthes perugiae and four permutations was used to analyze the temporal changes in allelic frequencies.

| Reciprocal gene flow between sampling sites
To estimate the contemporary reciprocal migration between the MEL and SFM sites in each sampling year, we used the microsatellite loci and the program BAYESASS 3.0 (Wilson & Rannala, 2003) with a burn-in period of 3,000,000 iterations, after burn-in of 30,000,000 iterations with sampling at each 100 iterations. The mixing parameters for allele frequencies, migration rates, and inbreeding coefficients were defined as 0.5, 0.3, and 0.7, respectively. Five independent runs were conducted, starting with different random seeds, and the results were expressed in terms of the average value of these five independent runs.

| Changes in the odc genotype frequency over time
To

| Temporal changes in the physicochemical features of water
The first PCA performed with the data from both MEL and SFM sites

| Changes in odc expression over time
The the same year was significantly higher before (18.7) than after (3.37) the collector was operational.

| Microsatellites as control loci and gene flow
The summary statistics for the microsatellite loci studied are shown in Appendix S3. Two monomorphic loci (Odon59 and Odon39) were excluded from further analysis. The results showed no significant departures from HWE at any sampling site or in any sampling year.
Furthermore, the F ST index showed no significant difference between the sampling years at each site (Table 1). The current reciprocal migration estimated between the affected (MEL) and unaffected (SFM) sites for different sampling years showed self-recruitment values >74% with migrants between sites ranging from 13% to 26% (Table 2). samples were observed between different sampling years (Table 3).

| Changes in odc variability over time
In contrast, significant differences were observed in all MEL samples (p < .05), except for the comparison between the years before  (Figure 6). In contrast, no significant differences in fitness were observed between the odc genotypes in the unaffected SFM sampling sites.

| D ISCUSS I ON
The main objective of this study was to correlate the odc expression and genotype frequencies of silversides to the reduction in pollution due to changes in public policies implemented to mitigate water pollution in the Maipo River Basin. The results showed apparent changes in the odc expression and genotype frequencies of silversides after the operation of the wastewater collector.
Field studies performed in natural environments have shown that domestic and industrial pollutants affect the natural populations. Domestic wastewater primarily contains microorganisms, degradable organic materials (feces, detergents, oil, and solvents), nutrients, and metals (Hence & Comeau, 2008), while industrial wastewater mainly contains pulp, paper, textiles, chemicals, pharmaceuticals, petroleum, tannery substances, and toxic synthetic and natural substances (Nath, Chakraborty, & Bhattacharjee, 2014).
In the case of the Maipo River Basin, we have no data on the changes in specific pollutants before and after the operation of the wastewater collector. However, EC and DO were among the important variables that explained the differences observed before and after the wastewater collector operation. EC is useful as a general measure of source water quality (Spellman, 2014). A natural system tends to have a relatively constant range of EC; thus, changes in this variable correlate with changes in water quality. Significant changes in conductivity may indicate the likely presence of a discharge or another source of pollution (Spellman, 2014). DO has a negatively proportional relationship with pollution, such that as the DO values decrease, the pollution levels increase (e.g., Sánchez et al., 2007).
Overall, we acknowledge that our chemical analyses were limited to only nine elements and do not necessarily represent all types of pollutants; thus, the physicochemical data were used as supporting evidence to understand the changes at the affected site over time.
Previous studies have attempted to associate wastewater with health in both humans and wildlife, suggesting an increased risk of cancer in wastewater treatment workers when compared to national risk averages (Hansen, Hilden, Klausen, & Rosdahl, 2003;Lafleur & Vena, 1991;Lagorio, de Santis, & Comba, 1996;Massaquoi et al., 2015). The effects of wastewater on the physiology and health of freshwater fishes have also been described. Liney et al. (2006) TA F I G U R E 5 Genotype frequencies observed in silversides collected during different sampling dates from the (a) affected (MEL) and (b) the unaffected (SFM) sites. Bars differing in letters represent statistical difference (p < .05), and for panel (b), nonsignificant differences were detected exposed the juveniles of Rutilus to wastewater, which resulted in male feminization, alteration of the gonads, and alterations in the tubular diameter of the hepatic ducts, among others. In general, water contamination has also been related to the presence of some neoplasms (Black, 1983;Lauren et al., 2010;Sonstegard, 1977) and papillomas in fish (Grizzle, Melius, & Strength, 1984).
Despite the significance of pollution as one of the primary factors affecting the freshwater diversity (Vörösmarty et al., 2010), studies correlating the selection of specific alleles or genotypes to pollution are limited. For example, the selection of CYP1A alleles has been described for Fundulus heteroclitus (Williams & Oleksiak, 2011) and Poecilia reticulata (Hamilton, Rolshausen, Uren Webster, & Tyler, 2017) inhabiting contaminated areas. Williams and Oleksiak (2011) detected CYP1A  In particular, ornithine decarboxylase is the first rate-limiting enzyme in polyamine biosynthesis that catalyzes the decarboxylation of l-ornithine to putrescine. The polyamine biosynthetic pathway is a critical regulator of cell growth (Jhingran et al., 2008) and plays an important role in neoplastic transformation and tumor growth (Pegg, 1988). Furthermore, its enzymatic activity has been used as a biological marker for evaluating tumor growth and aggressiveness (Deng et al., 2008). Sato et al. (2018) demonstrated that odc plays an important role in the reduction of methylmercury toxicity in mice, while Zahn et al. (1981) reported high odc activity and polyamine production in sponges exposed to benzo[α]pyrene. However, field and laboratory studies have shown contrasting odc expressions in the presence of pollutants. High level of odc activity was observed in fetal and newborn rat brain exposed to hypoxic periods and the  (Oleksiak, 2008). Taking into account the biological importance of odc and considerable reduction in its expression at the same location after wastewater reduction, it is important to follow its expression progression over time using real-time PCR (e.g., Rojas-Hernandez, Véliz, & Vega-Retter, 2019). On the other hand, it is important to note that our study was a field work in which the factors affecting odc expression could not be treated separately. Thus, further studies in controlled conditions, where the factor of interest is isolated from other factors, such as temperature water regime, parasites, etc., are required. In addition, it is necessary to perform expression analyses of each genotype under different pollution conditions to establish the complete relationship between selection and gene expression.
The results of our study suggest that the implementation of public policies enhanced the water quality of the Maipo River Basin, resulting in a significant increase in the habitat quality of native silversides. However, our findings do not imply that the Maipo River Basin has returned to its original state because only one gene that responds to a specific type of pollutant was studied; thus, a specific study of fish health and population size is needed to confirm that the changes detected here also evidence a recovery of the populations.

ACK N OWLED G EM ENTS
Authors thanks C Quezada-Romegialli, S Scott, R Gauci, M Briones, J Cortes, and P Acuña for fieldwork. Thanks to P Muñoz-Rojas for laboratory assistance. Authors thanks financial support of Millennium Nucleus of Ecology and Sustainable Management of Oceanic Islands (ESMOI), CONICYT PIA Apoyo CCTE AFB170008, Fondecyt 11150213 and CONICYT-FONDEQUIP EQM150077.

CO N FLI C T O F I NTE R E S T
None declared.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon request. and Pierre Duchesne (third unofficial advisor). I learned with them that it is necessary to work hard to achieve our goals, an important lesson for both personal and academic growth. The second phrase was during the summer sampling when the lobster war exploded in one of the sampling sites in New Brunswick. This was another big history never completely told. Many thanks Louis and feliz cumpleaños!