Comparative population genetic structure of redbelly tilapia (Coptodon zillii (Gervais, 1848)) from three different aquatic habitats in Egypt

Abstract Recently, tilapia have become increasingly important in aquaculture and fisheries worldwide. They are one of the major protein sources in many African countries and are helping to combat malnutrition. Therefore, maintenance and conservation genetics of wild populations of tilapia are of great significance. In this study, we report the population genetic structure and genetic diversity of the redbelly tilapia (Coptodon zillii) in three different Egyptian aquatic environments: brackish (Lake Idku), marine (Al‐Max Bay), and freshwater (Lake Nasser). The habitat differences, environmental factors, and harvesting pressures are the main characteristics of the sampling sites. Three mitochondrial DNA markers (COI: cytochrome oxidase subunit I; the D‐loop; CYTB: cytochrome b) were used to assess population structure differences among the three populations. The population at Lake Nasser presented the highest genetic diversity (H d = 0.8116, H = 6), and the marine population of Al‐Max Bay the lowest (H d = 0.2391, H = 4) of the combined sequences. In addition, the phylogenetic haplotype network showed private haplotypes in each environmental habitat. Results presented here will be useful in aquaculture to introduce the appropriate broodstock for future aquaculture strategies of C. zillii. In addition, evidence of population structure may contribute to the management of tilapia fisheries in Egyptian waters.

The goal of this study was to assess population structure in C. zillii by comparing its genetic characteristics in three different environments. Three Egyptian aquatic habitats differ in fish harvesting pressure and habitat characteristics. For example, tilapia is not a target fish for fishermen in Al-Max Bay, but it is the main target in Lake Nasser and Lake Idku. In addition, the impact of anthropogenic activities (enclosed by large fishermen communities and fish farms) and sea levels affected the demographic population structure in Lake Idku and the north Delta lakes (Malm & Esmailian, 2012). However, Lake Nasser is surrounded by small fishing communities, and fishing pressures are influenced by factors such as high temperatures and fish transportation. Additionally, the Aswan High Dam isolates Lake Nasser from other Egyptian aquatic environments. Therefore, we expected some differences in population structure and genetic diversity among C. zillii. The results of this study are the first genetic study on the Egyptian populations of C. zillii based on three mtDNA markers (COI, D-loop, and CYTB), and these data will be useful to assess population structure, which is needed for aquaculture and fisheries management.

| DNA extraction and sequencing
Total genomic DNA was extracted from muscle tissue (~25 mg) Table 1 of denaturation at 94°C for 45 s, optimum annealing (Tm D-loop = 52°C and Tm COI and CYTB = 50°C) for 45 s, and extension at 72°C for 1 min and a last extension step at 72°C for 10 min. The amplification products were examined by electrophoresis (1.5% agarose gel). They were purified with Exonuclease I and Alkaline Phosphatase Shrimp (Takara, Japan) and were incubated at 37°C for 20 min, followed by deactivation at 83°C for 30 min. Cleaned PCR products were sequenced using an ABI 3730 Genetic Analyzer at Fasmac Co., Kanagawa, Japan.

Sequence
Haplotype sequence data from this study were deposited in GenBank under accession numbers KY465475 -KY465492. Distributions and frequencies of haplotypes for the studied locations (Lake Idku, Al-Max Bay, and Lake Nasser) in each genetic marker are shown in Table 1 and However, haplotype diversity and nucleotide diversity of the brackish water population in Lake Idku were higher than in the freshwater population based on COI (H d = 0.7048; π = 0.0025) and lower using other markers (Table 1). Among all populations, CYTB indicated significantly lower haplotype diversity (0.0000-0.2651) compared to other genes studied here (D-loop and COI). Interestingly, results based on CYTB showed low or monomorphic results of the genetic indices for C. zillii among all populations. This gene has some limitations as a phylogenetic marker, and it fails to resolve the lower genetic relationships in cichlid fishes because of its slow mutation rate (Farias, Orti, Sampaio, Schneider, & Meyer, 2001). Therefore, CYTB seems to be a poor marker to study the population genetics of C. zillii. Therefore, we focus primarily on results based on the D-loop and COI markers, which show variability consistent with previous studies (e.g., Wu & Yang, 2012 Phylogenetic networks of the haplotypes were identified using the median-joining algorithm for 8,7,3, and 17 haplotypes of D-loop, COI, CYTB, and joined sequences, respectively (Figures 3 and 4). The haplotype network had relatively star-like shape in two of the three markers with some unique haplotypes at the edges of each network (Figure 3a and b). The Hap_1 haplotype was dominant in all three sites. According to the D-loop haplotype network, the Hap_4 haplotype was dominant in Lake Idku and Al-Max Bay (Figure 3a). In addition, the haplotype network of the COI genetic marker presented fewer unique haplotypes (Hap_1, Hap_6, and Hap_7 among all sites ( Figure 3b). However, the lowest number of haplotypes was obtained using the CYTB genetic marker (Figure 3c). The joined sequences (i.e., using all markers combined) showed more distinction among the three sites, with 7, 2, and 4 unique haplotypes from Lake Idku, Al-Max Bay, and Lake Nasser, respectively ( Figure 4). Overall, the haplotypes networks according to two variable markers (D-loop and COI), marine, and brackish sites show very different haplotype distributions, and that in the freshwater site, almost all haplotypes are present. However, the freshwater site presents several private combined haplotypes, and none in common with the marine site (because they present very different D-loops).
Pairwise Φ ST values were estimated to assess the population genetic structure of C. zillii among the three aquatic environments (  (Table 2). In addition, the concatenated sequences showed significant Φ ST values among all populations as shown by D-loop and COI markers ( Table 2).
The mismatch distribution analysis for both mtDNA markers (COI and D-loop) revealed multimodal pattern among the observed distribution values (Appendix S1). The isolation by distance (

| DISCUSSION
We found clear genetic population structure in the redbelly tilapia, (freshwater, brackish, and marine) using three mtDNA markers (COI, D-loop, and CYTB). The genetic isolation of the freshwater population (Lake Nasser) from the brackish and marine populations might be due to anthropogenic factors such as the Aswan High Dam (Figure 2) and fishing activities, as well as habitat differences (Hassanien & Gilbey, 2005). In the Dead Sea system (the Kishon River and Jordan River), hypersaline waters have been shown to create biological barriers for C. zilli populations (Szitenberg et al., 2012). A similar pattern of habitat-driven genetic isolation was reported in Nile tilapia (Oreochromis niloticus) (Hassanien & Gilbey, 2005). As aquaculture and other anthropogenic activities continue to expand, it will be important for management entities to consider the impacts of these activities on genetic diversity and connectivity of target species.
The saltwater population (Al-Max Bay) presented the lowest ge- in Lake Idku (a semiclosed system). The demographics of Lake Idku are changing rapidly because the lake lost about 319.3 km 2 of surface area between years 1800 and 2010 (Shawer & Ibrahim, 2010). The reduced area (~27%) of the lake is mainly caused by human impacts in the surrounding area (Malm & Esmailian, 2012;Shawer & Ibrahim, 2010).
However, Lake Nasser (the second largest artificial lake in Africa) presented positive values of Fu's F S test, indicating overdominant selection or a recent bottleneck in its C. zillii population. It might be that the population of C. zillii in Lake Nasser has a stable overdominant equilibrium.
Additionally, two characteristics of Lake Nasser affect population structure of tilapia in all Egyptian waters: 1) the Aswan High Dam was built in 1970, creating a barrier to exchange between populations on either side of the dam, and 2) the presence of various dendritic inlets (Khors) or side projections of the main lake (Ryder & Henderson, 1975 Nile River has begun, and these are likely to further affect the diversity, distribution, and population structure of aquatic species in Lake Nasser and other Nile River populations. Therefore, the present study may serve as an important scientific resource for future studies to monitor the influence of existing and new dams. Such structures appear to have the ability to influence connectivity and subsequent genetic structure of the target populations. Our preliminary findings may support fisheries managers, for example, to understand the population structure and diversity of C. zillii in Nile waters. Loss of genetic diversity may lead to reduced resilience in the face of unforeseen potential ecological pressures, such as the emergence of diseases or susceptibility to pollutants. We hope that regional authorities will consider and/or monitor population diversity as part of adaptive management plans in this commercially important species.

ACKNOWLEDGMENTS
The authors would like to thank Dr. Steven D. Aird and the two anonymous reviewers for their suggestions and comments of this manuscript.
This study was funded by internal grants from the National Institute of Oceanography and Fisheries, Egypt ("Study of the Environment and Fisheries of Lake Nasser"); Okinawa Institute of Science and Technology Graduate University, Japan; and King Abdullah University of Science and Technology, Saudi Arabia

CONFLICT OF INTEREST
None declared. Taha