Kinship Analysis of Offspring of the Giant South American River Turtle (Podocnemis expansa) Using Microsatellite DNA Markers
Abstract
The giant South American river turtle (Podocnemis expansa) is the largest extant member of the suborder Pleurodira and the largest freshwater chelonian in South America. Owing to its size, its meat is sought for consumption and trade in the Amazon region. The aim of the present study was to investigate the reproductive behavior of 2 different populations of P. expansa. Allelic frequency variation was studied in 6 DNA microsatellite loci of recently hatched offspring from 8 nests on a beach in the municipality of Oriximiná in the state of Pará, Brazil. Multiple paternity was identified in all nests studied. Between 5 and 10 fathers contributed to each nest. Compared with previous studies, a greater frequency of multiple paternity and a greater number of fathers contributing to each nest were found in the present investigation. The results suggest that populations from different locations may exhibit different paternity patterns due to ecological or biological differences. By comparison with previous studies, we suggest that the differences in the number of nests, proportion of offspring per nest, and number of loci analyzed in each study may influence the frequency of multiple paternity detected.
The giant South American river turtle (Podocnemis expansa) is the largest member of the suborder Pleurodira and the largest freshwater chelonian in South America (Pritchard and Trebbau 1984). This turtle has a wide, smooth carapace with black and gray coloration (Vogt 2008) and occurs in the Amazon, Orinoco, and Essequibo river basins (Pritchard and Trebbau 1984). It has a relatively prolonged sexual maturity and the replacement of individuals in populations is low because of the high mortality rate of offspring (Pritchard and Trebbau 1984).
Due to its large size, which exceeds 90 kg in body weight and 80 cm in carapace length (Pritchard and Trebbau 1984), P. expansa is heavily exploited for food and trade for the consumption and direct sale of its meat, viscera, and eggs, or for the use of its carapace in craftwork (Cantarelli 2006). These factors contribute to the listing of P. expansa as Endangered (International Union for Conservation of Nature 2016). Although some aspects of the reproductive behavior of P. expansa have previously been described (Valenzuela 2000; Pearse et al. 2006b), studies on this species are lacking from throughout most of its range. Thus, there is a need to identify and understand the ecological factors linked to its reproductive behavior to provide information that can be used for the adequate management of the species.
We present comparative findings on 2 different populations of P. expansa using larger and more homogenous samples sizes among nests, compared with previous studies. The aim of the study was to investigate the reproductive behavior of P. expansa determined by the degree of kinship of recently emerged hatchlings from a nesting beach in the municipality of Oriximiná (state of Pará, northern Brazil), as inferred from hypervariable microsatellite loci.
METHODS
Blood samples were obtained from recently hatched specimens of P. expansa at Jacaú beach on the Trombetas River, the municipality of Oriximiná, the state of Pará, Brazil. A total of 221 individuals distributed among 8 nests were evaluated. Approximately 50 l of blood was collected from each individual through the venous puncture and stored in 500 μl of ethanol at 4°C (Avery and Vitt 1984). All individuals were then released at their original capture sites.
DNA extraction was performed using 2% CTAB (cetyl trimethylammonium bromide) extraction buffer (Doyle and Doyle 1987). Extraction efficacies and DNA concentrations were determined by electrophoresis in 1% agarose gel, which was examined under an ultraviolet light transilluminator. We followed the protocol described by Schuelke (2000) and used nested polymerase chain reaction (PCR) to amplify hypervariable regions for 6 pairs of oligonucleotides. Two of these were developed for Podocnemis unifilis (Fantin et al. 2007) and 4 others were developed for P. expansa (Sites et al. 1999; Valenzuela 2000; Table 1). Amplification reactions were adjusted to a total volume of 13.5 μl containing 1.5 μl of 10× (NH4)2SO4 buffer, 2.7 mM MgCl2, 0.2 μM reverse primer, 0.14 μM forward primer, 0.14 μM TET-6 M13 primer, 0.2 mM (each) dNTP, 0.4 U of Taq DNA polymerase, and approximately 100 ng of genomic DNA. The initial denaturation temperature was 94°C for 1 min; followed by 25 cycles at 94°C for 30 sec, 55°C for 30 sec, and 68°C for 30 sec; followed by 20 cycles at 94°C for 30 sec, 50°C for 30 sec, and 68°C for 30 sec; with a final extension at 72°C for 15 min.
Amplified DNA fragments were separated by electrophoresis on a 1% agarose gel, after which all PCR products were then diluted in distilled water in a ratio of 1:100, followed by the addition of the size marker ROX pUC-19, as modified from the protocol described by DeWoody et al. (2004). Genotyping was performed in an ABI 3130XL automatic DNA sequencer and analyses of allele frequencies for each locus were performed with the aid of the GeneMarker V2.2.0 program (Applied Biosystems).
Paternity estimates were based on the minimum method of allele counts (Myers and Zamudio 2004), which presupposes a Mendelian distribution of alleles in the offspring. However, this method does not estimate an actual number of contributing parents to a single nest; it is conservative because it cannot distinguish among parents sharing homologous alleles. Thus, multilocus combinatorial analyses were also performed using the program Kinalyzer (Berger-Wolf et al. 2007), to estimate relationships between true siblings and half-siblings of the same nest and to assign paternity and infer mating systems. The inference of families of true siblings is possible even without having knowledge of the parental genotype, as this program uses information on shared alleles at multiple loci among individuals from each nest.
RESULTS
Based on the minimum method allele counts, multiple paternity was found in all nests analyzed (Table 2). When it was not possible to infer parental genotypes, we inferred a minimum of 2 paternal contributions to the same nest when at least 5 alleles were found for the same locus among the offspring: 2 from the mother and 3 (or more) from at least 2 contributing fathers. When one of the maternal alleles could be identified, the occurrence of 4 different alleles of the same locus was considered indicative of multiple paternity. However, when both maternal and paternal alleles could be identified through the examination of homozygous offspring, the presence of 3 alleles segregating at one locus confirmed multiple paternity.
The number of alleles per locus ranged from 4 to 22, with locus 91 (Nest 1) being the least polymorphic and locus 1 (Nest 5) the most polymorphic. Nest 4 had the fewest paternal contributors (5 fathers) and Nest 5 had the most (10 fathers). Table 2 summarizes the number of alleles per nest, the inferred number of males that contributed to the offspring of each nest, and the combination of all loci used to infer sibling groups, as estimated using Kinalyzer.
DISCUSSION
Polyandry and multiple paternity have been found in a wide range of animal taxa, including mammals such as Mus musculus musculus (Thonhauser et al. 2014); insects of the order Hymenoptera (Jaffé et al. 2012); crocodilians such as Melanosuchus niger (Muniz et al. 2011), Crocodylus porosus (Lewis et al. 2013), and Caiman crocodilus (Oliveira et al. 2014); amphibians such as Chiromantis xerampelina (Byrne and Whiting 2008) and Crinia georgiana (Roberts et al. 1999); and birds, including Ficedula albicollis (Michl et al. 2002), Emberiza schoeniclus, and Malarus cyaneus (Griffith et al. 2002). In chelonians, high levels of extra-pair paternity have been found in multiple species: Chrysemys picta (Pearse et al. 2001, 2002), Caretta caretta (Moore and Ball 2002), Chelonia mydas (Lee and Hays 2004; Ekanayake et al. 2013), and Emys blandingii (Refsnider 2009), and in other species of the genus Podocnemis, namely P. unifilis (Fantin et al. 2008), Podocnemis sextuberculata (Fantin et al. 2008), and Podocnemis erythrocephala (Fantin et al. 2010).
Different hypotheses have been suggested to explain the polyandrous behavior exhibited by species representing different phyla of animals. Different costs and benefits are derived from polyandry, so many factors may contribute to the evolution of this type of reproduction. Researchers suggest that polyandry and multiple paternity increase the odds of fertilization, the acquisition of more competitive gametes, and the viability of the offspring through increased genetic diversity (FitzSimmons 1998; Jennions and Petrie 2000; Uller and Olsson 2008; Thonhauser et al. 2014).
According to Yasui (1997), the probability of acquiring genes that increase offspring fitness is higher when females are promiscuous (the “good genes” and “compatibility of genes” hypotheses). Myers and Zamudio (2004) suggest that polyandry could have evolved in environments in which optimal environmental conditions are only intermittently available, thus favoring promiscuous females. According to Slatyer et al. (2012), polyandry is favored by natural selection when mating with multiple males provides females with more nutrients contained in sperm. The same authors also suggest that such substances may increase in number and/or quality of eggs, which then provides an effective increase in the clutch size.
Multiple paternity within P. expansa.
The first evidence of multiple paternity in P. expansa was demonstrated by Valenzuela (2000) in nests on the Caquetá River in Colombia. Subsequently, Pearse et al. (2006b) found evidence multiple paternity in P. expansa nests on the Orinoco River in Venezuela (Table 3). Sample sizes between these studies likely account for the considerable differences in frequency of multiple paternity of P. expansa. Valenzuela (2000) studied 2 nests with 19 and 46 offspring and found multiple paternity in both, while Pearse et al. (2006a) analyzed 32 nests containing 9 to 76 offspring per nest and found multiple paternity in only 10%. These differences in frequency of multiple paternity might be explained by any combination of different factors, including perhaps the dramatic decline in nesting females on the single nesting beach sampled in Venezuela.
First, both nests studied by Valenzuela (2000) in the Rio Caqueta basin had moderate to large clutch sizes, while Pearse et al. (2006b) showed that in the Rio Orinoco basin, nests with smaller clutch sizes had single paternity and multiple paternity was documented only in the larger clutches. Low fecundity of many of the Orinoco female P. expansa may be an artifact of the low statistical power of the combination of loci used in this study; rare alleles would likely be missed.
A second factor that may contribute to the differences between these river basins would be impacts of human exploitation. Historically in the Orinoco basin, the numbers of P. expansa nesting females decreased drastically from 330,000 (Humboldt 1820) to 13,800 (Ojasti 1967). Although P. expansa nesting is protected by the Ministry of the Environment on a single beach on the Orinoco River with only ∼ 1000 females (Pearse et al. 2006b), Mogollones et al. (2010) affirmed that this turtle population seemed to be stable by the time captive rearing and reintroduction programs were initiated in 1989. Despite these conservation efforts, the earlier estimates of the severe population decline of nesting P. expansa on this single beach constitutes a severe bottleneck and if it was sufficient to eliminate low-frequency alleles, then many males would share only high-frequency alleles at most loci and the markers used by Pearse et al. (2006b) likely would not have sufficient resolution to identify low-frequency cases of multiple paternity.
Regarding estimates of the number of paternal contributors to P. expansa nests, Valenzuela (2000) found a minimum of 2 fathers in one nest and 3 in the other, with 19 and 46 hatchlings, respectively. Pearse et al. (2006b) found only 3 of 32 multiple-sired clutches, with a minimum of 3 fathers contributing to clutches of 20 hatchlings, 71 hatchlings, and 70 hatchlings. In our study, all clutches had multiple fathers, with a minimum contribution of 5 males in Nest 4 and a maximum of 10 males in Nest 5. In part, some of the differences among these 3 studies likely reflect differences in sampling efforts, including variation in 1) the number of nests sampled, 2) clutch size per nest, 3) number of loci sampled, and 4) number of alleles per locus. Some combination of these variables will influence paternity estimates; for example, in a demographic study of the North American turtle Chrysemys picta, Pearse et al. (2002) showed that the likelihood of detection of multiple paternity increased with the number of nests and number of offspring per nest sampled.
Among biological variables, genotypic diversity of males in a population will exert a direct influence on the calculation of the paternal contribution to offspring. When any particular locus with few alleles is included, different individuals within the breeding population in question are likely to sometimes share the same genotype, which can lead to an underestimation of the number of males that contribute to the production of offspring. This error can be minimized by sampling many nests, many hatchlings per nest, and more variable loci; all of these variables will improve estimates of genetic variability among the males within a breeding population, which would improve estimates of male contributions to single clutches.
Multiple paternity in the genus Podocnemis.
Paternity studies of different species of the genus Podocnemis also report the existence of polyandry. Fantin et al. (2008) analyzed the reproductive behavior of P. unifilis on a beach of the Amazon River (Brazil) and found that at least 2 males contributed to all of the clutches tested. These same authors also reported multiple paternity in P. sextuberculata nests on the Amazon River in the municipality of Barreirinha, Brazil; at least 3 males contributed to the offspring in a single nest, and at least 2 males sired the offspring of all other nests. Fantin et al. (2010) found that P. erythrocephala is also promiscuous, reporting multiple paternity in 83% of nests analyzed in 2 locations of the Brazilian Amazon. Single paternity was found in one nest of this species, demonstrating the possibility of differences in the paternity pattern among populations of the species (Table 3).
Our study underscores the importance of considering ecological differences among populations. The scarcity of resources is related to a reduction in population density and the number of copulations (Engqvist 2011). The higher the population density, the more likely the chances of encounters that could result in mating (Taylor et al. 2014). According to Crim et al. (2002), sex ratio is also an important factor to be considered, given that populations with male-biased sex ratios are more likely to exhibit multiple paternity.
In kinship analysis, we suggest that larger and more homogenous sample sizes among nests be considered in order to minimize among-locality ecological or biological variation when comparing different populations of P. expansa, given that such variation could influence the rate of multiple paternity found.
Contributor Notes
Handling Editor: Peter V. Lindeman
