Conservation Genetics of the Asian Giant Soft-Shelled Turtle (Pelochelys cantorii) with Novel Microsatellite Multiplexes

Simple Summary Pelochelys cantorii is critically endangered and rarely seen in the wild, and only 13 adults are being kept in captivity in China. For the purpose of reinforcing the conservation and management of P. cantorii, the Pearl River Fisheries Research Institute (Chinese Fishery Academy of Sciences) successfully bred more than 800 turtles from 2015 to 2020. In this study, we developed and characterized 10 simple sequence repeat markers from the RNA transcriptome of P. cantorii, established two multiplex PCR systems, and obtained the genetic structure and genetic diversity of the artificially conserved population. The aim was to obtain viable second-generation P. cantorii with the highest genetic diversity to implement population recovery plans for this species. Abstract To understand the genetic structure of the protected turtle species Pelochelys cantorii we used transcriptome data to design more than 30,000 tri- and tetranucleotide repeat microsatellite primer pairs, of which 230 pairs were used for laboratory experiments. After two screenings, only 10 microsatellite markers with good specificity, high amplification efficiency, and polymorphisms were obtained. Using the selected primers, two multiplex PCR systems were established to compare and analyze the genetic diversity of artificially assisted breeding generations from four parents (two females and two males) continuously bred over two years. A total of 25 alleles were detected among the 10 microsatellite loci of the offspring. The polymorphism information content (PIC) was 0.313–0.674, with an average of 0.401, among which two loci were highly polymorphic (PIC ≥ 0.5). The number of alleles was 2–5 and the number of effective alleles was 1.635–3.614. The observed heterozygosity was 0.341–0.813, with an average of 0.582, whereas the average expected heterozygosity was 0.389–0.725, with an average of 0.493. Moreover, on the basis of Nei’s genetic distance and the Bayesian clustering algorithm, the 182 offspring were divided into two subgroups, and the results corresponded to the two maternal lines. This is the first study to investigate the molecular markers of P. cantorii, and the results obtained can be used to protect genetic resources and provide a genetic basis for the design of population recovery plans.


Introduction
The Asian giant soft-shelled turtle (Pelochelys cantorii) belongs to the order Testudines (family: Trionychidae), one of the largest inland aquatic turtle species. It is an important indicator of ecological health in the Pearl River and the south of the Yangtze River in China. This species also has a long history and is of great scientific value in paleogeography and paleontological evolution. In the past, P. cantorii was widely distributed in southeastern China, but due to excessive economic development, its habitat has continuously deteriorated, and its population has been greatly reduced. Currently, P. cantorii is critically Animals 2022, 12, 3459 2 of 12 endangered and rarely seen in the wild, and only 13 adults are kept in captivity [1]. To strengthen its protection policy, China listed the turtle as a key aquatic wildlife protection animal at a national level in 1989. The World Conservation Union approved P. cantorii as an endangered species in 2000, and it was later added to a page in Appendix II of the Convention on International Trade in Endangered Species of Wild Fauna and Flora treaty in 2003.
For the purpose of reinforcing the conservation and management of P. cantorii, the Pearl River Fisheries Research Institute (Chinese Fishery Academy of Sciences) successfully bred more than 800 turtles from 2015 to 2020 using four sexually mature turtles (two females and two males). Twenty healthy P. cantorii, aged 45 years and weighing 1.5-2 kg, were selected for a rewilding adaptive protection test in 2020 [2]. Nonetheless, the genetic structure and diversity of the artificially conserved population are still unclear; therefore, scientific management is urgently needed to obtain viable second-generation P. cantorii with high genetic diversity and restore the wild population through artificial propagation and release.
In this study, we obtained transcriptomic data for P. cantorii. The main goal was to develop microsatellite multiplexes for P. cantorii to evaluate the genetic diversity and structure of artificially assisted breeding generations and to design the most feasible population recovery plan for this species.

Experimental Materials
The experimental materials were obtained from a P. cantorii breeding and protection base in Gaoming, China, which maintains four parents (two females and two males). There were a total of 182 offspring (103 individuals born in 2019 and 79 in 2020). All of them underwent artificial incubation, whereafter umbilical cords were collected after hatching. The umbilical cords were soaked in anhydrous ethanol and stored at −20 • C.
Permission for this work was obtained from the relevant staff at the Gaoming breeding and protection base. We only used the umbilical cord as experimental material, which can naturally fall off, to prevent influencing individual biological behaviors.

Genomic DNA Extraction
Umbilical cords were cut to a mung bean size (approximately 30 µg), and a MicroElute Genomic DNA Kit (OMEGA Bio-Tek, Inc., Norcross, GA, USA) was used for DNA extraction according to the manufacturer's specifications. The optical density (DNA absorbance ratio) and concentration of the extracted DNA were measured using a Nano Q microspectrophotometer (BoAo, USA). Then, 1% agar gel electrophoresis was performed to test DNA integrity. The DNA solutions were stored at −20 • C until use.

Design and Screening of Microsatellite Primers
Using transcriptome data produced in our laboratory (unpublished), tri-and tetranucleotide repeat microsatellite sequences were identified, and primers were designed using Primer Premier 5.0 [19]. Thereafter, eight turtle samples were randomly selected to preliminarily detect primer specificity and polymorphism at the optimum annealing temperature. The selected primers were combined to construct a multiplex PCR system. Notably, the lengths of the PCR amplification products of the same joint primers did not overlap. During primer synthesis, different joints (PET, VIC, and NED) were added at the 5 end of each forward primer. In addition, the synthesized PET, NED, and VIC sequences were labeled with red, black, and green fluorescence, respectively (Table 1). Before PCR amplification, all primers were diluted to 10 µmol/L and mixed at a 1:40 ratio for each pair of forward and reverse primers. All three fluorescent connectors were diluted to 20 µmol/L. The established 10 µL multiplex PCR system comprised the following: 5 µL Applied Biosystems Multiplex PCR Master Mix; 2 µL mixed primers (forward primer and reverse primer); 0.2 µL fluorescent connector, 1.8 µL deionized water; 30-150 µmol/L DNA; 1 µL. The amplification procedure was as follows: initial denaturation at 94 • C for 5 min; denaturation at 94 • C for 30 s, annealing at 60 • C for 45 s, and extension at 72 • C for 70 s (24 cycles); denaturation at 94 • C for 30 s, annealing at 53 • C for 40 s, and extension at 72 • C for 30 s (eight cycles); final extension at 72 • C for another 10 min. A 2 µL volume of PCR product was mixed with 8 µL of molecular weight marker (GeneScan 500 LIZ)™ and a deionized formamide mixture (1:100). After further incubation for 5 min at 95 • C, the mixture was cooled on an ice plate for rapid denaturation. Thereafter, capillary electrophoresis was performed using an Abi 3130 multifunctional genetic analyzer (Applied Biosystems, Waltham, MA, USA). Finally, Peak Scanner Software v1.00 was used for genotyping.

Data Analysis
POPGENE v1.32 [20] was used to calculate the number of alleles (N a ), effective alleles (A e ), observed heterozygosity (H o ), expected heterozygosity (H e ), and Shannon diversity index of the microsatellite loci. The polymorphism information content (PIC) of the microsatellite loci was calculated using CERVUS v3.0 [21]. Calculation of the inbreeding coefficient (F IS ) among turtles was performed using GenAlEx 6 [22]. POPGENE v1.32 was used to treat each sample as a population to calculate Nei's standard genetic distance for each turtle, and dichotomous difference clustering (evolutionary tree) was constructed with MEGA 5.0 [23] to obtain the classification relationship between each of them. Structure v2.3.4 [24] was used to simulate the number of subgroups, and the Bayesian clustering algorithm was used to calculate the clustering status and blood composition of each turtle. Mapping was obtained by CLUMPP; K values were selected from 1-7, and each value was repeated six times, preheated 50,000 times, discarded, and followed by 100,000 formal calculations. The package structure results (K = 1-7) were submitted to http://taylor0.biology.ucla.edu/structureHarvester/ (accessed on 6 January 2020) and concluded as the best K value.

DNA Extraction and Quality Control
DNA integrity was evaluated using 1% agarose gel electrophoresis. When the band pattern was clear and bright (Figure 1), DNA integrity was considered good, the original figure is shown in Figure S1. The 260/280 ratio of DNA ranged from 1.82-1.92, indicating that DNA integrity was good and the purity was high, which met the requirements of the subsequent experiments.

Design and Screening of Microsatellite Primers
More than 30,000 tri-and tetranucleotide repeat microsatellite primer pairs were designed. Trinucleotide repeat microsatellite primers comprised 67.73% of the total, and the number of (AAT)n sequence motifs was the largest, accounting for 6.00%. Tetranucleotide microsatellite primers consisted of 32.27% of the total, and (AAAG)n sequence motifs were the most numerous, accounting for 6.67%. Among these, 230 primer pairs were randomly selected. In total, 153 pairs were amplified to form stable bands after the first screening. Eight individual samples were randomly selected for detection, and 10 pairs were polymorphic, with a polymorphism ratio of 6.53%. Relevant information on the 10 microsatellite primer pairs is shown in Table 2.

Design and Screening of Microsatellite Primers
More than 30,000 tri-and tetranucleotide repeat microsatellite primer pairs were designed. Trinucleotide repeat microsatellite primers comprised 67.73% of the total, and the number of (AAT) n sequence motifs was the largest, accounting for 6.00%. Tetranucleotide microsatellite primers consisted of 32.27% of the total, and (AAAG) n sequence motifs were the most numerous, accounting for 6.67%. Among these, 230 primer pairs were randomly selected. In total, 153 pairs were amplified to form stable bands after the first screening. Eight individual samples were randomly selected for detection, and 10 pairs were polymorphic, with a polymorphism ratio of 6.53%. Relevant information on the 10 microsatellite primer pairs is shown in Table 2.

Construction of Multiplex PCR
By optimizing the parameters of multiplex PCR, two groups of microsatellite multiplex PCR systems were established, each containing five microsatellite loci. The specific parameters of the system are shown in Tables 3 and 4, and the partial electrophoresis patterns of each multiplex PCR group are shown in Figures 2 and 3. Table 3. First group of multiplex PCR primers.

Locus
Primer Sequence Joint Sequence Fluorescent Label

Genetic Relationship Analysis
The genetic distance between each individual was calculated using POPGENE. The individual clustering diagram based on the evolutionary tree method using MEGA showed that the entire progeny population could be divided into two subgroups (Figure 2). Beyond that, it was also concluded that when the subgroup value was 2, it could be measured using Structure software. Each individual was represented by a vertical bar partitioned into segments according to the proportion of genome belonging to each of the clusters identified (K = 2) using Structure (displayed from left to right in the order of 1-183). The number of subgroups was also the closest to reality, and the degree of individual hybrids is shown in Figures 4-6. Individuals numbered 1-103 were offspring from 2019 and 104-182 were offspring from 2020. partitioned into segments according to the proportion of genome belonging to e clusters identified (K = 2) using Structure (displayed from left to right in the o 183). The number of subgroups was also the closest to reality, and the degree of hybrids is shown in Figures 4-6. Individuals numbered 1-103 were offspring and 104-182 were offspring from 2020.      [25]. Twelve microsatellite markers obtained from an RNA transcriptome were constructed from blood cells of Tachypleus tridentatus, which could also be applied to the analysis of its genetic diversity [26]. According to Lindqvist et al. [27], microsatellite markers of tri-and tetranucleotide repeats are more suitable for the large-scale automatic analysis of fluorescent labels. Lu et al. [28] also showed that microsatellite markers with triand tetranucleotide repeats in C. carpio had more abundant polymorphisms. In vertebrate genomes, microsatellite tetranucleotides (GATA/AGAT) are the most common [29]; therefore, microsatellite enrichment of many species has subsequently been carried out using GATA/AGAT sequence motifs as a probe. In this study, more than 30,000 tri-and tetranucleotide repeat microsatellite primer pairs were selected and designed from the transcripts of P. cantorii. The AAAG (n) sequence motif was the most prevalent, accounting for 6.67% of the total, whereas GATA and AGAT sequence motifs only represented 1.92% and 1.57%, respectively. This may be related to the study design criteria and species specificity.

Genetic Structure of the Soft-Shelled Turtle Population
Compared with other testudines, the population of P. cantorii observed in this study had fewer alleles. For example, genetic differentiation of Geochelone nigra using 10 microsatellite loci showed 12-37 alleles in each microsatellite seat in various groups, with an average of 21.1 [30]. In addition, the number of alleles determined for Malaclemys terrapin ranged from 8-14 in five microsatellite loci, which was an approximate average of 10.67 [31]. The number of alleles (A) at a single locus in Mauremys mutica ranged from 5-26, with an average of 14.190 [18]. Schultz et al. (2009) believed that a PIC > 0.5 is high, 0.5 > PIC > 0.25 is moderate, and a PIC < 0.25 is low. In this experiment, the number of alleles detected at each locus was 2-5 [32]. The PIC was 0.313-0.674, of which two were highly polymorphic (PIC ≥ 0.5). The Shannon diversity index was 0.7549. The H o was similar to that of nine populations of T. tridentatus (0.46-0.57) and higher than that of the endangered Cuora flavomarginata (0.032-0.936), with an average value of 0.329 [33,34]. The data were also similar to the average H o (0.512-0.627) of 218 individuals from four Pelodiscus sinensis populations [16]. These results indicate that the genetic diversity of P. cantorii is moderate. However, it is worth noting that the number of alleles and available alleles, H o , H e , Shannon diversity index, and other data from 2019 were all higher than those from 2020. The reasons for this may be the following: (1) the number of umbilical cord samples used varied-in 2019 and 2020, there were 103 and 79 samples, respectively; (2) these experiments took samples from the same four parents (two females and two males), and every parent mating combination was haphazard; (3) genetic diversity analysis was performed only on offspring from 2019 and 2020, which is a short sampling period.
In this experiment, MEGA and Structure software were used to cluster the conserved population. The results showed that the population could be divided into two subgroups, which was consistent with the actual situation of producing offspring from only two maternal parents. MEGA's evolutionary tree results show that the population could be further divided into four subsets, consistent with the actual situation in which all four parents participate in reproductive activity and produce theoretical offspring groups (M 1 × F 1 , M 1 × F 2 , M 2 × F 1 , and M 2 × F 2 ). The fact that the four subgroups had different numbers of progeny may be due to male competition during mating, including female selection preference, sperm motility difference, fixed collocation, and other factors. This phenomenon indicates that there are dominant individuals among the parents, but the emergence of such individuals may cause problems, such as migration and a reduction in genetic diversity in the offspring population. For the breeding of second-generation offspring in future strategies, individuals of different subgroups can be selected according to clustering data, and gene exchange within the population can be artificially mediated to prevent inbreeding decline.

Conclusions
The breeding period we used to develop a conservation population of P. cantorii was short, and sexually mature individuals are very rare in China. The Pearl River Fisheries Research Institute (Chinese Fishery Academy of Sciences) collaborating with Gaoming's breeding center for P. cantorii has had no genotype supplementation from individuals of other regions in the short term. For all offspring derived from the four initial parents, germplasm resources are very limited, and the number of close relatives may increase in the future. Therefore, the introduction of sexually mature individuals from other regions may be an effective method to improve the level of genetic diversity, strengthen gene exchange among populations, and maintain the genetic potential of the population in prospective studies. At the same time, on the basis of molecular markers of parental genotype files and the genetic distances and relationships of the parents used to set up the scientific breeding program, we propose that it is possible to ensure that the offspring inherit the full genetic variation of the parents. This will allow them to produce offspring with rich genetic diversity and facilitate the supplementation of artificial breeding populations into the natural population.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available in this article and supplementary material.