Isolation and Characterization of 32 Microsatellite Markers in Hermann's Tortoise, Testudo hermanni (Testudinidae)
Abstract
We used 454-sequencing to identify and characterize 32 microsatellite loci for the Western Hermann's tortoise (Testudo hermanni hermanni) for which primer sequences, characteristics, and multiplexes are provided. These markers have been tested for one continental (Var) and one insular (Corsica) French population as well as for the Eastern Hermann's tortoise (Testudo hermanni boettgeri). For these 3 populations, 24 loci were polymorphic. These microsatellite markers will be particularly useful to assess the genetic diversity and characteristics of the different Hermann's tortoise populations and will have direct applications for its protection, notably with respect to reintroduction and conservation programs.
The Hermann's tortoise (Testudo hermanni) is distributed in the Mediterranean area and is represented by 2 subspecies: the Eastern (T. h. boettgeri) and the Western (T. h. hermanni) Hermann's tortoise (Fig. 1). The former is widely distributed in the Balkan Peninsula, whereas T. h. hermanni is highly fragmented and found in continental (Spain, France, Italy) and insular (Balearic Islands, Corsica, Sardinia, Sicily) populations (Cheylan 2001). Testudo h. hermanni is particularly threatened by several factors, including reduction of suitable habitat, wildfires, unauthorized translocations, and anthropogenic pressures such as agriculture practices, urbanization, and illegal collection (Cheylan 1984; Bertolero et al. 2011; Couturier et al. 2014). These threats are particularly affecting continental populations, whereas island populations appear less affected (Zenboudji et al. 2016).
Citation: Chelonian Conservation and Biology 17, 2; 10.2744/CCB-1257.1
The distribution area of T. hermanni has been reduced from a continuous Mediterranean fringe to a few patches (Fig. 1), mainly because of climatic changes since the end of the Paleolithic (Morales Pérez and Sanchis Serra 2009), then accentuated by anthropogenic pressure since the Holocene. The Hermann's tortoise numbers have declined greatly, more particularly in continental Western European countries (Spain and France), and it is classified as near threatened by the International Union for Conservation of Nature (van Dijk et al. 2004). Consequently, conservation programs in different European countries have created several nature reserves within the frame of the Natura 2000 network and established translocation projects devoted to the protection of this emblematic reptile. In France, programs such as the European program LIFE+ Nature (2010–2014; Celse et al. 2014) or the Plan National d'Actions (2009–2014; Cheylan et al. 2009), recently renewed for the next 10 yrs (2017–2026; Celse et al. 2016), are devoted to conserving the last native French continental (Var) and insular (Corsica) populations of Hermann's tortoise.
Most continental populations are recognized as native, whereas the question is more debated for insular populations (Perez et al. 2013; Zenboudji et al. 2016). Moreover, the Hermann's tortoise has been translocated for conservation purposes (e.g., 7 official projects in Spain [Bertolero 2014; Bertolero et al. 2018]), but also uncontrolled releases took place (e.g., Cutuli et al. 2013; Soler et al. 2013). There is a long history of humans using the Hermann's tortoise as a pet, thus leading to an illegal trafficking where the impacts on survival and genetic mixing have been difficult to evaluate.
Genetic tools are particularly effective to reveal the characteristics and functioning of populations, and thus, to answer crucial questions about species conservation (Zenboudji et al. 2016). In the era of next-generation sequencing and avalanche of molecular data (such as single nucleotide polymorphisms), microsatellites still appear as useful markers because they can easily be used for analyzing new samples and also because they represent a time- and cost-effective method for estimating molecular diversity. However, until now only 7 microsatellite loci were described for the Hermann's tortoise (6 loci in Forlani et al. [2005] and 1 locus in Farke et al. [2015]), which is a limited number to provide statistical significance and answer specific questions regarding genetic variability and population structure. In the present study, we report 32 new microsatellite markers obtained for the Western Hermann's tortoise by next-generation sequencing, on which 15 loci have already been used to analyze the genetic diversity of 8 Hermann's tortoise populations (Zenboudji et al. 2016).
Materials and Methods: Obtaining Microsatellites
Microsatellite markers were isolated by a method coupling microsatellite enrichment and next-generation sequencing (Malausa et al. 2011). In order to guarantee defining species and not individual-specific primers, total genomic DNA from several tortoises were pooled. Fifteen individuals originating from 5 different sampling sites (3 samples from each population of Var, France, as well as from Albera, Ebro Delta, and South and North Minorca, Spain) were used. Our work received the official authorizations from the regional governments of Catalonia and the Balearic Islands (Generalitat of Catalonia and the Govern de les Illes Balears, permits SF/009, SF/304 and 65/2008) as well as from the French Direction Départementale des Territoires et de la Mer (arrêté préfectoral 83-2011/82). Concerning export and import permits, the Hermann's tortoise is registered on the list 2 and 3 of the Convention on International Trade in Endangered Species of Wild Fauna and Flora, thus allowing intracommunity trades between the different country members of the European Union, such as between Spain and France.
For each tortoise, 200 μl of blood was collected by a subcarapacial venipuncture from the nape of the neck (which does not necessitate the head to be in extension) or coccygeal vein, as described in Hernandez-Divers et al. (2002), and immediately preserved in 96% ethanol or in a saline solution (Queen buffer). As recommended in Beaupre et al. (2004), no anesthetic was used to minimize disturbance and handling time, and the puncture wound was cleaned using an antiseptic solution before releasing the tortoise.
DNA extractions were performed from blood samples preserved in ethanol or in Queen Buffer using the DNeasy Blood and Tissue Kit (Qiagen). Our final DNA sample (containing ∼ 1.8 mg DNA) was sent to the biotechnology company Genoscreen (Lille, France) that produced sequences by high-throughput sequencing on 454 GS-FLX Titanium platform. Before sequencing, genomic DNA was fragmented and enriched in 8 types of microsatellite motifs (TG, TC, AAC, AAG, AGG, ACG, ACAT, ACTC). A total of 29,174 raw sequences were obtained, among which 147 sequences have been validated as containing a microsatellite motif (100 di-, 28 tri-, and 19 tetra-nucleotide motifs). From Genoscreen, we received a selection of potential primer pairs designed for each of the 147 microsatellite loci, from which we selected 45 loci (15 di-, 18 tri-, and 12 tetra-nucleotide motifs) to assess for amplification and polymorphism.
Validation of Microsatellites and Multiplex Design
Each locus was tested for amplification by polymerase chain reaction (PCR) using the best primer pairs selected by Genoscreen, and PCR products were visualized on a 2% agarose gel. At this step loci that gave only one band were kept, whereas loci that produced multiple bands despite different increases of the annealing temperature were eliminated. PCRs were conducted in a 10-μl reaction volume containing 2 μl of H2O, 5 μl of the Master Mix (2×) solution from Qiagen, and 1 μl of each primer at 2 μM and 1 μl of DNA. All microsatellite amplifications were run using the following conditions: an initial denaturation step of 15 min at 95°C followed by 30 cycles of denaturation (30 sec at 94°C), annealing (1 min 30 sec at 52°–64°C; see Table 1), elongation (1 min at 72°C), and ended with a final elongation step of 30 min at 60°C.
Multiplexes were then built based on annealing temperature. Each forward primer was labeled with 1 of the 4 specific fluorescent dyes (FAM, VIC, NED, PET; ThermoFisher Scientific, Life Technologies SAS, Courtaboeuf, France) chosen according to the size of each locus of the multiplex in minimizing overlap between markers. Multiplex PCRs were conducted in a 10-μl reaction volume containing 3 μl of H2O, 5 μl of the Master Mix solution from Qiagen, 1 μl of a primer mix (each primer at 2 μM), and 1 μl of DNA using the same conditions as previously described.
PCR products were genotyped on a 16-capillary sequencer (3130 xl Genetic Analyzer, Applied Biosystems) at the “Genotypage-Séquençage” platform of the LaBex CeMEB (the Centre Méditerranéen Environnement Biodiversité, Montpellier, France). Electrophoregrams were visualized using GeneMapper v4.5 (Applied Biosystems) and 2 independent readings were performed by 2 people in order to minimize genotyping errors.
Data Analysis
Multiplexed microsatellites were tested for 2 Testudo hermanni hermanni populations from France (Var and Corsica) as well as for 15 samples (unknown origin) of T. h. boettgeri. For each locus, Micro-Checker v.2.2.3 (Van Oosterhout et al. 2004) was used to check for the presence of null alleles in each population. Several genetic diversity indices were calculated with Genalex v6.4 (Peakall and Smouse 2006): the number of alleles (Na), the observed heterozygosity (Ho), the expected heterozygosity (He) under the Hardy-Weinberg equilibrium (HWE), the inbreeding coefficient (FIS), and probability of deviation from HWE.
Results and Discussion
From the initial 45 selected microsatellites, 13 loci were eliminated because amplification was not specific no matter what the PCR conditions (several bands were generally observed). Among the 32 resulting loci, 10 multiplexes have been defined (Table 1), each including between 2 and 4 loci. The multiplex number 10 here includes only 1 locus (Ther93) because it is usually amplified with 2 other microsatellites already published (Test10 and Test56; Forlani et al. 2005; see Zenboudji et al. 2016). For each locus, 5 individuals of T. h. hermanni were tested and the electrophoregrams between uni- and multiplexes were compared for genotype consistency. For 10 loci, these comparisons revealed systematic genotype inconsistency (unreadable or multiple peak profiles); and for these loci, PCRs were then performed for each locus independently before the genotyping step. In Table 1 each locus is noted “pre-PCR” when the multiplex is performed before the PCR (mix of different loci in one PCR) or “post-PCR” when the multiplex is constituted after the PCR step just before genotyping, thus representing 22 and 10 loci, respectively. For each locus, DNA sequence from which primers were designed has been deposited in GenBank (Accession No. MG204067 to MG204098; see Table 1).
The 32 loci in 10 multiplexes were then analyzed for 2 French populations of T. h. hermanni—1 continental (Var, 85 individuals) and 1 insular (Corsica, 73 samples)—as well as tested for cross-amplification with T. h. boettgeri (15 individuals). Among the 32 loci, 8 have been detected as monomorphic for all samples of the 3 populations tested: Ther13, Ther26, Ther36, Ther38, Ther45, Ther49, Ther55, Ther82. However, 4 of them (Ther13, Ther26, Ther36, Ther55) were detected polymorphic for some individuals not included in the present study. High genetic divergence was previously detected between T. h. hermanni and T. h. boettgeri, as well as between populations of T. h. hermanni (Perez et al. 2013; Zenboudji et al. 2016); therefore, we cannot exclude the possibility that the 4 monomorphic loci might be polymorphic for other populations of both subspecies. Consequently, all 32 loci were maintained for publication (Table 1).
For the 3 populations analyzed, genetic diversity indices of the 24 polymorphic loci are given in Table 2. Micro-Checker detected the presence of null alleles for the 2 populations of T. h. hermanni: 8 loci (Ther17, Ther20, Ther23, Ther40, Ther51, Ther69, Ther74, Ther112) for the Var, 2 loci (Ther106, Ther112) for Corsica, as well as 3 loci (Ther42, Ther57, Ther112) for T. h. boettgeri. This number can be considered as surprisingly large given that the microsatellites have been defined specifically for T. h. hermanni. However, Micro-Checker detected the presence of null alleles on the basis of an excess of homozygotes (Van Oosterhout et al. 2004), which is also the signature of other genetic occurrences (lack of admixture, low population numbers, population bottlenecks, etc.). Var and Corsica have been shown to be genetically heterogeneous and made of several genetic groups (see Zenboudji et al. 2016), which might explain the presence of null alleles. Moreover, with the exception of Ther112, no other locus showing signs of null alleles is common to the 3 groups tested. Thus we feel confident that these loci can be used for subsequent analysis, although we advise to check again for null allele in each local population. The number of alleles detected varies from 1 to 21 for T. h. hermanni and from 2 to 16 for T. h. boettgeri. Two loci (Ther48 and Ther93 for T. h. hermanni, and Ther74 and Ther93 for T. h. boettgeri) are characterized by a large number of alleles. For the same 17 loci analyzed (see Table 2), the number of polymorphic loci is smaller for the insular Corsican population (12 loci) compared with the two other ones for which all loci appeared polymorphic. However, the mean number of alleles (over all 24 loci) is comparable: 4.6 for the Var, 4.2 for Corsica, and 4.7 for T. h. boettgeri. Observed (Ho) and expected (He) heterozygosities are similar between the 3 populations tested (see Table 2) and ranged from 0 to 1 and from 0.012 to 0.92, respectively. Deviations from HWE are mainly observed in the Var population (10 loci), whereas 4 and 6 loci are detected as significantly deviating from HWE for Corsica and T. h. boettgeri, respectively (Table 2). As expected, there is a strong correlation between these deviating loci and those detected with null alleles by Micro-Checker.
Our study developed 32 new microsatellites for the Hermann's tortoise, among which 24 loci are polymorphic for both subspecies (T. h. hermanni and T. h. boettgeri). The 8 remaining loci are totally functional and potentially polymorphic in other populations (in particular only 15 individuals were tested for T. h. boettgeri). A previous study (Zenboudji et al. 2016) analyzed 15 of the 32 microsatellites published (Ther4, Ther17, Ther20, Ther23, Ther40, Ther48, Ther51, Ther57, Ther69, Ther73, Ther94, Ther101, Ther106, Ther110, Ther112) on 357 individuals from 30 localities, representing a sampling nearly exhaustive (only the Majorca island was missing) of the entire range of the subspecies. This study revealed a strong genetic structure between geographic areas (Albera, Ebro Delta, Minorca, Var, Corsica, Continental Italy, Sardinia–Sicily; see Fig. 1), each of them being characterized by only 1 unique genetic lineage (with the exception of Minorca, for which 2 lineages have been identified). The 32 microsatellite markers characterized will be useful to provide answers to essential questions concerning evolution, characterization, and conservation of the Hermann's tortoise, such as the following.
Evaluation of the extent of hybridization between T. h. hermanni and T. h. boettgeri. Signs of admixture between the 2 subspecies have been evidenced for continental Italian and French populations, whereas insular ones appear rather preserved (Zenboudji et al. 2016). This study needs to be completed at a finer scale in the different populations.
Estimation of the genetic integrity of T. h. hermanni populations. As previously stated, the Hermann's tortoise has been the subject of human-mediated translocations. This practice continues today and the consequences on the genetic mixing between populations, as well as between wild and domestic tortoises, still need to be estimated.
Identification of individuals for translocation programs. Each geographical population is genetically unique; therefore, translocations should be done with individuals genetically identified as similar. Such programs concern the repopulation of areas after wildfires, the restocking of declining populations, or the reintroduction of new populations.
Evaluation of the genetic heterogeneity and connectivity of populations at a fine scale. In the context of habitat fragmentation, the objective is to identify the main features (roads, rivers, forests, cultivated areas, etc.) constituting a barrier to tortoise dispersal and thus potentially responsible for restraining gene flow between genetic groups.
In conclusion, these 32 new microsatellite markers will provide a useful tool to better evaluate the genetic variability and structure of Hermann's tortoise populations and consequently to propose adapted measures of conservation (see the different actions of the Plan National d'Actions; Celse et al. 2016) for this endangered species.
Distribution area of the 2 Hermann's tortoise subspecies (modified from Zenboudji et al. 2016).
Contributor Notes
Handling Editor: Jeffrey E. Lovich
