Interstitial Telomeric Repeats Are Rare in Turtles

Telomeres are nucleoprotein complexes protecting chromosome ends in most eukaryotic organisms. In addition to chromosome ends, telomeric-like motifs can be accumulated in centromeric, pericentromeric and intermediate (i.e., between centromeres and telomeres) positions as so-called interstitial telomeric repeats (ITRs). We mapped the distribution of (TTAGGG)n repeats in the karyotypes of 30 species from nine families of turtles using fluorescence in situ hybridization. All examined species showed the expected terminal topology of telomeric motifs at the edges of chromosomes. We detected ITRs in only five species from three families. Combining our and literature data, we inferred seven independent origins of ITRs among turtles. ITRs occurred in turtles in centromeric positions, often in several chromosomal pairs, in a given species. Their distribution does not correspond directly to interchromosomal rearrangements. Our findings support that centromeres and non-recombining parts of sex chromosomes are very dynamic genomic regions, even in turtles, a group generally thought to be slowly evolving. However, in contrast to squamate reptiles (lizards and snakes), where ITRs were found in more than half of the examined species, and birds, the presence of ITRs is generally rare in turtles, which agrees with the expected low rates of chromosomal rearrangements and rather slow karyotype evolution in this group.


Introduction
Telomeres are regions of repetitive DNA motifs and associated proteins localized at the edges of chromosomes. They play a crucial role in maintaining the chromosome structure as a single unit, preventing fusions of free "sticky" chromosome ends, or degeneration and loss of genetic information during the replication events [1]. In vertebrates, telomeric regions consist of long tandem repeats of the TTAGGG motif, which was first characterized in humans [2] and seems to be extremely conserved across all studied vertebrates [3]. Telomeres are formed and preserved by telomerase, a reverse transcriptase able to add new TTAGGG repeats based on an RNA template and to compensate for the normal shortening of telomeres after each replication event [4][5][6][7]. However, the presence and, therefore, activity of telomerase varies across cell types and developmental stages [8]. Notably, In this study, we explored the distribution of telomeric repeats from representative species across the turtle phylogeny by analyzing 30 species from nine families using fluorescence in situ hybridization with a probe specific for the telomeric repeats. We combine our results with previously published data, and present an overview of the distribution of ITRs across turtles and compare their frequency to other reptile lineages.

Sampling
We studied the distribution of telomeric (TTAGGG) n repeats in the karyotypes of 30 turtle species representing nine families ( Table 1). Blood samples were collected from captive-bred or legally imported turtles from the Münster Zoo (Germany), Plzeň Zoo, Prague Zoo (both Czech Republic), the Museum of Zoology, Senckenberg Dresden (Germany), Turtle Island (Austria) and private keepers. Experiments with animals were performed by accredited researchers (LK: CZ02535, MR: CZ03540). When needed, turtles were temporarily kept in the animal facility of the Faculty of Science, Charles University (accreditation No. 37428/2019-MZE-18134). The species identity of studied turtles was verified in morphologically challenging taxa using mtDNA sequences that were compared to the sequences published in the revisionary works by Fritz et al. [86] for the genus Cyclemys, Petzold et al. [87] for Pelomedusa, and Ihlow et al. [88] for Malayemys. The Galápagos giant tortoises used in the present study from the Prague Zoo are of known provenance and species identity. All experimental procedures were carried out under the supervision and with the approval of the Ethics Committee of the Faculty of Science, Charles University, approved by the Committee for Animal Welfare of the Ministry of Agriculture of the Czech Republic (permit No. MSMT-34426/2019-7). Table 1. Summary of species analyzed in this study.

Suborder Family Species Sex
Genes 2020, 11, x FOR PEER REVIEW 3 of 18

Sampling
We studied the distribution of telomeric (TTAGGG)n repeats in the karyotypes of 30 turtle species representing nine families (Table 1). Blood samples were collected from captive-bred or legally imported turtles from the Münster Zoo (Germany), Plzeň Zoo, Prague Zoo (both Czech Republic), the Museum of Zoology, Senckenberg Dresden (Germany), Turtle Island (Austria) and private keepers. Experiments with animals were performed by accredited researchers (LK: CZ02535, MR: CZ03540). When needed, turtles were temporarily kept in the animal facility of the Faculty of Science, Charles University (accreditation No. 37428/2019-MZE-18134). The species identity of studied turtles was verified in morphologically challenging taxa using mtDNA sequences that were compared to the sequences published in the revisionary works by Fritz et al. [86] for the genus Cyclemys, Petzold et al. [87] for Pelomedusa, and Ihlow et al. [88] for Malayemys. The Galápagos giant tortoises used in the present study from the Prague Zoo are of known provenance and species identity. All experimental procedures were carried out under the supervision and with the approval of the Ethics Committee of the Faculty of Science, Charles University, approved by the Committee for Animal Welfare of the Ministry of Agriculture of the Czech Republic (permit No. MSMT-34426/2019-7).

Sampling
We studied the distribution of telomeric (TTAGGG)n repeats in the karyotypes of 30 turtle species representing nine families (Table 1). Blood samples were collected from captive-bred or legally imported turtles from the Münster Zoo (Germany), Plzeň Zoo, Prague Zoo (both Czech Republic), the Museum of Zoology, Senckenberg Dresden (Germany), Turtle Island (Austria) and private keepers. Experiments with animals were performed by accredited researchers (LK: CZ02535, MR: CZ03540). When needed, turtles were temporarily kept in the animal facility of the Faculty of Science, Charles University (accreditation No. 37428/2019-MZE-18134). The species identity of studied turtles was verified in morphologically challenging taxa using mtDNA sequences that were compared to the sequences published in the revisionary works by Fritz et al. [86] for the genus Cyclemys, Petzold et al. [87] for Pelomedusa, and Ihlow et al. [88] for Malayemys. The Galápagos giant tortoises used in the present study from the Prague Zoo are of known provenance and species identity. All experimental procedures were carried out under the supervision and with the approval of the Ethics Committee of the Faculty of Science, Charles University, approved by the Committee for Animal Welfare of the Ministry of Agriculture of the Czech Republic (permit No. MSMT-34426/2019-7).

Chromosome Preparation and Staining
Chromosome suspensions were prepared from whole blood cell cultures. Briefly, the cultivation medium consisted of 90

Fluorescence In Situ Hybridization with Probes for Telomeric Repeats
The (TTAGGG) n probe for telomeric sequences was prepared and labelled with dUTP-biotin by PCR, using the (TTAGGG) 5 and (CCCTAA) 5 primers, without DNA template following the protocol of Ijdo et al. [89] and Rovatsos et al. [27]. The probe was diluted in a hybridization buffer (50% formamide in 2 × SSC, pH 7) and stored in the freezer for further use.
The slides with chromosome spreads were aged either overnight at 37 • C or for 1 h at 60 • C. The FISH experiment was conducted in two days. During the first day, the chromosome spreads were treated with RNase (100 µg/mL) for 1 h at 37 • C, 0.01% pepsin for 10 min at 37 • C, post-fixed in 1% formaldehyde solution for 10 min and dehydrated in a 70-85-95% ethanol series. Once dried, the slides were denatured in 70% formamide for 4 min at 75 • C and dehydrated once more with the ethanol series. Meanwhile, the probe was denatured for 6 min at 73 • C and kept in ice for at least 10 min. To each slide, 10 µL of probe was added, covered with a coverslide and incubated overnight at 37 • C. During the second day, the slides were washed 3 times with 50% formamide/2 × SSC solution for 5 min at 37 • C, two times with 2 × SSC for 5 min and once with 4 × SSC/0.05% Tween 20 (Sigma-Aldrich, St. Louis, MO, USA) for 5 min. The slides were incubated in 4 × SSC/5% blocking reagent (Roche) for 45 min at 37 • C and then in 4 × SSC/5% blocking reagent containing avidin-FITC (Vector laboratories, Burlingame, CA, USA) for 30 min at 37 • C. The fluorescence signal was twice amplified by the fluorescein-avidin D/biotinylated anti-avidin system (Vector Laboratories, Burlingame, CA, USA). After this treatment, the slides were dehydrated in ethanol series, air-dried and stained with Fluoroshield with DAPI (Sigma-Aldrich, St. Louis, MO, USA).
For each specimen, at least 20 images were obtained using a Provis AX70 (Olympus, Tokyo, Japan) fluorescence microscope equipped with a DP30BW digital camera (Olympus, Tokyo, Japan). The pictures were obtained in BW and later superimposed in color with DP Manager imaging software (Olympus, Tokyo, Japan).

Distribution of ITRs across the Turtle Phylogeny
Data for the topology on the karyotype of telomeric sequences from the species studied here was supplemented by previously published records in order to reconstruct the phylogenetic history of the presence/absence of ITRs in turtles ( Figure 1). The phylogenetic trees by Pereira et al. [90] and Kehlmaier et al. [91] were used for this phylogenetic reconstruction in the software Mesquite v3.61 [92].

Karyotype Description
We analyzed 30 species with karyotypes ranging from 2n = 26 in Peltocephalus dumerilianus to 2n = 66 in the softshell turtles Apalone ferox and Lissemys punctata. To the best of our knowledge, the karyotypes of the following 12 species are presented here for the first time (

Karyotype Description
We analyzed 30 species with karyotypes ranging from 2n = 26 in Peltocephalus dumerilianus to 2n = 66 in the softshell turtles Apalone ferox and Lissemys punctata. To the best of our knowledge, the karyotypes of the following 12 species are presented here for the first time (

Presence of ITRs
We analyzed the topology of TTAGGG repeats across a wide range of species, covering nine out of the 14 extant families of turtles. All examined species possessed multiple TTAGGG repeats in the expected terminal telomeric positions (Figures 3 and 4). The telomeric signals looked brighter on microchromosomes than on macrochromosomes in almost all species.
We detected ITRs in only five out of the 30 examined species and they were located exclusively in the centromeric positions ( Figure 3). ITRs were detected at the centromeres of chromosomes 1 and 2 in Claudius angustatus and Staurotypus salvinii (Figure 3a,b). In Podocnemis unifilis, ITRs accumulated at the centromeres of chromosomes from the seven largest pairs (Figure 3c), confirming reports from previous studies [76,78].

Presence of ITRs
We analyzed the topology of TTAGGG repeats across a wide range of species, covering nine out of the 14 extant families of turtles. All examined species possessed multiple TTAGGG repeats in the expected terminal telomeric positions (Figures 3 and 4). The telomeric signals looked brighter on microchromosomes than on macrochromosomes in almost all species.

Distribution of ITRs across the Turtle Phylogeny
Pooling together our results with data from previous studies, ITRs were identified in 10 out of 55 examined turtle species. The phylogenetic pattern of the topology of TTAGGG motifs across turtle genomes suggested that ITRs evolved independently at least seven times (Figure 1).

Discussion
All examined turtle species possessed multiple telomeric repeats in terminal positions. We detected ITRs in five species, increasing the total number of known turtle species with ITRs to 10. We detected ITRs in only five out of the 30 examined species and they were located exclusively in the centromeric positions ( Figure 3). ITRs were detected at the centromeres of chromosomes 1 and 2 in Claudius angustatus and Staurotypus salvinii (Figure 3a,b). In Podocnemis unifilis, ITRs accumulated at the centromeres of chromosomes from the seven largest pairs (Figure 3c), confirming reports from previous studies [76,78]. In Chelonoidis duncanensis, ITRs were detected in the centromeric regions of chromosome pairs 1, 4, 8 and 9 (Figure 3d). In Stigmochelys pardalis, ITRs were present in the centromeres of chromosome pairs 6, 8 and 9 (Figure 3e).

Distribution of ITRs across the Turtle Phylogeny
Pooling together our results with data from previous studies, ITRs were identified in 10 out of 55 examined turtle species. The phylogenetic pattern of the topology of TTAGGG motifs across turtle genomes suggested that ITRs evolved independently at least seven times (Figure 1).

Discussion
All examined turtle species possessed multiple telomeric repeats in terminal positions. We detected ITRs in five species, increasing the total number of known turtle species with ITRs to 10. ITRs seem to be restricted to the centromeres, except for the Y chromosome of Elseya novaeguineae, where telomeric-like sequences were amplified in the intermediate position of the p-arm (Figure 1). Interestingly, as in birds [93][94][95][96][97] and squamate reptiles [28,51,98], in almost all examined turtle species the telomeric signals look brighter on microchromosomes than on macrochromosomes. This stronger signal brightness has been previously reported in other vertebrate lineages, including birds [39] and squamates [28,48,98], and it can be caused by a higher number of telomeric repeats on microchromosomes [94]. The species analyzed in the present study cover a large part of turtle phylogenetic diversity from both suborders (Pleurodira and Cryptodira) and nearly the whole variability in the diploid chromosome numbers known in turtles. Despite this variability, only 18.2% of the studied turtle species have ITRs detectable with molecular cytogenetic methods ( Figure 1). Phylogenetic distribution suggests that ITRs evolved independently within turtles at least seven times.
Kinosternidae are reported to have undergone minimal chromosomal changes [99,100], where the main difference between the subfamilies Staurotypinae (2n = 54) and Kinosterninae (2n = 56) is found in a pair of microchromosomes. Centromeric ITRs have been described for the two species of the genus Staurotypus and their sister taxon Claudius angustatus (Figure 3a,b and Figure 1). The only other species from the family Kinosternidae analyzed up to now, Sternotherus odoratus (2n = 56), does not have any ITRs (Figure 1). It seems that the centromeric ITRs are an apomorphy of the subfamily Staurotypinae and cannot be directly connected to chromosome rearrangements leading to formation of the derived karyotype of the family Kinosternidae or its subfamily Staurotypinae.
A burst of ITRs emerged in Chelydra serpentina (Chelydridae), where they are present in seven chromosome pairs, although this species possesses with 2n = 52 the putatively ancestral diploid chromosome number for turtles [76] (Figure 1).
The presence of ITRs in centromeres in Podocnemis unifilis (Figure 3c) [76][77][78] and P. expansa [78] was suggested to be a remnant of ancestral fusions leading to the lower diploid chromosome number in the family Podocnemididae [76][77][78]. Nevertheless, the situation is more complicated. In P. expansa, Noronha et al. [77] detected telomeric sequences only in the terminal positions, while Cavalcante et al. [78] reported ITRs with a similar pattern as in the closely related P. unifilis. ITRs can contain a different number of telomeric-like repeats, which might be reflected in the intensity of the probe signal, and there might be interindividual variability in the number of repeats. Nevertheless, such a strong difference within the same species is surprising. None of the two studies reported a genetic identification of the examined specimens (e.g., mtDNA sequence per specimen), leaving open the possibility that perhaps due to taxonomic misidentification, different taxa were examined in each study. Regardless of this questionable report, the lack of ITRs in Peltocephalus dumerilianus, another member of family Podocnemididae, suggests that the presence of ITRs is not directly connected to fusion events. Supporting this view, ITRs have not been found in Pelomedusa variabilis (Figure 1), as well as in Pelomedusa subrufa [76] (Figure 1), even though it is possible that a different taxon was studied because formerly all ten currently recognized species [87,101] were lumped together as Pelomedusa subrufa. In any case, the genus Pelomedusa belongs to the family Pelomedusidae, the sister taxon of Podocnemididae. Pelomedusids possess a similar or even lower diploid number compared to podocnemidids, and the reduction in chromosome number can thus be a synapomorphy of Pelomedusidae and Podocnemididae. If in the common ancestor of these two families the fusions responsible for the reduction of the chromosome number would have led to ITRs in the centromeric region, the ITRs had to be retained for a long evolutionary time in some species, while they were lost in others.
A connection between ITRs and chromosomal fusions is also not supported for the family Testudinidae. Previous studies did not detect ITRs in Chelonoidis carbonarius and C. chilensis [72,75]. In Chelonoidis duncanensis, we detected centromeric ITRs in four chromosome pairs (Figure 3d), although all three species of the genus share the same chromosome number 2n = 52. We also detected ITRs in centromeres of three chromosome pairs in Stigmochelys pardalis, another species sharing the same chromosome number (2n = 52) typical for the family Testudinidae (Figure 3e). The phylogenetic position and the different topology of the accumulation of telomeric-like repeats suggest that ITRs evolved within testudinidids independently in C. duncanensis and S. pardalis (Figure 1).
ITRs occur only rarely in turtles, especially in comparison to squamates [28]. Wherever present in turtles, ITRs are restricted to centromeric regions and to the non-recombining parts of sex chromosomes ( Figure 1). The distribution of ITRs agrees with the observation that turtle karyotypes are very conserved [85] and that centromeres and non-recombining parts of sex chromosomes are the most dynamic parts of genomes [116,117]. Centromeric ITRs do not seem to be directly related to interchromosomal rearrangements in turtles. If ITRs were formed during fusions, they had to be subsequently lost in several cases. Such a scenario was supported, for instance, by Nanda et al. [39] for birds, possibly due to selection for smaller genome size counteracting the accumulation of repetitions and resulting in the disappearance of ITRs. That the presence of centromeric ITRs is not related to rearrangements is further supported by the fact that in species possessing ITRs, the number of chromosomes with ITRs does not agree with the estimated number of interchromosomal rearrangements for the respective karyotype. However, our conclusions on the homology of chromosomes with ITRs are based on the comparison of chromosome morphology, and should be further tested using more informative approaches, such as comparative BAC-FISH or chromosome painting.
Interestingly, although ITRs are generally rare in turtles, whenever they occur, they tend to be present in centromeres in more than one chromosome pair. ITRs are present in two chromosome pairs in the kinosternids Claudius angustatus and Staurotypus salvinii, in three chromosome pairs in Stigmochelys pardalis, in four chromosome pairs in Chelonoidis duncanensis, and even in seven chromosome pairs in Podocnemis unifilis and Chelydra serpentina (Figure 1; [76]). This observation can be attributed to a phenomenon known as "centromere homogenization" potentially caused by the involvement of telomeric-like repeats in retrotransposons colonizing centromeres [118]. Another interesting observation is that only centromeric ITRs were detected in turtle autosomes. One reason can be that centromeric ITRs are the most common type, at least in squamates [28]. It is therefore likely that they would be overrepresented also in the few cases of ITRs found in turtles. However, it is also possible that in turtles, events leading to ITRs within chromosome arms are rarer. Intermediate ITRs have been found in squamates [28] and in birds [39], and it was speculated that they were formed due to chromosome inversions, which are common in these groups. Turtles, thus, may be less prone to intrachromosomal inversions than squamates [119] and birds [97,120]. As these inversions are important for the evolution of postzygotic reproductive isolation mechanisms, and thus speciation [120][121][122][123][124], this feature may contribute to the generally low species diversity in turtles [70] and their slow molecular evolution [125][126][127]. The low rate of inversions in the karyotype evolution of turtles could explain why they exhibit a high incidence of genomic introgression and why they extensively hybridize (and are often able to produce fertile offspring), even after many million years of divergence, as documented in chelid, cheloniid and geoemydid turtles [81,91,[128][129][130][131][132][133][134][135][136].
In turtles, the TTAGGG motif is restricted only to the centromeric regions. The lack of a strong phylogenetic signal in the distribution of ITRs across all studied species (Figure 1) indicates that the telomeric-like sequences in the centromeres of turtles are not connected to chromosomal rearrangements. Nevertheless, the identification of ITRs is a more complicated process. Genome sequencing projects revealed that short arrays of ITRs, below the detection efficiency of in situ hybridization methods, are common in vertebrate genomes [19][20][21][22][23][24][25][26]. Such s-ITRs are expected to exist in turtle genomes as well. Furthermore, long ITRs, e.g., derived from chromosome fusions, are non-essential, non-coding regions, and like other microsatellite motifs, they can degenerate over time by reduction of copies and/or by accumulation of mutations, which will complicate their detection by in situ hybridization or bioinformatic methods. Previous studies in rodents [27] and birds [68,69] documented that the TTAGGG motif can be part of a longer satellite sequence occurring in centromeres and other heterochromatic regions. Notably, a telomeric-like TCATGGG motif forming long tandem repeats has been identified in the pericentromeric regions of Drosophila hydei, and this motif does not seem to have a telomere-related function [137], suggesting that this could also be the case in turtles with ITRs.
In conclusion, the centromere organization and sequence composition in turtles are currently not well known. This calls for further research to understand whether or not the TTAGGG motif in the centromeres of turtles has a non-telomeric origin, e.g., using long-read throughput sequencing of the centromeric regions of phylogenetically informative species.