Evolutionary History of the PER3 Variable Number of Tandem Repeats (VNTR): Idiosyncratic Aspect of Primate Molecular Circadian Clock

The PER3 gene is one of the clock genes, which function in the core mammalian molecular circadian system. A variable number of tandem repeats (VNTR) locus in the 18th exon of this gene has been strongly associated to circadian rhythm phenotypes and sleep organization in humans, but it has not been identified in other mammals except primates. To better understand the evolution and the placement of the PER3 VNTR in a phylogenetical context, the present study enlarges the investigation about the presence and the structure of this variable region in a large sample of primate species and other mammals. The analysis of the results has revealed that the PER3 VNTR occurs exclusively in simiiforme primates and that the number of copies of the primitive unit ranges from 2 to 11 across different primate species. Two transposable elements surrounding the 18th exon of PER3 were found in primates with published genome sequences, including the tarsiiforme Tarsius syrichta, which lacks the VNTR. These results suggest that this VNTR may have evolved in a common ancestor of the simiiforme branch and that the evolutionary copy number differentiation of this VNTR may be associated with primate simiiformes sleep and circadian phenotype patterns.


Introduction
The light/dark cycle that arises from the geospatial relationship between the Earth and the Sun has been shown to be the strongest environmental clue, or zeitgeber, for the entrainment of circadian rhythms in many organisms, including human beings [1,2]. Virtually all physiological processes, such as hormone secretion, sleep, and body temperature which oscillate with circadian periodicity, are regulated by the light/dark signal [3]. The processing of this signal is associated with the expression of a group of genes, collectively known as clock genes, which are directly involved in the regulation and maintenance of the circadian rhythms [4,5]. Due to the fact that variations in these genes, such as polymorphisms and mutations, are associated with aberrant or differential expression of circadian behaviors (reviewed in [6]), the clock genes are strong candidates for natural selection associated with the light/dark cycle signal.
Period circadian clock 3 (PER3) is one of the clock genes engaged in the core mammalian molecular circadian system [7,8]. In humans, this gene is located on the short arm of chromosome 1 (1p36) [9], and its 18th exon contains a polymorphic variable number of tandem repeats (VNTR) composed of a 54-bp motif that is repeated either four or five times [10]. This length polymorphism in the PER3 gene has been described in few primate species, such as chimpanzees, gorillas, gibbons, langurs, marmosets, and orangutans, but not in non-primate mammals, such as rats, dogs, and mice [11]. Several reports have shown associations between this VNTR and human circadian rhythm phenotypes, including morning/evening preferences, delayed sleep phase disorder (DSPD), and homeostatic regulation of sleep [10,[12][13][14][15][16], highlighting an important link between genetics, the sleep/wake cycle and adaptation to the light/dark cycle.
Due to the importance of the PER3 VNTR for regulation of sleep and circadian rhythms in humans, it is a matter of interest to investigate the hypothesis that this genomic structure is exclusive to primates. This would allow rationalizing whether this VNTR is associated to some special feature on the primate circadian system. Since, with very few expections, Simiiforme primates are essentially diurnal or cathemeral animals [17,18] and Prosimians are almost all nocturnal, it would be pertinent to ask whether this VNTR is associated with activity phase allocation during the day or the night. The present study aimed to analyze and compare the PER3 VNTR regions of several primate species and other nonprimate mammals.

Biological Material from Primate Species
Blood samples were collected from 129 individuals belonging to 13 New World primate species (Table 1)  All animals were housed under natural conditions of temperature and humidity in a light-dark cycle. The cages were constructed in brick and wire mesh measuring 0.962.061.8 m, equipped with wood, rope, perches, basket, concrete platforms, cooler box and nest platforms. Water was available ad libitum, and food was provided twice a day: in the morning hours between 7:00 and 9:00 am, and in the afternoon between 1:00 and 3:00 pm. The feed consisted of seasonal fruits and a rich-in-protein preparation that supplemented the diet. Candies were also sporadically offered to animals. At the end of blood sampling the animals were not euthanized. During blood collection, the animals were immobilized by suitable equipment. All animals were in excellent health.
After collection the animals received a bounty of candy. The experimental procedures were in accordance with the Guidelines for the Care and Use of Mammals in Neuroscience and Behavioral Research of the National Research Council.

Extraction, gene amplification, and sequencing
Blood samples were collected into PAXgene Blood RNA vacutainer tubes (PreAnalytiX, Hombrechtikon, Switzerland), and frozen at -20uC. Total RNA or genomic DNA (gDNA) were extracted from the animals' white blood cells using the PAXgene 96 Blood RNA Kit (QIAGEN, Hilden, Germany). The target cDNA or gDNA regions were amplified by PCR using the primers described by Jenkins and colleagues (59 AGCAGYTCACCSTTR-CAGTT 39 and 59 GGYACCTGGTATGTCATGAGAA 39) [11] or the following pair of primers: 59 GACTAA-CAGGTGGGTGGCA 39 and 59 CAGAACTTTTTGGGGT-GAC 39. The PCR amplicons were sequenced, and the sequences obtained were submitted to NCBI GenBank (Table 1).

In silico analysis an comparisons of the VNTR region sequence among primate and non-primate species
Multiple alignments of the PER3 nucleotide sequences obtained were performed using the ClustalW algorithm [19]. The alignments were then visually examined to ascertain the presence of the VNTR and the number of repeats in each primate species. In addition, the PER3 protein sequences from several mammalian species, including prosimians (Table 2), were obtained from the Ensembl Genome Browser Database (http:// www.ensembl.org) and were examined for the presence of the target VNTR by aligning the human PER3 protein sequence and the orthologous region from the selected mammals. In silico searches for repetitive elements surrounding the VNTR were conducted using the RepeatMasker 3.1.8 online software (http:// www.repeatmasker.org).

Results
The DNA analysis of the PER3 VNTR region belonging to the New World primates revealed that all of the studied individuals carried this locus in their genomes, although the number of repeats varied among the different species (Figure 1 and Figure S1). The tufted capuchins and Kuhl's owl monkeys analyzed had only two repeats. The brown woolly monkey, the two grivets and the redhanded tamarin had three repeats, whereas the two black howlers, the two red-faced spider monkeys and the common squirrel monkey had four. The emperor tamarin had five repeats, the two Goeldi's marmosets had six, and the white-headed marmoset had seven ( Figure 1).
The analysis of the PER3 protein sequences from several mammalian species revealed that the amino acid sequence in this region differed both in size and sequence among species (Figure 2). No full repeats of the 18 amino acids of this VNTR were found among the prosimians and non-primate mammals and, although some of the non-primate mammalian sequences aligned with one of the repeat units of the human VNTR, the alignment score was low. Only a few species exhibited scattered amino acid identity to the amino acids present in a human unit of the VNTR sequence. Thus, the non-primate mammals examined in the present study had neither repeats of this VNTR nor a single sequence identical to a human unit of the 18 amino acids. We would like to emphasize that the in silico analysis of the PER3 VNTR region of three species of prosimian primates (Tarsius syrichta, Otolemur garnettii, and Microcebus murinus) also revealed an absence of repeats ( Figure 2).
In addition, the bioinformatics analysis showed that the PER3 VNTR is surrounded by transposable elements. The DNA transposon tigger transposable element derived 7 (TIGD7) is inserted upstream of the VNTR at the end of the intron and a few nucleotides before the beginning of the first repeat. Downstream of the VNTR, a LINE-1 element (L1ME1) is inserted near the end of the last repeat ( Figure 3). Interestingly enough, although the prosimian Tarsius syrichta contains these two transposable elements in this region of the gene, such species does not have the VNTR. Moreover, the non-primate mammals analyzed did not exhibit transposable elements nor VNTR.

Discussion
The present study demonstrates that the VNTR in the PER3 gene is a genomic structure present in all Simiiforme primates studied, although the number of repeats is different, depending on the species. We have found that the number of repeats range from 2 to 11 ( Figure 1). Observation of these results indicates that other similar primates, not studied here, may also carry the VNTR, thus indicating that this structure is present exclusively in primates [11]. Platyrrhini (New World monkeys) exhibits greater variability. For example, the species belonging to genus Aotus (the only established nocturnal simiiforme primate) exhibit only two repeats, whereas genus Saguinus carries up to 11 repeats.
In silico analysis of the flanking regions concerning the 18th exon of PER3 revealed that both introns adjacent to the VNTR contain transposable elements (Figure 3). The sequences of the two transposable elements surrounding the PER3 VNTR region are highly similar in all the primates analyzed, including the prosimian Tarsius syrichta, and are absent in non-primate mammals. This finding suggests that these elements were present in the common ancestor of Tarsiiformes and Simiiformes, which probably lived during the Eocene Period [20,21], and that the insertion of the transposable elements occurred prior to the Table 2. Non-simiiforme primates and non-primate mammal species analyzed in the present study. appearance of the PER3 VNTR. Moreover, the absence of these sequences in non-primate mammals indicates that their insertion into these loci was a primate lineage-specific event.
Transposable elements are known to produce genomic instability, rearrangements, genetic innovation [22][23][24][25][26][27][28], and are associated with the generation of tandem repeats [29]. Therefore, one might tentatively propose that the primitive insertion of these transposable elements in the primate lineage caused the emergence of the VNTR in the PER3 gene in this Order.
The results of the present study, combined with data generated in a previous report [11], enable a reasonable reconstruction of the primate PER3 VNTR evolution (Figure 4). The interpretation of these results largely follows the explanation provided by Jenkins and colleagues [11] who suggest that the VNTR was derived from a single ancestral unit in the common ancestor of primates, carnivores, and rodents. However, data from the additional species analyzed herein and the additional finding of transposable elements flanking the VNTR region suggest that, in fact, the VNTR derived from a single copy that was present in the common ancestor of Simiiformes, and that a duplication event most likely occurred in primates before Catarrhini and Platyrrhini diverged, approximately 43.5 million years ago [20,21]. The repeat thus created was then subject to a variable number of posterior duplications and, perhaps, deletions.
Tandem repeats exhibit mutation rates much higher than those of other genomic regions. When they are located inside genes and regulatory regions, tandem repeats may influence the properties or functions of these genes, such as binding sites, chromatin structure, and, ultimately, gene transcription [30]. A VNTR that is located within a coding region may affect the stability and/or activity of the gene product [31,32]. Although molecular mechanisms underlying gene functional changes mediated by tandem repeats are generally poorly understood, tandem repeat domains in proteins are usually involved in protein-protein interactions and fine-tuning protein conformation [33]. Nevertheless, it is unlikely that variations in tandem repeats produce completely new characteristics; rather, they allow the fine-tuning of specific phenotypes, including those related to behavior, physiology, and morphology, due to the wide range of allelic variation [34], thus switching from most night to most day activity patterns; in the case of primates, would fit with the fine-tuning phenotype proposal. A clear documented case for this phenomenon, for instance, is that variations in the number of repeats of a hexanucleotide in exon 5 of the Period gene influence the mechanism of temperature compensation in Drosophila melanogaster [35,36].
The uniqueness of the PER3 VNTR in primates and its strong association with circadian phenotypes and homeostatic sleep regulation in humans indicate that this VNTR is associated with a  Evolutionary History of the PER3 VNTR in Primates special feature of the primate circadian system or sleep. When compared with other mammalian species, Simiiforme primates exhibit such special features as consolidated monophasic or biphasic sleep [37] and a diurnal activity pattern. The owl monkeys from the genus Aotus are an exception between the Simiiformes [38], and the Prosimians, more primitive primates, are mostly nocturnal. Interestingly, the present study showed that the analyzed Aotus species possesses only two repeats of the PER3 VNTR, and the Prosimian species, none.
It is generally assumed that the ancestor of mammals -and of primates in particular -was a nocturnal creature [39]. Although most of the recently evolved mammals are nocturnal, Simiiformes definitively are not. Repetitive sequences may provide a high degree of evolutionary flexibility, allowing adaptive accommodations at a minimal cost to the genetic function. Thus, perhaps the PER3 VNTR is involved as a part of the mechanism that provides evolutionary flexibility that allowed the occupation of different portions of the light/dark niche -in the case of Simiiformes, the lightened portion of the cycle.
While some genomes used for the sequence assembling have high sequencing depths, such as the genomes of human, chimp, mouse and dog, others, such as the gray mouse lemur genome, present only low levels of sequence coverage. Although a complete coverage of the genomes used for the study is obviously preferable, the usage of lower-redundancy genomes may be the only alternative when investigating non-laboratory model species [40]. While low-percent coverage brings some limitations for sequence analysis, utilization of only deeper-covered genomes would substantially reduce the number of targets, crashing the possibility of a phylogenetic approach.
In summary, the present study demonstrates that the PER3 VNTR is a genomic structure found exclusively in the primate PER3, a gene found only among vertebrates. This VNTR is flanked by two transposable elements that are phylogenetically older than the VNTR itself and that may have been involved in the emergence of this structure. The present findings also show that the nocturnal Tarsius does not possess this VNTR and that primates of the genus Aotus, which are nocturnal, have the smallest number of copies of the VNTR of almost all the primates analyzed.   Evolutionary History of the PER3 VNTR in Primates